If you reset the ladder every year then for a few months of the year there will be a lot of wildly one-sided matches. It also can't be much fun as a 1700 elo player to take a few months off and then suddenly have to spend a lot of time crushing people to get back up to 1700. And the other way around as well, people that found their 400 elo suddenly need to lose a lot again.
Incidentally, I just re-watched SoTL's take on this question just a few days ago, so I have his conclusions fresh in mind - as well as his in depth explanation of how the elo system works. He suggested more placement data be used before dropping new players into the ladder, such as placement matches vs AI, data from other ladders the player might have played in (team games primarily) etc. Personally though, I think the easiest solution would be to have an higher, but steadily dropping K value for the placement matches. Presently the 10 placement matches a new player plays through all have a K value of 100, which means if you match up against a 1000 player in your first match and lose, you'll lose 50 elo points. That will only get you to 950 obviously, so if your true elo is 500, you'll have to lose a lot of games. If the first placement match had k value 600, and then was reduced by 50 for each of the following placement matches or something like that, getting to true elo would be much quicker. Sure you might bounce to 1300 right away if you're lucky or play out of your mind somehow, but that will quickly get corrected.
I reckon its a matter having the ai compete in ranked. And putting your starting elo matching that of the ai corresponding to your best offline performance vs ai 1v1 with ranked settings... I dont see any drawbacks here!
Sounds good, except for one problem: The high K value affects the opponent too. Imagine being a 1100 player that's been playing for a good while matched with a 1000 beginner and winning 300 points. That would make you jump to 1400, way above your level, then you're forced to deal with losing over and over again to actual 1400 players because of the low K value that's applied in those situations. It would take a huge loss streak to get you back where you belong, so you have the same problem you saved the 1000 beginner from. It feels like every solution has a major drawback like this...
@@danieltoth9742 Just don't apply the K value to the established player, not even sure this is the case today. But I definitively agree that the real solution is getting to your real ELO fast within the first 5 matches, when you loose 5 and you're at your ELO then you will have more fun. Any playing around with the ELO numbers doesn't make too much sense, like a reset Hera suggested, the ranked ladder would be unplayable for some weeks due to all the mismatching ELO level matches, except if the reset means that everybody has to replay their placement matches, but even like that, not sure it would be playable for some time.
@@mramosotero Are you suggesting the established player should gain fewer points than the beginner lost? If so, that would take points out of the system, which gradually lowers the average rank -- meaning those overpowered 1000 players are going to be 900 players... then 800... then 700, and so on. They would follow the beginners in a race to the bottom, making everyone feel like they lose more than they win.
@@danieltoth9742 Yeah sure, the ELO is based on points exchange between players, but it's not worse than making beginner players start at 500 ELO for example. Back in the years 2000s the starting ELO in MSN Gaming Zone was 1600, but for me that doesn't change anything, so even if the system would lose some point over time il wouldn't be a great deal. Nevertheless, one solution would be to keep a points credit for the beginner player that would be "consumed" over many games. For example, the new player loses 500 points in 5 games, but the opponents only gain a total of 100 from those games (20 each), so he keeps 400 points that will be credited to his next opponents over the next X matches. Then when he loses a match the opponent would get double the points (which doesn't make a huge difference at an established level) until the credit is depleted.
Grubby, the Warcraft 3 player, would argue against your proposition. And I would agree with him. Ladder resets just make the placement matches a mess until the ladder is settled again, having more unfair matches during that time.
i´m not too sure if you are right about this one. there are hard resets and soft resets. hard resets would reset your mmr and ranking points etc. you basically start from the beginning. soft resets just reset your rank. the community made ladder in warcraft 3 "w3champions" does soft resets and it works great
There are people who persevere... people who quit after initial losses...And then there is our Hero "lucaitasa" having fun at 22 elo ( o7 ) playing there for years :_D
Played a guy at 1050 the other day who did a perfect 2 group scout rush into good castle time with Knights all whilst full house/palisade walling his base.
I used to think so too, and then I played a game vs 1800 elo player (as a 1000 elo player). The 1800 elo player basically has 2-3 scouts in my base when I'm halfway up to feudal age and haven't even started building walls or making a barracks. The skill level is definitely much higher at greater elos, and the scout rush at lower elo is nowhere near perfect.
Minor point: i think ur underestimating how many ppl play custom games, set up team lobbies w friends or simply play unranked. Like the choice is not binary bw ranked and campaigns
But the skill difference is even bigger in custom games. Lobbies will be called “noobs only” and nobody knows what that means. For some peope everything below 900 is noob level, for some it’s 1000, for some it’s 1200.
"1300 already start doing micro and build orders." I havent played the game in months, but to my knowledge even at 1000 people try to have perfect build orders and very good level of micro, hit and run and saving each unit in the whole army etc.
I agree. But it's also very inconsistent. If I get spanked it tends to be by a RUclips build order. And if they play naturally I often do the spanking. Generally speaking though most people at 1k can do a basic build order for every opening unit. @@M0rdFustang
As someone who hates micro and likes to just sit back and have a nice good old fashioned macro game. I notice I have to deal with archer micro bullshit (stutter step) down at 950 elo, even had monks once. Used to be that people didnt do that shit until about 1100 elo.
I think a 1000 elo can execute a perfect buildorder without being able to analyze if that is the right strategy to apply. In addition he/she would do it without being able to analyze why he/she won or lost. So, it’s like they are playing a different game. You could have a perfect buildorder but it could not make sense because of your civ , the opponent civ, the map, the behavior of your opponent, the strategy of your opponent , the unit each one is opening, if there is water or not in the map, etc etc.
@@franok9733 at 1k elo' though the build order they follow is normally civ specific. If they go random I find they know roughly what each civ should open with and do the basic build. It's genuinely a lottery in terms of what you might face 900-1300 elo'.
You’re mistaken, Hera. The placement matches would be hell for half of the community. They just need to do the same system for new players that they do in chess, where new players games count for a lot more elo points up and down so that you find your elo more quickly. Or something along those lines however it works. If you open a new chess account you will be roughly around your elo after 5-10 games. Nobody loses their first 20 games in a row that way.
I think the important thing is that it counts more for the new player, but less for their opponent. I think in the Elo version used in AoE the win/loss is the same on both, so everyone playing against several new players would immediately be catapulted up, and then have the same problem of being in the wrong ladder region. (And people playing against new smurfs would have the opposite problem.)
1000 Elo is indeed a bit brutal right now. I just picked up AoE2 seriously a few months ago, but i played campaigns as a kid and played Starcraft 2 ladder for quite a while all the way back in the Wings of Liberty era, so I've been around RTS for quite a while and am familiar with the importance of build orders, worker production, expansions, etc. I still ate shit against 1000 elo players for several matches. I only got down to like 800 elo before bottoming out, i can't imagine how rough it must be for someone new to RTS who loses for 20+ straight. I don't like the reset, that just creates a couple weeks of volatility while everyone wins/loses their way back to their true elo and then we're back to status quo. I think a better solution is increasing the ego changes you get early on (ie losing your first match knocks you down 2-300 points) so that you more quickly find your level without an epic losing steak. Alternatively, you could let players sort themselves a bit with questions like "what is your experience level with RTS games" and then start them at 600, 800, 1000, or 1200. That might cause some smurfing, but with aggressive early elo settings that would only work for a few games per account.
- Start players on a low score - Create a lot of score out of air if they win placement games - Keep track of how much score is in the game as a result of their account - While there is more score from them than there should be, destroy extra score per loss (which the opponent doesn't get). While there is less score from them than there should be, create extra score per win (which the opponent doesn't lose) That keeps the total Elo score at 1000 per player while allowing players to start at 500 Elo if they're new
This is a good idea. I would just add a possible improvement. Take for example a player that starts at 700 ELO so 300 ELO is destroyed into thin air (which we need to refund to the ladder somehow). During the placements, we would be giving this player more ELO than their opponents lost, so that the ELO can be given back to the system, and taking less from them than their opponents get. This means that the player would have to lose more than win, in order to stay at 700 (assuming that is indeed their true ELO and we placed them correctly). This would still feel to the player that they need to win more than lose in order to get to their ELO. Of course, this is much better than losing long streaks, but maybe we can do better. Maybe we can limit the amount of ELO generation or destruction during this phase to some small level, so that the win rates of this player are not far from 50%. All the ELO that does not get resolved in due time is simply sprinkled to over all the other players, which means even if this player stopped playing after initial match, and got 700 ELO, the remaining 300 ELO would stay in the ladder, and the average would still be 1000.
Same. I’m no 1500 Elo but, I was constantly at 1200 and I took two months off and I have played for two weeks and I am back to 1k. I don’t think I have gotten that bad but, everyone is getting better.
The elo gets deflated as the average player skill increase, a 1500 from 2 months ago is not the same as a 1500 from today. You're probably no longer a 1500 elo anymore, probably a 1100-1200. I always check aoe2insights when I start a custom game with a specific skill range precisely because of that. Some people are just happy to reach certain ranking and stop playing but that ranking means nothing if it's not constantly updated.
i had the same experience, i was playing aoe2 like 3 years ago and i climbed to 1446 elo back then, when i came back i did some little training with AI and study to just get familiar and i started to play ranked games, first game i won and then the lose streak began, i was losing way more than i was winning and i dropped back to 1220 elo. After a week of play i climbed back and 2 weeks later i almost hit 1600 elo for the first time. Which is nice but i have to play very well to bet my average opponent of 1550 elo. The weird thing is that when im really on it i win so easily even with 1500+ elo but then i get to 1600 and i start losing again. Mostly because my build orders are wack.
@@miguelbalderas905 Dropping 200 Elo can happen even if you are constantly playing, thats most of the time just temporarely if you have some days you just play worse than normal.
Here's another take: The issue is not necessarily losing several games in a row, but that there is a lack of communication between new players and experienced players. I am at 1000-1100 elo with about 1000 games and so an average player. I'd like to know if the person I'm playing against is going through placement matches at the START of the game rather than at the end when I see how many games they played. That way I could go easy on them and rather chat with them, offer some encouraging words and ask how things are going. I think that would be a way of helping new players feel welcome despite losing a lot in the beginning.
I really like these videos with the continuous gameplay in the background and the chill voice other. Keep up the good work and let's hope to convert some more of these campaign players someday!
Best solution: 1. require art of war completed before multiplayer 2. Have a selector that asks, "What would you classify yourself as?: Complete beginner, beginner, moderate, experienced" 3. That selection puts you at 500 elo / 800 elo / 1000 elo / 1200 elo respectively
Smurfs would have a field day with this one. And new players could genuinely misjudge their skill level, like imagine if most new players read about others having a bad experience they are all going to choose the lowest option and it would get overcrowded quickly.
I like this. If you self score experienced and lose its your own fault but you're also the person who doesn't care and will keep grinding. I think just leave it up to people to chose where they enter. Why we forcing people to come into the game at national level, its ok to want to play community level and naturally climb. Also why don't we take the medals from Art of War. Like if you get no medals (i cant remember how medals work lmao) you can only place in beginner.
@kirtanamrita2302 I don't really get these comments, I don't think I've ever come across this problem, maybe I'm lucky but I don't think it's an argument for not making changes... like can't you just do it now anyway? So they'll still do their thing, but this just makes is easier for everyone else. I don't see a meaningful downside
I rarely play 1on1, mostly teamgames where mit ELO fluctuates between 1050 and 1200. But a few weeks ago I played a noob 1vs1 and I didn‘t know, gave it my best and picked the poor guy apart…I felt really sorry after seeing it was his first game. The gap between a beginner and 1000 ELO is huge!
I am a good chess player, not titled but honestly I am at least a national master strength on a good day. I am ranked 99.8%. Chess has two factors in the elo system ranking when you start playing. One of them is a provisional factor. If you start playing as a brand new player and demolish a 1200 player, demolish a 1700, demolish a 2200, you might be playing a 2400 player very shortly, whoever put together the elo ratings and how it evolves did an atrocious job. It would never fly in chess. If you start losing 5 guys in a rows at 1100, 900,700, 500, 300, you might be playing like a 200 elo player after just a few losses and those last games even if you lose would not be stomps. The algorithm for rankings is just really messed up.
A guy getting stomped 20 games in a row is just a sign that the system is broken. Honestly, that’s something that kept me from playing for a while. I just didn’t have the mental fortitude to deal with the rust that I’ve built up in a stressful aoe2 environment.
@@neildutoit5177 But what about reintroducing that elo to the system during the early games of the player? Using 700 as an example, lets say that for the first 10 games, they win 40 elo if they win, but go down only 10 if they lose. I dont lnow if those are the right numbers, you would need to find the starting point where, on average, a new players wins half of their placement matches, and then adjust the the elo reward to match that so that, on average, each new player is still introducing 1000 elo to the system. It would help overqualified new players to climb faster, but it would help new players get some wins early by matching them with weaker oponents. Or is there something im missing?
The solution for this is very simple: up the k-value for the first 10 matches massively, and up the k-value for a players' next 20-30 matches a decent amount. The k-value determines how "swingy" your elo is and how much rating you win/lose from a loss. This way a noob goes from 1000 starting -> 850 -> 700 -> 600 -> 500 and stabilizes there as opposed to going 0-20 to get down to his "true" elo of 500. Similarly experienced players with new accounts would shoot up to 1500+ quickly and not noob stomp as much.
I've played the game on and off for years. Not competitively, but I knew the basic build orders, unit counters and what all the upgrades do, etc. When I jumped into Ranked games in DE a couple months ago, I got smacked down from 1000 to 700 pretty instantly before I managed to level off.
I have same kind of background. I got interested in ranked play but I did not want to lose all placement games so I trained against AI, getting build orders right and practicing my multitasking. My first ten games were 5-5 and I'm happy with that. All that time spent playing campaigns and against AI on and off for years definitely helps but ranked is slightly different beast to tame.
Hera: "1300 know the basics pretty well and begin to execute build orders" Me at 800-900 watching my opponents getting to castle age at 14-15 minutes and wrecking me shortly after: 💀
Watching this video has given me a lot of confidence and I do have a suggestion at the end of this comment as a potential solution. I recently (June 2024) got back into the game after 20+ years away. In those 20 years, I've become a history teacher, chess coach, and eSports coach and I've used my knowledge of chess and attempted to apply it to AoE2 through casting T90 Community Cup and AoE2 Specialist Cup games as well as participating in said Specialist Cup. When I took my talents to the ladder, I won my first game in my first game but then proceeded to lose the next 9! Your advice on having fun first is great because as a coach, that is what I stress to my students, but I often forget that for myself. This is truly a game similar to chess where you have to lose a lot of games before you win a few, so thank you for your encouraging words. One teacher to another: being a calm encouraging presence is key and you've got that! MY IDEA: A season reset is a decent idea, but instead I'd like to see a ranking system based on chess, which uses ELO as their ranking system. In that system you'd have Classes of players akin to chess so Class F would be 0-1000 ELO, Class E would be 1001-1199, and in increments of 200 until you run out of letters and get into Master categories at 2000 ELO. Then you'd have like say Expert from 2000-2199, Master from 2200-2399, and Grandmaster from 2400 onward with Super Grandmaster around 2600. Each level has a FLOOR which can be achieved and a player cannot go below that once they achieve and retain a certain class for a certain period of time. This would also have the benefit of preventing sandbagging. No system is perfect, but this would be an interesting way to go about things I think. Again, thank you for your time and your talents!
as someone who studies lots of statistics The ELO system is basically a mathematical model in which, under the premise that higher elo means higher % of winning a match vs people with lower elo. This difference can be graphed and modeled and, as usual, this difference follows a normal distribution and you can assest a proportional reward/punishment for wining or loosing. Which is why, for example, an elo 1200 winning vs a elo 1500, the player with 1200 elo will win a loooot of points and the other player likewise will loose a lot of points as well, this is because the statistical probability that a 1200 beats an 1500 is pretty small, on the same side, if the 1500 elo wins, he will win very very few points. But this concept must embrace also the concept of iterations, you can never see an algoritm and understand properly this concepts without the concept of repeating the same "game" of probabilities time after time And, after various iterations, you will loose some, win some, the trend will dictate your ELO. In the perfect system, that only mathematics can imagine, infinte numbers of iterations will leave you your "true and objective ELO" But that is of course imposible, materially, technically, economically with that in mind, we know that it is imposible to create the perfect algoritm with perfect representation of each individual skill on any given time, only trough time and iterations we can, to some extent, "measure it"
Go statistics!! I think they're using a logistic distribution instead of normal, which is actually normal😅. There's a lot of things to consider for a rating system, inflation, min rating, commutativity, the famous K factor, sigma values etc. However if you know exactly what you don't want to happen with your rating system, it's not that hard to engineer one. As you said Elo can't messure real "strength" and no 1 dimensional rating system can (maybe infinite dimensional), but it's good enough for most cases, in fact it's better than any tournament outcome ranking (Unless it's a round robin and the points you get from each match are your rating system usual points, but that makes a 10 player tournament 45 or 90 matches long)
@@EneldoSancocho yeah indeed it is often use the LOG function to soften the bases, but like you said it is not that hard to implement some quirks or restricción in an ELO algoritm
@@sebastianpalominos3706 Man I'm honestly happy every time I encounter someone who studies math. Probability and statistics are beautiful subjects and they are full of counter intuitive problems (like 2 envelope paradox and such), they also are related to so many applications in different fields. If you are interested: there are alternative formalisms for prob theory for which you kind define a uniform over the integers 😃
What you say resonate a lot with me too. I played a little when i was a kid. Last year i started playing a lot with 3 friends who where noob too and I demolished them. I played basic campagnes, played even a lot vs difficult IA.. But online vs real players is different. I got crushed my firsts 7 rankeds and imediatelly stopped playing online. I needed a full month training, making around like 100 games, restarting every time i was feudal too late, watching tutos, to have my first win. And even then, it was close.. So i didnt continue for the moment. I feel like i need more training. I think even 800 elo is too high for new players. I dont think there one global solution, but maybe one big step can be to propose to new accounts to play 5-10 games vs adaptives bot. So if you are a noob like me, after a few games, you can start a bit more at your level. The big dream could be to have real short recaps in the end, on your weakness during the game. So beginner can know easily what to work on first (Late ages, how much time exactly when your Forum was not working, bad balance of ressources in early..)
I'm a long time player (started following the scene before the first Hidden Cup to give you context). I stopped playing for a long time and only recently got back into it. The whiplash and skill level differences between then and now is insane. I think a large part is because of great creators like Hera making high-level insight available to all. But it creates a weird dynamic where people like me (that appreciate competitive play but like to mess around while playing) keep bouncing between stomping the competition and getting stomped. Reworking the ladder system to have a more nuanced view of skill would be a good start to address this. Or pushing the unranked scene more (it's so much fun and I barely knew about it till I came back to AOE).
Great topic, thanks for bring attention to it. I would love to see ranked seasons with some minor reward at the end showing your season/rank or highest rank achieved badge, etc. The details for the ranked system would need to be dialed in first. I hope the devs will gather ideas from the community, then build a new ranked system. For example- your last seasons rank can be taken into account when you play your new season placement matches to avoid a noob stomp. Since were on the topic, lets get some updated menus as well. The main menu is.. fine, but the lobby menu and ranked menu feel a little "windows 98". Are the devs limited by the software, or just trying to keep the nostalgic feel?
As someone who just started playing age of empires 2, I have zero desire to get into the multiplayer. I know I will get absolutely demolished if I try it. Single player provides enough entertainment for me and I can challenge myself by increasing the difficulty without the stress of facing a veteran player
I played 4 times vs incredible strong players (compared to me), the last one harassed me via chat, and so I went away from multiplayer. Playing vs AI is fun too and watching pro gamers shows the skill ceiling.
well said. been playing ranked 1v1s as long as Hera has known of me and longer even. over the years i've managed to enter a stable 900-940 elo, but that has come with so so many ups and downs a *metric shit ton* of watching everyone around me get better very quickly. these days 900 means something; 1k3 *really* means something again, well said. good video addressing a point that often goes undiscussed
There is only one solution... lose more elo per game lost in a row. A multiplier of some kind, losing 10 games in a row would have you at 500 elo as opposed to 800 or whatever it is now
@@neildutoit5177when someone loses x elo after losing, the winner does not gain x elo. The lower the rating of the person who lost, the less elo gain the winner gets
No sir, not per lost games in a row. They need a provisional factor like they do in chess. When you are starting where your rating can really bounce around while they try to find your true rating. It should not take 20 games. They should be pretty close after like 5 games.
I basically spent 3-4 weeks researching the game and practicing build orders before I even went into a ranked game. I think if I went in completely blind with only my knowledge of playing as a child I would have lost 20 games as well.
Hello Hera, I am new in game (1 month) and I am 920elo atm. I started to watch your videos about "How to win in low elo" and etc, and I was just keep getting rekt :D. And believe me, I was following every instruction! I really feel that ~900 elo players should be worse than that, or this is how it used to be in Moba games that I used to play. Then I start thinking that this elo, while it is actually low, it might not considered that low in AoE2. xD *Edit: You really helped me to reach this elo in a month, by just watching your streams, not the videos about low elo :D
i recently began playing aoe 2 again after 15 years and first played unranked and after 20 matches started rank with around 900, I resigned multiple games so my Elo goes down intentionally so I play against my own caliber, right now im 600 and have 9W and 18L overall in ranked and I am enjoying the matches :) ty Hera for the videos. I would love to see more in depth videos about Civs and how to play them.
So you're telling me if I went back to ranked after not playing ranked in a couple years I would get wrecked? I'd have to check but I think I'm currently around 1000 elo? All joking aside, I have been playing through campaigns lately because I got all the dlcs and I wanted to play them all, so I definitely don't feel like a new player (600+ hours on DE); I just haven't played against a real person since I got back into the game a month ago (don't worry I do intend on playing ranked soon and won't mind losing if that's what happens; I can always lean on my Huns if necessary!) Also I'd say that from anecdotal data it seems like there are a lot of og players from back in the day (like my dad) who simply doesn't play online because he isn't interested in playing fast enough to keep up, so I'd imagine that players like him probably will stick to playing against the ai
Different solution, just do what League does, have Bronze 3 for true beginners and have people rank up if they win a bunch and rank down if they keep losing, that way you always get fair games but the starting rung of the ladder is true beginners so no one gets discouraged
I am currently at about 750 elo and most matches I encounter are people with +200 ranked matches, and I'm a newcomer, never played AoE (I did play some AoM but they play very differently), bought the game in the 25th anniversary which was like 1 month ago and I've been playing online 3 weeks now and Im watching content and striving to do better every time, every match, I'm seeing myself progress trough not wasting food, walling my base, choosing better my armies and my upgrades, and all due to the vids of the AoE 2 community that helped me. And yes, 1000 elo seems like its gonna take a bit of work XD
I'm glad that you are informed on what happening on the ladder at lower ranks than your own. I haven't played in a very long time but when I did today, I got regularly beaten by people less than 200 Elo below me.
I pushed through the loss streak 1 year ago with the mentality that every loss is a lesson, I checked replay with superfast forward brief analyses, what my opponent did and how he overwhelmed me. Then I went next. Lost like 8 games until my first win. After I stabilize my elo, I started learning more and absolutely enjoying the 1v1 multiplayer (I rarely play 2v2 or 4v4 - too much rage quit dumb-dingos). Currently sitting on ~1200 playing casually few games a week (pref. Arabia). I find the most important think is build order and get the feel how to distribute your economy to being able to produce vils and your army spam choice (including how many military production building you should add once you develop eco more) and being able to do tech switch and adjust resource gathering ratios accordingly - this is what I find to be the core of all the game, also most demanding on multitasking. I don't focus on micro very much yet, but eventually I'll get there once I'll be more efficient with eco management -> more time to focus on the army and become a micronerd community member. Great video as always Mr. Hera. P.S.>> I encourage all of you, people who read this, don't give up with multiplayer, once you get better, it's a lot of fun and the community is super nice on discords so you can discuss and develop your own playstyle with favorite civs. Stop caring too much about your ELO, winning or loosing, be positive, call the GG every game, you can check with your opponent after game, maybe he'll give you some tips for the future (happened to me many times) and trust me, if you suffer the ranked anxiety, this will really eliminate most of it, after all, it's just a game and it's okey to loose and raise stronger next games. ☺
Yes best advice .. people stop care about elo. Care about having a good game . If your elo is higher just means more chance to lose. Your ELO will match your ability after time giving you a better experience winning 50% of your games means you are at your current skill level.
Agree with having more in-game communication systems. I find that I enjoy the post-game analysis with my friends and discussing what went well and what went wrong. I would like to do that after ranked games as well with my opponents at times. If the post game screen had an AI analysis of the game and tips to get better and ways to share/ communicate with opponent & friends, even your losses might end up teaching you something and you would feel more positive about the experience.
I'd always wanted to play ranked but didn't want to get destroyed. So I watched a bunch of build order tutorials and created my own spreadsheet showing what to do with every villager up to Castle Age (based on research I found online) and practiced this against the AI a few times to get the muscle memory for it (and watched the games back on CA to analyse what I did). After that, I jumped into my first ever ranked game and made some silly mistakes but nothing too major with the build order. However, I lost because the opponent was simply more aggressive and spent their resources better, which I only learned through watching the game back on CA (also it was Arena and I stupidly let the enemy's troops into my base so I just GGd). Then I went into my second game and played a lot better, making a note to actually spend my resources (especially building multiple military buildings for faster unit production). I got very lucky spotting a couple of the enemy's attempts at sneak attacks, and just managed to out-play them and eventually win. It was so satisfying getting the win, and it was only really possible because I'd done so much prep beforehand. If I didn't know anything about build orders then yeah, I'd have been crushed, because both my opponents had enough of a grasp of the early game that I'd have fallen behind significantly. I agree it's absolutely wild to start at 1000 ELO. Like, I know if I got matched with a player who was playing their very first ranked game with zero prep or research, I'd likely destroy them. But equally, if I got matched with a 1000 ELO player with multiple games who knew what to do then yeah, it'd be a very difficult and different game. But equally, you only get better by playing opponents better than you - there's no point constantly playing against people below your level.
i had played the game in single player for almost 20 years before ever trying an online game and it was in a custom lobby with friends, we played quite a few matches over a couple months that way and i noticed that, while i had the game knowledge, i was really not used to playing against humans but over those months i suddenly got way better than i ever was at the game and eventually i tried playing ranked where i played like 6 matches, won 3 of those and then decided it was too stressful for me lol point is, i think a lot of new ranked players had similar experiences, getting good at the game while playing with friends and then actually coming into ranked fairly competent, completely unfair towards truly new players
I play for 3 years now, watch every Hera video, check meta play on T90‘s Channel with Red Phosphorus but still can‘t exceed 1400 ELO. The Online Community is insanely strong!
People are obbssed about their score instead of understanding their score is their current ability and will give them the best matches .. do you want to be 1900 and lose every match ...
I took a break from the game about 6 months ago and came back to find kind of the reverse effect for me. I jumped from 1200ish elo to 1600 after a couple weeks. I would go against either extremely strong opponents or those that would feel more like beginners still. It might not happen as often but I feel like your thoughts on 1000 elo also apply in other areas too. :)
What if you actually start at 0 Elo and there is no reset until you reach 1k? Everybody above 1k gets the seasonal reset and maybe there are even leagues... like if you break 1300 you only get reseted to 1300 and if you reached 1600 you get reseted to that, never lower, so you don't have to grind your way back up to your current elo range?
just recently returned to the game after a very long absence (stoped playing when hunwars where a thing) because it seems I can enjoy playing vs a lot of civs again. the "art of war" challenges do help to get back into the game for sure but they don't go far enough. like getting into feudal/castle with x vills asap is nice and all but they don't teach anything what happens after that. so can my vills just keep sitting on wood/food for the most part or should I shift them because I want to add TC/castle/stables+units or whatever? if so, when do I need to shift them to get my desired result? that is for everyone to find out himself. now making buildorders is not much of a problem for me since I used to do that back then as well but for realy new players, that may be to much. especialy considering that you can comfortable beat hard DE AI without much of a plan. I did so yesterday on arabia with only: wall your base + reach castle with a few archers for defence asap into boom. as nice as it was, I doubt I would see much light when starting ranked.
I've never played AoE online and, thinking about it, the game came on disks way back when I first started playing and I didn't even have internet at home. Therefore, whenever I played AoE as a kid, it was offline and that's just how I've always remembered playing the game, so, even after getting it again on Steam, I continued playing like I always had without any interest in online play.
I think i can relate to this. Was a new AoE player myself, literally brand new 2-3 months ago, but i sort of had an idea for the game from just enjoying watching high level play and having played the campaigns as a kid. I learnt build orders and everything before going ranked and even then i still got my ass kicked most of the time. But i think the problem is that 1k elo is inconsistent. I would meet a guy that had no idea what he was doing, and next game i was playing terminator and got my ass kicked HARD. after ~20 games i ended up clearing up my dark ageand naturally climbed back up to 1k elo. Took me arround 50 or 60 games to break past 1050 because it just was so wildly inconsistent. 50 games and a month later i'm sitting at ~1250 and i'm still having some games where i match with way higher skill players but overall its okay. Its just the 1000-1050 range is completly messed up. Most games nowdays im still loosing due to experience gaps and knowing what units and strategies to go for rather than skill (which is where your discord is actually very nice, always learning), so i can only imagine how bad it can be for someone stuck at 1k elo with 1000+ games
Thanks for the video Hera! Im 1350+ elo since 2021 i just dont find the way to improve. Im spending less time playing now because work and life, but it wasnt a matter of grinding, i just coudnt got better.
I started 0-2. And quit. But I went to the quick play mode (no idea how that matchmaking works) and played like 100 games with good results (probably less than 50-50, but it felt like progress). That helped me get the confidence back to jump to the ranked matches again. I think clarifying that progression can make sense for some people. Quick play felt stress free, and the vast majority o the community is helpful if you ask, so you actually learn from people that are better than you
I have some background in game theory, and math overall. One solution to try is just add + 200 rating points to everyone. The reason for this is that AOE2 doesn't have adequate incoming number of players every year so there is no "feeding". Normally new players are supposed to come in that give rating points to good players and their rating goes up. But due to shortage of new players, everyone just exchanges points among themselves, even though community is getting stronger, their sum of ELOs is constant. So add a community growth value every year to every player.This doesn't have to be constant for example +200 at 1000 rating, +50 at 2000 rating, +10 at 3000 rating. Once the incoming number of players are high enough to compensate for community skill growth this adjustment can be stopped. A stagnant game where new players are not coming but skill keep raising we have to keep doing this.
Exactly. There is hardly any new players coming into ranked and is exactly the problem. You could also give more for win than loss which would push better players higher than 1000 .. who cares if the highest ranked is 5000 or 20000 it doesn't matter as long as the score can reflect the skill difference.
I was waiting for this video... When I started about a year ago, I got steamrolled for about 20 games, before being able to adjust. I'm still sub 1000 and players at this level know how to wall, some even quick wall. They do follow build orders, and they are really fast in aging up. It's very competitive.
I definitely agree with heras point in ELO. I was a 1000 Elo player back when DE was first released and I was doing ok with some win and some loss. I recently got back to the game after practicing against AI, using the same build order (fast castle boom), and got stomped for like 10 games in a row.
Been thinking that for a long time, glad to see a big name addressing this issue. I think setting initial ELO to 500 solves it short/mid term. Now beginners will play against other beginners or "actual" 500s, while actual 1000s continue playing among themselves. Long term perhaps more changes are needed. I think something like making the first 5 games worth triple the points, and the next 10 games worth double should do the trick.
This is very true, after playing SCBW for 20 years I thought i'd reach like 1300 ELO within a month or two... but the game is very complex, gotta balance between 4 different resources, build orders are quite complex, there are tons of things to know, what unit counter what unit in which numbers. so far most games are either weird rushes i have no clue to deal with and I lose (lmao sicilian dungeon rush is ridiculous), or we turtle till castle/imperial and I usually win because I got much better late game mechanics but my unit composition is usually awful and it's still a complex game to figure out. The 1000 ELO guy in AoE2 has some very very decent knowledge of how to play an RTS. I mean even at 900 ELO people know how to push deer properly in dark age, know interesting rush build orders etc. Respect to AoE2 players, seriously.
An easy to implement solution would be to regularly adjust all players ELOs so that the average of the distribution is 1000 ELO. After that put in place, you can give newcomers a free pass (ina certain range) on where they want to start on the ladder. A recommendation system based on single player history also makes sense.
You can do an ELO inflation correction, like FIDE and national chess federations did. And maybe have K factor work better, that you have smaller changes if you're an active player, and if inactive you'll drop quickly due to higher K factor.
Some good points. I haven't played ranked since DE release and I haven't played aoe2 for a couple of years now. I'm still learning all these recently added civs. I really would like to play some matches again, but I think my almost 1300 rating from back in the day is a bit inflated now and I will get my ass kicked. I'm not even able to do a proper Dark Age anymore ^^" ... so I'm staying in my comfort zone in SP and practice against the AI haha
I'll be honest. I regularly watch you and T90, and never played an online game. I really like playing with my friends or vs AI, even managed to beat extreme AI couple of times, but I just prefer not to play with strangers. When I watch and try to evaluate my hypothetical ELO based on the players I watch I would guess it's a little below or around 1000. The fun fact is that I tried twice to play on-line ranked and start this whole ELO adventure and both of the times the servers were undergoing maintenance, so I guess it's not meant to be
If you consistently beat hardest or extreme AI you are around 1k in my experience. It's just that AI usually does the same things and doesn't wall so you'll have to adjust your own game play a bit when going ranked.
Playing with strangers is how you learn the game. You do not have to play against Extreme AI either. Just make a lobby “4 vs Hard AI” and play. You will eventually learn what other people do, adapt it into your gameplay, and become much better. Watching videos means nothing because, until you can quantify how good that player really is, you cannot even begin to compare yourself to what you see in videos. It’s like watching the English Premier League or La Liga - we have no concept of how good these footballers really are until we see them in person.
This is just how competitive games go. Especially when the casuals leave. I was gold level in Starcraft 2 in 2010 and obviously I had no idea what I was doing in Age of Mythology back in like, '01.
I'm a 20+ year single player only player, and viewer of the competitive scene. A few years ago I tried my hand at multiplayer to see how I would do. And I basically won streak until about 1300, some players thought I was smurfing because I only had a few games but played like I knew a thing or two lol Most of my play experience is 1v1 hard-extreme AI. Mulitplayer AOE2 is just too stressful for me, when I play a game I want to relax, not get white knuckle sweaty, which is essentially AOE2 competitive lol
The emergence of streaming and collaboration has allowed this. When everyone played on MSN Zone from 2000-2008, there was no streaming. You had recorded games but it was view lock at best, there was no Capture Age, no analytics, no APM, what you saw in a standard rec is what you saw - the closest to watching real time was playing co-op and the 'honor' system that those co-oping wouldn't interfere. RUclips was in its infancy, expert players for the most part were still rather unknown aside from what records of them you could download. I achieved a 2100 rating on MSN Zone, but with all of that said above and the emergence of that today, people applying what they know NOW about the game would easily crush most of the competition.
I'm a fan of dirt simple solutions. If someone never played ranked before, the game could just ask them: "How experienced are you in AoE2?" Let there be three or four broad categories, like: 1. I have never played AoE2 before. 2. I know the basics and have a little experience. 3. I have a lot of knowledge and experience. Depending on the answer, put them into Elo 500, 750, or 1000. Add some "placement games" logic on top of that (rapid Elo adjustment) and people should land in the correct spot pretty quickly, ideally not having to lose more than 2-3 games in a row at the start. For example, if the first few games give or take 50 Elo, then a completely new player (starting at 500) losing 3 games in a row would already put them down to 350 Elo.
I will say. I started playing 2 years ago and immediately dropped to like 400 elo. One of my friends who plays kept me going to help me through the first games. I’m 1200 now and still grinding. It’s hard. But I’m very competitive so I’ve had a lot of fun with it
I don't have a lot of time to play the last few years... I feel like my knowledge of the game and my mechanics are becoming slowly better after 20y of playing but my elo just keeps on averaging or dropping from 1100 to 900-1000. You are absolutely right about quality of 1000 ELO player.. one thing that I do notice is the amount of all-in strats or one trick civs is very high. I'm all for the reset proposition!
The ELO system has two main things: → Good players have something they can show off. You try to climb the ladder to get a higher ELO-number. → You get to play against people of a similar skill as you. I'd argue that the second one is more important than the first one, and it's what we should focus on. Currently we got the problem that "average new players" (e.g. around ELO 500) take a long series of losses until they are paired with other similarly-skilled players. Giving new players a lower starting position will help for a while, but with enough new players joining at 500, that will become the new average, as Hera mentioned. I'm not sure how a complete reset is supposed to help. Just erasing all history will make it so _everyone_ gets very one-sided matches after every reset (or after joining again after a break). It might make it "exciting" if you just value the experience of climbing the ladder, but it won't help against having one-sided games for new players. (It does help against the "camping at a high elo", but I'm not sure how much of a problem that is.)
I'm late to the discussion as always, but I still want to throw in my 2 cents: Resetting the ladder is likely not a good idea, I'd even expect it to have the adverse effect. We just saw this in RBW, starting everyone at 1k means that new players will very likely just face players far above them. It won't just be a loss anymore, it will be a demolition that is far harder to stomach. They'll feel like they can't even compete. Some might not even know who this Hera guy is, lose to him, and then decide to never play again. That can still happen today, but it's a nigh impossibility. And doing challenges as the higher rated player or such is not really good either, if they realize you're trolling on them, they'll be even more likely to avoid the ladder. And not everyone is as masochistic as me, and don't get me wrong, if I didn't value Hera's time, I would dm him right now to "1v1 me bro". But it would likely turn away even more potential ranked players. Starting them off at 500 or so wouldn't work either. As mentioned in the video, it would only lower the ELO of the average player, but not their rank. ELO is meant as a long term ranking system, where the majority will, after hundreds of games, end up at 50% win rate. A lower starting point would have the same effect as the inflated TG ELO before, just in the other direction. What I would suggest is changing the algorithm for the first 10 placement games. During these games, the range of ELOs that is included is a lot wider. In a regular match, a 1k player will almost always play someone 900-1100, but in placement games, it could easily become 600-1400. How about we make it so that during those placement games, there's a higher chance to match with someone lower than your not yet confirmed rating? Like, a 80% chance you match with a lower rated player, 20% higher rated. Even if you lose your first few games, with the wide ELO range, it makes it so you will almost certainly play with players below 1k, and that barrier will go down with you. And if you happen to win, well, then there's really no change, you get the ELO, get up the Rankings, and match with those around your rank - just more likely to be a bit below you. I do always feel bad when I realize I met a newbie, and sometimes I even got them on their first game. It's hard to make it a fair fight for them, and motivate them to go on. This game is about slow and steady improvement at every level, and I agree that those first games are a harsh teacher. But I also think 20 losses straight is an outlier. Maybe if the game recommended "beating the hard AI" before going on ranked would also be an option? That was when I joined the ranked ladder, and I ended up almost exactly 1k after my placement games. So yeah, those are my thoughts. Now 1v1 me bro 11
I played AOE2 when it came out at the age of 10. I just got it back after watching you, t90, and Viper. I had the basics down and figures i would ROCK as soon as i started. I lost my first 12 games. lol That being said I am 850 ish elo, 41 wins and 34 losses.... I have had real good dark/feudal ages, maybe 40 seconds idle TC time into castle, trying to optimize Eco micros, following build orders, luring boars, pushing dear. Most of the time my opponent is doing the same. I wanted to quit those first 12 games, but I didn't. That initial elo placement is ROUGH and if i had to do that every so often I don't know if i would keep playing the game. just my 2 cents
I think New Player Lobby like in Voobly would be a even better idea. (In this case a separate ladder for new players, maybe you need to reach a certain rank or maybe 100 games to get to regular ladder)
i actually think the starting elo should go up rather than down. if the average elo is 1000 and that level is so far above the beginner level, i feel like just a range of 1000 is not enough to accomodate for all the skill levels inbetween 0 and 1000. no that does not fix the issue of having to lose 20 games first before reaching your normal elo, but Dawnrim's suggestion here in the comments would help a lot with that. as for me, ive been fighting this battle for years. i used to be 1600 hitting 1700 a few times, but i easily burn out on any game i play actively and take months off, and every time i come back i drop 200 elo and have to learn tens of new things in order to get back to that 1600 level, be it new build orders that up a villager faster every year or simply just rules of thumb. its not just being rusty like reddit likes to claim. nowadays, im just getting too old and gave up actively trying to improve, and notice more and more people gaining advantage over me. even just yesterday, i was losing a feudal battle on teamgame flank against a 1300 (1v1 elo). luckily i'll keep enjoying this game no matter how far my elo drops.
having this problem rn, me and a bunch of friends who all played aoe2 as kids started playing team games against AI last year and eventually that got kinda boring. so i suggested we switch to ranked, and we just keep getting absolutely stomped over and over again, to the point where 2 of my friends don't really wanna play anymore at all. it's strange because i know there's other guys like us out there who probably feel the same way, but by the time we'd reach each other's elo level where we could comfortably play against each other a lot of us will have dropped out due to the frustration. the worst part is that the whole reason we went for ranked instead of just joining lobby games was that we didn't wanna get stomped by high level players, we just assumed that the matchmaking would figure out we're not competitive players but oh boy were we wrong.
it will eventually, but the problem is it takes some games, also problem is if the elo of your team is wildly different, also team game elo is sometimes a bit misleading
I have observed this myself. I dont really play AoE, but some 3 years ago I played for a little bit and also did some ranked games and won against 1200 ELO players with reasonable ease. I dropped back into some online play earlier this year and was beaten handsomly at the same rank. I wanna say Im at pretty much the same skill level as 3 years ago. Still easily beat the hard AI, which is really the only thing I play. So yeah the average skill definitely increased a lot.
An easy fix would be to start the ELOs lower. What you're saying is bang on. When me and my friends were playing ranked games years ago, we were steady at 1100-1200, now we're all sat at 900 when coming back to the game lol. If the game wasn't competitive, it wouldn't thrive. So starting people at maybe 700 ELO to find their level is maybe a good move
After much consideration I started playing ranked about a month ago, i knew I was gonna lose quite a few at the beginning but slowly started to settle around 800-850 elo. Suddenly in one week I just went down to 700 elo. A streak of 7 losses in a row and some of them just really bad losses. Don’t really know if I really hadn’t gone to my real Elo and now I am but those losses do make you consider if ranked is still worth the frustration that sometimes comes with it
I started 0 20 or something like that too. I remember saying to myself ''you just need to tank it until you get better bro'' I'm actually 100 loses 110 wins or something like that, around 1030 elo. At least I can put some fight now!
Something else to keep in mind, we have an "unranked" matchmaking system that uses a hidden ELO. It's basically more streamlined version of the ranked matchmaking we already have but simply because it doesn't have a visible ELO a lot of more causal players will just go there. I discovered the difference in skill/tryhard when I was experimenting with strategies and just wanted to mess around and I was surprised by how big of a difference there was. 1000 ELO in "unranked" is probably still newbie kind of players or at least it feels that way.
It's so much simpler than that, the first X number of matches have boosted elo influence. instead of 20 matches to find where people land, it should take 5-10. Additionally having UI that explains to people briefly about the elo system and that "we're starting you at the middle point, and it should take about 5 games to settle you in." so players know that there's an end in sight and something to work towards.
I've definitely been there. usually a teamgame player of 3v3 and 4v4s, have dabbled a bit into 1v1s. about 1/3rd of my games feel fair, or I have the feeling I did something obviously wrong that I can improve on. The other 2/3rd of games I'm getting completely sweeped by a player that feels way better than those other opponents, like its an entirely different beast. Ended up quitting ladder alltogether, it just wasn't fun with how big the skillgap was.
Recently you said that the average 1000ELO player today would beat Viper 5 years ago. It seems this is true to an extent, and it makes me sad. My micro is decent and my build orders need work but my dark age is at least reasonable - but I can't compete at 1000ELO. I've been playing AoE2 since its original launch, I love multiplayer AoE and I've been trying to learn from watching pros since 2012 and have gotten a lot better, but my decision making is terrible because I have no ability to predict what my opponent will do. That's my weakness in literally everything. And I'm so competitive that I take losses really hard. I can only do like one ranked match at a time before I have to go mope around for a while at how terrible I am for something I've put thousands of hours into, and I don't want to know my actual ELO because it's going to take too many upsetting losses to get there. Anyway. The barrier to entry into ranked is very difficult for AoE2. I've played far more ranked AoE4 (even though the poor balance keeps me from taking that game seriously); it's easier on the psyche to do 10 placement matches and get categorized than just to start at a standard rank which is apparently too good, and make you lose indefinitely until you reach your true rank.
Ive always been a TGs scrub and I noticed this when they reset the TGs Elo. I used to be ~2200 before and I faced scrubs most of the time. After the reset my peak has been 1500 and 90% of the times these 1400-1500 peps are also 1500+ in 1v1. Idk if I need to play some 1v1s at this point. I don't really feel overwhelmed by my opponents but I usually don't do any dmg to them in feudal and the lot. I also feel like I dominate post-imp late game but it's all throughout castle when I lose the most.
I also think some sort of Beginner Sandbox would be good. Like a test you have to do before you start. (Match vs. AI where different metrics are measured like apm, timings, ressource-balance etc. together with the difficulty you can beat the AI. If you beat an extreme AI, you're probably no Low Elo Legend.)
This is why when anyone comes to the Steam forum asking about ranked, about five people say. Go watch Hera, TheViper, etc pro player, watch Spirit of the Law number crunching and theory, and T90/Memb for casted games. Go practice the Art of War and try to beat the extreme AI or at least the hard one. Because doing anything else than that is going to result in some crushing defeats and you'll have no idea why. Also I think some people need watch Spirit of the Law's video on ELO and how it works, because seeing some comments that don't understand why lowering the starting point does not work.
In sc2 u have placement matches and seasons. If you havent played before you are going to start of playing vs the avg mmr, and for every loss during the placemnet matches the mmr loss/gain will be amplifyied. So if you lose all of you matches, every match would be vs a much weaker opponent.
Great video! I have a solution: when I start the game after not playing for a while, I quit games voluntarily to lose Elo on purpose. I do that until I am sure i won't lose many games in a row. Hence I would advise beginners to quit 10 to 20 games in a row on purpose to have this Elo climbing satisfaction when they decide to start trying for real.
Everyone should start in 500 ELO. I felt like that at the beginning and that's the reason I only play team games now. Also agree with the seasonal ranked, but with long periods of time and not total restart. For example, 1000 ELO could restart in 500, 1500 restart in 1K and 2K in 1500. Thanks for this video. I hope Microsoft hears you.
Placement matches for new players would be the best solution I think. New players would play 5 games which would set their starting ELO. This would start new players at an appropriate ELO. On a side note having ranks displayed for ELO - bronze, iron, silver, gold, would be a fun way to incentivize more ranked play.
I have over 1000 hours combined age of empires to HD edition and definitive edition. I consider myself a new player when it comes to ranked. I consider myself below when it comes to rank. The reason why is because I had to practice so much learn build orders to be able to maintain and do something in a ranked game. When I play a one V one it’s very intense and I’m sweating, so hard to get a minor advantage over my opponent and one or two mistakes that I make will cost me the game hence why I play team ranked instead solo ranked
I am one of the people that went like 3-20 in ranked before finding my rank. AOE2 was my first RTS and even knowing a lot of what to do isn't enough to save my 6 APM ass. It's pretty hard to motivate oneself to play when you know you have a fair few losses ahead of you before you start finding people around your skill level. I think the best solution would just be making point losses steeper for new accounts with some system to detect if they're getting stomped, so instead of taking 10 games to drop to 700 elo it takes like 3-5. Cause I don't think this issue is wanting to be at a certain elo number, but instead wanting to find matches which are relatively even.
Hello Hera, I am from the Sc2 Community and played Age 2 only as a kid. After playing the sc2 ladder I learned a lot of things for RTS in general and tbh I think Sc2 are in a lot of ways harder (more micro and faster pacing) and after playing Age 2 ladder the first time I won because I had a faster play style and looked up some conters like pike -> horse -> skirms etc. And only this was enough to win. Now I think many player go to Age 2, because there is RTS left with a big community and if you were good in your rts you probably can learn the ways of age 2 very fast what leads to good players joining the ladder and destroing the match making. Like I was Diamond in Sc2 and than came to the 1000 ELO Age2. Sry for my english
Starcraft is a joke of an RTS. Its all tactics, zero actual strategy. Also SC2 is way easier than AOE2. And Hera disagrees about pikes countering knights ;)
those players today are insane. ive got achievement from HD edition defeat 1v7 hardest AI, ive got half of campaigns on hard difficulty and im stuck at 1000 elo
I've never played a campaign or AI highest elo 1600 1v1 and TG, been playing 2.5 years My point is only way to learn is playing higher elos than yourself then studying the replays, or just watching replays of higher elo players than yourself , For me I'd always watch about 200 elo above where I was at, it's more relatable. Now I'm finally at 1600 watching the Pros is more slightly relatable, but I've realised I'll never be higher than 1700, never ever ever, not getting younger, not getting faster, don't kid yourself, eapm is legit. Someone's who's max eapm is 60 will never beat a 200 eapm player with tbe and same knowledge. The top % of players have something we don't, they are special and no amount of training can make up for God given gifts....
The problem with Ladder 1v1 is that at ranges 1000-1200 which is the starting area, are many smurfs. I would say around 30-40% at least. People with higher ranks that dont like tgs and wanna chill or build a new account. I consistently beat in tgs 1400-1500 elo players in 1v1 fights (zero pocket help) with ease and 2v1 sometimes, and lost most game in 1v1 with 1200-1300 players. Then you check their total games and its like 50 or 200. 200 games you dont even know the civs well yet. The second problem is that when a player advances from 1200 > 1300/1400 he meets really good players, will lose some games, drop back to 1200 as expected, then meet the smurfs and drop more. What I would do to minimise this is take the total games of players into account, adding BOTH 1v1 and TGs. That way you wont be matched at least with a good player with new account which will cut down the smurfed games a lot. Voobly had a similar mech where people with less than a number of games, never match against people with more games, regaldless of ELO.
If you do a reset, Hera could be matched against a player who was 200 elo before the reset. It would lead to more mismatched games until elos drifted back to where they started.
If you reset the ladder every year then for a few months of the year there will be a lot of wildly one-sided matches. It also can't be much fun as a 1700 elo player to take a few months off and then suddenly have to spend a lot of time crushing people to get back up to 1700. And the other way around as well, people that found their 400 elo suddenly need to lose a lot again.
I agree, I used to play overwatch, and grinding back to where I was wasn't fun but a chore, and I didn't feel too good about crushing weaker players.
Rollercoasterman!
It still feels like a chore if you try new civ and then if you switch back to frank maining you need to grind 400 elo back
@@biegaliusz4439 we have no sympathy with the franks picker, despite what Hera keeps saying 11
Rollercoasterman plays aoe2! Also a 1999 game. A collab with hera would be great, teach him to coast!
Incidentally, I just re-watched SoTL's take on this question just a few days ago, so I have his conclusions fresh in mind - as well as his in depth explanation of how the elo system works. He suggested more placement data be used before dropping new players into the ladder, such as placement matches vs AI, data from other ladders the player might have played in (team games primarily) etc.
Personally though, I think the easiest solution would be to have an higher, but steadily dropping K value for the placement matches. Presently the 10 placement matches a new player plays through all have a K value of 100, which means if you match up against a 1000 player in your first match and lose, you'll lose 50 elo points. That will only get you to 950 obviously, so if your true elo is 500, you'll have to lose a lot of games. If the first placement match had k value 600, and then was reduced by 50 for each of the following placement matches or something like that, getting to true elo would be much quicker. Sure you might bounce to 1300 right away if you're lucky or play out of your mind somehow, but that will quickly get corrected.
I reckon its a matter having the ai compete in ranked. And putting your starting elo matching that of the ai corresponding to your best offline performance vs ai 1v1 with ranked settings... I dont see any drawbacks here!
Sounds good, except for one problem: The high K value affects the opponent too. Imagine being a 1100 player that's been playing for a good while matched with a 1000 beginner and winning 300 points. That would make you jump to 1400, way above your level, then you're forced to deal with losing over and over again to actual 1400 players because of the low K value that's applied in those situations. It would take a huge loss streak to get you back where you belong, so you have the same problem you saved the 1000 beginner from. It feels like every solution has a major drawback like this...
@@danieltoth9742 Just don't apply the K value to the established player, not even sure this is the case today. But I definitively agree that the real solution is getting to your real ELO fast within the first 5 matches, when you loose 5 and you're at your ELO then you will have more fun. Any playing around with the ELO numbers doesn't make too much sense, like a reset Hera suggested, the ranked ladder would be unplayable for some weeks due to all the mismatching ELO level matches, except if the reset means that everybody has to replay their placement matches, but even like that, not sure it would be playable for some time.
@@mramosotero Are you suggesting the established player should gain fewer points than the beginner lost? If so, that would take points out of the system, which gradually lowers the average rank -- meaning those overpowered 1000 players are going to be 900 players... then 800... then 700, and so on. They would follow the beginners in a race to the bottom, making everyone feel like they lose more than they win.
@@danieltoth9742 Yeah sure, the ELO is based on points exchange between players, but it's not worse than making beginner players start at 500 ELO for example. Back in the years 2000s the starting ELO in MSN Gaming Zone was 1600, but for me that doesn't change anything, so even if the system would lose some point over time il wouldn't be a great deal.
Nevertheless, one solution would be to keep a points credit for the beginner player that would be "consumed" over many games. For example, the new player loses 500 points in 5 games, but the opponents only gain a total of 100 from those games (20 each), so he keeps 400 points that will be credited to his next opponents over the next X matches. Then when he loses a match the opponent would get double the points (which doesn't make a huge difference at an established level) until the credit is depleted.
Grubby, the Warcraft 3 player, would argue against your proposition. And I would agree with him. Ladder resets just make the placement matches a mess until the ladder is settled again, having more unfair matches during that time.
it would really be nice to see some collaboration between the two of them, Hera's micro who have been amazing in Warcraft 3.
@@BelialTnTn Or even a discussion about this topic in particular. They both seem interested in it.
i´m not too sure if you are right about this one. there are hard resets and soft resets. hard resets would reset your mmr and ranking points etc. you basically start from the beginning. soft resets just reset your rank. the community made ladder in warcraft 3 "w3champions" does soft resets and it works great
There are people who persevere... people who quit after initial losses...And then there is our Hero "lucaitasa" having fun at 22 elo ( o7 ) playing there for years :_D
That guy must be trolling. Mean APM in one of his last matches was 0.44.
@@BillyBlaze6907 there are people with disabilities playing online games.
@@karnazacss 0.44 is one action every two minutes.
Played a guy at 1050 the other day who did a perfect 2 group scout rush into good castle time with Knights all whilst full house/palisade walling his base.
I do that type of stuff every game and I’m 960
Maybe it wasn't perfect but you were just way slower than him but yeah breaking out of 1000 elo can be tough these days.
I used to think so too, and then I played a game vs 1800 elo player (as a 1000 elo player). The 1800 elo player basically has 2-3 scouts in my base when I'm halfway up to feudal age and haven't even started building walls or making a barracks. The skill level is definitely much higher at greater elos, and the scout rush at lower elo is nowhere near perfect.
He was not playing perfectly
That is 100% a Hera viewer right there.
Minor point: i think ur underestimating how many ppl play custom games, set up team lobbies w friends or simply play unranked. Like the choice is not binary bw ranked and campaigns
But the skill difference is even bigger in custom games. Lobbies will be called “noobs only” and nobody knows what that means. For some peope everything below 900 is noob level, for some it’s 1000, for some it’s 1200.
@@BillyBlaze6907 I remember on voobly “noobs only” lobies would have people with a few thousand matches played
"1300 already start doing micro and build orders."
I havent played the game in months, but to my knowledge even at 1000 people try to have perfect build orders and very good level of micro, hit and run and saving each unit in the whole army etc.
Agreed, I think Hera needs to watch more games at that elo, they're way better than hes giving em credit for lol
I agree. But it's also very inconsistent. If I get spanked it tends to be by a RUclips build order. And if they play naturally I often do the spanking. Generally speaking though most people at 1k can do a basic build order for every opening unit. @@M0rdFustang
As someone who hates micro and likes to just sit back and have a nice good old fashioned macro game. I notice I have to deal with archer micro bullshit (stutter step) down at 950 elo, even had monks once. Used to be that people didnt do that shit until about 1100 elo.
I think a 1000 elo can execute a perfect buildorder without being able to analyze if that is the right strategy to apply. In addition he/she would do it without being able to analyze why he/she won or lost. So, it’s like they are playing a different game. You could have a perfect buildorder but it could not make sense because of your civ , the opponent civ, the map, the behavior of your opponent, the strategy of your opponent , the unit each one is opening, if there is water or not in the map, etc etc.
@@franok9733 at 1k elo' though the build order they follow is normally civ specific. If they go random I find they know roughly what each civ should open with and do the basic build. It's genuinely a lottery in terms of what you might face 900-1300 elo'.
You’re mistaken, Hera. The placement matches would be hell for half of the community.
They just need to do the same system for new players that they do in chess, where new players games count for a lot more elo points up and down so that you find your elo more quickly. Or something along those lines however it works. If you open a new chess account you will be roughly around your elo after 5-10 games. Nobody loses their first 20 games in a row that way.
That is a placement system
@@tomyrobinson3196 I don’t know the terms and such, but I think that would fix it
I think the important thing is that it counts more for the new player, but less for their opponent. I think in the Elo version used in AoE the win/loss is the same on both, so everyone playing against several new players would immediately be catapulted up, and then have the same problem of being in the wrong ladder region. (And people playing against new smurfs would have the opposite problem.)
There needs to be a ladder and non ladder play option. That would solve this.
1000 Elo is indeed a bit brutal right now. I just picked up AoE2 seriously a few months ago, but i played campaigns as a kid and played Starcraft 2 ladder for quite a while all the way back in the Wings of Liberty era, so I've been around RTS for quite a while and am familiar with the importance of build orders, worker production, expansions, etc. I still ate shit against 1000 elo players for several matches. I only got down to like 800 elo before bottoming out, i can't imagine how rough it must be for someone new to RTS who loses for 20+ straight.
I don't like the reset, that just creates a couple weeks of volatility while everyone wins/loses their way back to their true elo and then we're back to status quo. I think a better solution is increasing the ego changes you get early on (ie losing your first match knocks you down 2-300 points) so that you more quickly find your level without an epic losing steak. Alternatively, you could let players sort themselves a bit with questions like "what is your experience level with RTS games" and then start them at 600, 800, 1000, or 1200. That might cause some smurfing, but with aggressive early elo settings that would only work for a few games per account.
yeah something similar to starcraft 2 would be great. Starcraft 2 ranking system is neat. I also played it for years and years.
bottomed out at 230 or so. Finally sort of on the way back at 370 or so now. Would hate a reset of the ladder......
- Start players on a low score
- Create a lot of score out of air if they win placement games
- Keep track of how much score is in the game as a result of their account
- While there is more score from them than there should be, destroy extra score per loss (which the opponent doesn't get). While there is less score from them than there should be, create extra score per win (which the opponent doesn't lose)
That keeps the total Elo score at 1000 per player while allowing players to start at 500 Elo if they're new
underrated comment
This is a good idea. I would just add a possible improvement.
Take for example a player that starts at 700 ELO so 300 ELO is destroyed into thin air (which we need to refund to the ladder somehow). During the placements, we would be giving this player more ELO than their opponents lost, so that the ELO can be given back to the system, and taking less from them than their opponents get. This means that the player would have to lose more than win, in order to stay at 700 (assuming that is indeed their true ELO and we placed them correctly). This would still feel to the player that they need to win more than lose in order to get to their ELO. Of course, this is much better than losing long streaks, but maybe we can do better. Maybe we can limit the amount of ELO generation or destruction during this phase to some small level, so that the win rates of this player are not far from 50%. All the ELO that does not get resolved in due time is simply sprinkled to over all the other players, which means even if this player stopped playing after initial match, and got 700 ELO, the remaining 300 ELO would stay in the ladder, and the average would still be 1000.
You don't even need to track per player. You can just track a total and compensate with random +1/-1s in any match at all
In Aoe2HD we had ALOT of smurfs who would love this.
@@crownofglory1264 why?
I swear 1500 turned into 1700 overnight. I took a break for a month and now my opponents all play perfectly and I can't win at my rank anymore
You also aren’t playing at a 1500 elo level likely
Same. I’m no 1500 Elo but, I was constantly at 1200 and I took two months off and I have played for two weeks and I am back to 1k. I don’t think I have gotten that bad but, everyone is getting better.
The elo gets deflated as the average player skill increase, a 1500 from 2 months ago is not the same as a 1500 from today. You're probably no longer a 1500 elo anymore, probably a 1100-1200. I always check aoe2insights when I start a custom game with a specific skill range precisely because of that. Some people are just happy to reach certain ranking and stop playing but that ranking means nothing if it's not constantly updated.
i had the same experience, i was playing aoe2 like 3 years ago and i climbed to 1446 elo back then, when i came back i did some little training with AI and study to just get familiar and i started to play ranked games, first game i won and then the lose streak began, i was losing way more than i was winning and i dropped back to 1220 elo. After a week of play i climbed back and 2 weeks later i almost hit 1600 elo for the first time. Which is nice but i have to play very well to bet my average opponent of 1550 elo. The weird thing is that when im really on it i win so easily even with 1500+ elo but then i get to 1600 and i start losing again. Mostly because my build orders are wack.
@@miguelbalderas905 Dropping 200 Elo can happen even if you are constantly playing, thats most of the time just temporarely if you have some days you just play worse than normal.
Here's another take: The issue is not necessarily losing several games in a row, but that there is a lack of communication between new players and experienced players.
I am at 1000-1100 elo with about 1000 games and so an average player. I'd like to know if the person I'm playing against is going through placement matches at the START of the game rather than at the end when I see how many games they played. That way I could go easy on them and rather chat with them, offer some encouraging words and ask how things are going.
I think that would be a way of helping new players feel welcome despite losing a lot in the beginning.
I really like these videos with the continuous gameplay in the background and the chill voice other. Keep up the good work and let's hope to convert some more of these campaign players someday!
Best solution:
1. require art of war completed before multiplayer
2. Have a selector that asks, "What would you classify yourself as?: Complete beginner, beginner, moderate, experienced"
3. That selection puts you at 500 elo / 800 elo / 1000 elo / 1200 elo respectively
Smurfs would have a field day with this one. And new players could genuinely misjudge their skill level, like imagine if most new players read about others having a bad experience they are all going to choose the lowest option and it would get overcrowded quickly.
@@theowlogram888 are the smurfs in the room with us right now?
I like this. If you self score experienced and lose its your own fault but you're also the person who doesn't care and will keep grinding.
I think just leave it up to people to chose where they enter. Why we forcing people to come into the game at national level, its ok to want to play community level and naturally climb.
Also why don't we take the medals from Art of War. Like if you get no medals (i cant remember how medals work lmao) you can only place in beginner.
Smurfs delight
@kirtanamrita2302 I don't really get these comments, I don't think I've ever come across this problem, maybe I'm lucky but I don't think it's an argument for not making changes... like can't you just do it now anyway? So they'll still do their thing, but this just makes is easier for everyone else. I don't see a meaningful downside
I rarely play 1on1, mostly teamgames where mit ELO fluctuates between 1050 and 1200. But a few weeks ago I played a noob 1vs1 and I didn‘t know, gave it my best and picked the poor guy apart…I felt really sorry after seeing it was his first game.
The gap between a beginner and 1000 ELO is huge!
I am a good chess player, not titled but honestly I am at least a national master strength on a good day. I am ranked 99.8%. Chess has two factors in the elo system ranking when you start playing. One of them is a provisional factor. If you start playing as a brand new player and demolish a 1200 player, demolish a 1700, demolish a 2200, you might be playing a 2400 player very shortly, whoever put together the elo ratings and how it evolves did an atrocious job. It would never fly in chess. If you start losing 5 guys in a rows at 1100, 900,700, 500, 300, you might be playing like a 200 elo player after just a few losses and those last games even if you lose would not be stomps. The algorithm for rankings is just really messed up.
A guy getting stomped 20 games in a row is just a sign that the system is broken. Honestly, that’s something that kept me from playing for a while. I just didn’t have the mental fortitude to deal with the rust that I’ve built up in a stressful aoe2 environment.
had this issue introducing my friends to the game, i think the ladder should start at 700 elo honestly
as noted, this is an unstable solution. the whole system will adapt until 700 is the new average.
@@neildutoit5177 Only, if the game suddenly gets flooded with thousands of new players. It doesn't. Unfortunately...
@@neildutoit5177 But what about reintroducing that elo to the system during the early games of the player? Using 700 as an example, lets say that for the first 10 games, they win 40 elo if they win, but go down only 10 if they lose.
I dont lnow if those are the right numbers, you would need to find the starting point where, on average, a new players wins half of their placement matches, and then adjust the the elo reward to match that so that, on average, each new player is still introducing 1000 elo to the system.
It would help overqualified new players to climb faster, but it would help new players get some wins early by matching them with weaker oponents. Or is there something im missing?
@@neildutoit5177 I don't get it, can you explain to me why that would happen?
The solution for this is very simple: up the k-value for the first 10 matches massively, and up the k-value for a players' next 20-30 matches a decent amount. The k-value determines how "swingy" your elo is and how much rating you win/lose from a loss.
This way a noob goes from 1000 starting -> 850 -> 700 -> 600 -> 500 and stabilizes there as opposed to going 0-20 to get down to his "true" elo of 500.
Similarly experienced players with new accounts would shoot up to 1500+ quickly and not noob stomp as much.
I've played the game on and off for years. Not competitively, but I knew the basic build orders, unit counters and what all the upgrades do, etc.
When I jumped into Ranked games in DE a couple months ago, I got smacked down from 1000 to 700 pretty instantly before I managed to level off.
I have same kind of background. I got interested in ranked play but I did not want to lose all placement games so I trained against AI, getting build orders right and practicing my multitasking. My first ten games were 5-5 and I'm happy with that.
All that time spent playing campaigns and against AI on and off for years definitely helps but ranked is slightly different beast to tame.
Hera: "1300 know the basics pretty well and begin to execute build orders"
Me at 800-900 watching my opponents getting to castle age at 14-15 minutes and wrecking me shortly after: 💀
Fantastic idea! Appreciate all good you do for the game, important stuff. We need players more than anything.
Watching this video has given me a lot of confidence and I do have a suggestion at the end of this comment as a potential solution. I recently (June 2024) got back into the game after 20+ years away. In those 20 years, I've become a history teacher, chess coach, and eSports coach and I've used my knowledge of chess and attempted to apply it to AoE2 through casting T90 Community Cup and AoE2 Specialist Cup games as well as participating in said Specialist Cup. When I took my talents to the ladder, I won my first game in my first game but then proceeded to lose the next 9! Your advice on having fun first is great because as a coach, that is what I stress to my students, but I often forget that for myself. This is truly a game similar to chess where you have to lose a lot of games before you win a few, so thank you for your encouraging words. One teacher to another: being a calm encouraging presence is key and you've got that!
MY IDEA: A season reset is a decent idea, but instead I'd like to see a ranking system based on chess, which uses ELO as their ranking system. In that system you'd have Classes of players akin to chess so Class F would be 0-1000 ELO, Class E would be 1001-1199, and in increments of 200 until you run out of letters and get into Master categories at 2000 ELO. Then you'd have like say Expert from 2000-2199, Master from 2200-2399, and Grandmaster from 2400 onward with Super Grandmaster around 2600. Each level has a FLOOR which can be achieved and a player cannot go below that once they achieve and retain a certain class for a certain period of time. This would also have the benefit of preventing sandbagging. No system is perfect, but this would be an interesting way to go about things I think. Again, thank you for your time and your talents!
as someone who studies lots of statistics
The ELO system is basically a mathematical model in which, under the premise that higher elo means higher % of winning a match vs people with lower elo.
This difference can be graphed and modeled and, as usual, this difference follows a normal distribution and you can assest a proportional reward/punishment for wining or loosing. Which is why, for example, an elo 1200 winning vs a elo 1500, the player with 1200 elo will win a loooot of points and the other player likewise will loose a lot of points as well, this is because the statistical probability that a 1200 beats an 1500 is pretty small, on the same side, if the 1500 elo wins, he will win very very few points.
But this concept must embrace also the concept of iterations, you can never see an algoritm and understand properly this concepts without the concept of repeating the same "game" of probabilities time after time
And, after various iterations, you will loose some, win some, the trend will dictate your ELO.
In the perfect system, that only mathematics can imagine, infinte numbers of iterations will leave you your "true and objective ELO"
But that is of course imposible, materially, technically, economically
with that in mind, we know that it is imposible to create the perfect algoritm with perfect representation of each individual skill on any given time, only trough time and iterations we can, to some extent, "measure it"
Go statistics!! I think they're using a logistic distribution instead of normal, which is actually normal😅. There's a lot of things to consider for a rating system, inflation, min rating, commutativity, the famous K factor, sigma values etc.
However if you know exactly what you don't want to happen with your rating system, it's not that hard to engineer one.
As you said Elo can't messure real "strength" and no 1 dimensional rating system can (maybe infinite dimensional), but it's good enough for most cases, in fact it's better than any tournament outcome ranking (Unless it's a round robin and the points you get from each match are your rating system usual points, but that makes a 10 player tournament 45 or 90 matches long)
@@EneldoSancocho yeah indeed it is often use the LOG function to soften the bases, but like you said it is not that hard to implement some quirks or restricción in an ELO algoritm
@@sebastianpalominos3706 Man I'm honestly happy every time I encounter someone who studies math. Probability and statistics are beautiful subjects and they are full of counter intuitive problems (like 2 envelope paradox and such), they also are related to so many applications in different fields.
If you are interested: there are alternative formalisms for prob theory for which you kind define a uniform over the integers 😃
What you say resonate a lot with me too. I played a little when i was a kid. Last year i started playing a lot with 3 friends who where noob too and I demolished them. I played basic campagnes, played even a lot vs difficult IA.. But online vs real players is different. I got crushed my firsts 7 rankeds and imediatelly stopped playing online. I needed a full month training, making around like 100 games, restarting every time i was feudal too late, watching tutos, to have my first win. And even then, it was close.. So i didnt continue for the moment. I feel like i need more training. I think even 800 elo is too high for new players.
I dont think there one global solution, but maybe one big step can be to propose to new accounts to play 5-10 games vs adaptives bot. So if you are a noob like me, after a few games, you can start a bit more at your level.
The big dream could be to have real short recaps in the end, on your weakness during the game. So beginner can know easily what to work on first (Late ages, how much time exactly when your Forum was not working, bad balance of ressources in early..)
I'm a long time player (started following the scene before the first Hidden Cup to give you context). I stopped playing for a long time and only recently got back into it. The whiplash and skill level differences between then and now is insane. I think a large part is because of great creators like Hera making high-level insight available to all. But it creates a weird dynamic where people like me (that appreciate competitive play but like to mess around while playing) keep bouncing between stomping the competition and getting stomped.
Reworking the ladder system to have a more nuanced view of skill would be a good start to address this. Or pushing the unranked scene more (it's so much fun and I barely knew about it till I came back to AOE).
Great topic, thanks for bring attention to it.
I would love to see ranked seasons with some minor reward at the end showing your season/rank or highest rank achieved badge, etc. The details for the ranked system would need to be dialed in first. I hope the devs will gather ideas from the community, then build a new ranked system. For example- your last seasons rank can be taken into account when you play your new season placement matches to avoid a noob stomp.
Since were on the topic, lets get some updated menus as well. The main menu is.. fine, but the lobby menu and ranked menu feel a little "windows 98". Are the devs limited by the software, or just trying to keep the nostalgic feel?
I'm ranked about 1000 elo and I rarely play against beginners. I think there are not many of them.
you just drop instant arround 100 elo after loosing the first game
As someone who just started playing age of empires 2, I have zero desire to get into the multiplayer. I know I will get absolutely demolished if I try it. Single player provides enough entertainment for me and I can challenge myself by increasing the difficulty without the stress of facing a veteran player
Me too. But sometimes I find a begginner in TG and is hard to see people bullyng them when is obviously a new player.
Even though beginners start at 1000 elo, they are usually matched against people at 800-900 elo for their first matches.
I played 4 times vs incredible strong players (compared to me), the last one harassed me via chat, and so I went away from multiplayer. Playing vs AI is fun too and watching pro gamers shows the skill ceiling.
well said. been playing ranked 1v1s as long as Hera has known of me and longer even. over the years i've managed to enter a stable 900-940 elo, but that has come with so so many ups and downs a *metric shit ton* of watching everyone around me get better very quickly. these days 900 means something; 1k3 *really* means something
again, well said. good video addressing a point that often goes undiscussed
about 1400 games played btw. ~50% win rate
Showing footage from viper's stream when talking about beginners 🤣
I did the same thing starting out. Crushed at 1000, drop to 700, now im at 1100.
Just keep going
Me, struggling vs moderate difficulty AI in the campaigns: Aha... yea
11
@@markusdegenhardt867811
Me too! I can usually smash a Hard AI though. Campaign difficulty is labeled lower than it actually is it seems
There is only one solution... lose more elo per game lost in a row. A multiplier of some kind, losing 10 games in a row would have you at 500 elo as opposed to 800 or whatever it is now
Or just do that for new accounts
The implication would be that you will gain more elo if you beat someone who has lost multiple games in a row. Which is obviously not good.
@@neildutoit5177how about: You lose 5 games in a row, then you get matched against a considerably lower Elo player?
@@neildutoit5177when someone loses x elo after losing, the winner does not gain x elo. The lower the rating of the person who lost, the less elo gain the winner gets
No sir, not per lost games in a row. They need a provisional factor like they do in chess. When you are starting where your rating can really bounce around while they try to find your true rating. It should not take 20 games. They should be pretty close after like 5 games.
I basically spent 3-4 weeks researching the game and practicing build orders before I even went into a ranked game. I think if I went in completely blind with only my knowledge of playing as a child I would have lost 20 games as well.
Hello Hera, I am new in game (1 month) and I am 920elo atm. I started to watch your videos about "How to win in low elo" and etc, and I was just keep getting rekt :D. And believe me, I was following every instruction! I really feel that ~900 elo players should be worse than that, or this is how it used to be in Moba games that I used to play. Then I start thinking that this elo, while it is actually low, it might not considered that low in AoE2. xD
*Edit: You really helped me to reach this elo in a month, by just watching your streams, not the videos about low elo :D
i recently began playing aoe 2 again after 15 years and first played unranked and after 20 matches started rank with around 900, I resigned multiple games so my Elo goes down intentionally so I play against my own caliber, right now im 600 and have 9W and 18L overall in ranked and I am enjoying the matches :) ty Hera for the videos. I would love to see more in depth videos about Civs and how to play them.
So you're telling me if I went back to ranked after not playing ranked in a couple years I would get wrecked? I'd have to check but I think I'm currently around 1000 elo?
All joking aside, I have been playing through campaigns lately because I got all the dlcs and I wanted to play them all, so I definitely don't feel like a new player (600+ hours on DE); I just haven't played against a real person since I got back into the game a month ago (don't worry I do intend on playing ranked soon and won't mind losing if that's what happens; I can always lean on my Huns if necessary!)
Also I'd say that from anecdotal data it seems like there are a lot of og players from back in the day (like my dad) who simply doesn't play online because he isn't interested in playing fast enough to keep up, so I'd imagine that players like him probably will stick to playing against the ai
1
Different solution, just do what League does, have Bronze 3 for true beginners and have people rank up if they win a bunch and rank down if they keep losing, that way you always get fair games but the starting rung of the ladder is true beginners so no one gets discouraged
I am currently at about 750 elo and most matches I encounter are people with +200 ranked matches, and I'm a newcomer, never played AoE (I did play some AoM but they play very differently), bought the game in the 25th anniversary which was like 1 month ago and I've been playing online 3 weeks now and Im watching content and striving to do better every time, every match, I'm seeing myself progress trough not wasting food, walling my base, choosing better my armies and my upgrades, and all due to the vids of the AoE 2 community that helped me. And yes, 1000 elo seems like its gonna take a bit of work XD
I'm glad that you are informed on what happening on the ladder at lower ranks than your own. I haven't played in a very long time but when I did today, I got regularly beaten by people less than 200 Elo below me.
I like the Elo Reset idea!
Man I feel this so much. I improved so much but the ladder grows ever so stronger.Its so insane how good everyone is at 1400+
I pushed through the loss streak 1 year ago with the mentality that every loss is a lesson, I checked replay with superfast forward brief analyses, what my opponent did and how he overwhelmed me. Then I went next. Lost like 8 games until my first win. After I stabilize my elo, I started learning more and absolutely enjoying the 1v1 multiplayer (I rarely play 2v2 or 4v4 - too much rage quit dumb-dingos). Currently sitting on ~1200 playing casually few games a week (pref. Arabia). I find the most important think is build order and get the feel how to distribute your economy to being able to produce vils and your army spam choice (including how many military production building you should add once you develop eco more) and being able to do tech switch and adjust resource gathering ratios accordingly - this is what I find to be the core of all the game, also most demanding on multitasking. I don't focus on micro very much yet, but eventually I'll get there once I'll be more efficient with eco management -> more time to focus on the army and become a micronerd community member.
Great video as always Mr. Hera.
P.S.>> I encourage all of you, people who read this, don't give up with multiplayer, once you get better, it's a lot of fun and the community is super nice on discords so you can discuss and develop your own playstyle with favorite civs. Stop caring too much about your ELO, winning or loosing, be positive, call the GG every game, you can check with your opponent after game, maybe he'll give you some tips for the future (happened to me many times) and trust me, if you suffer the ranked anxiety, this will really eliminate most of it, after all, it's just a game and it's okey to loose and raise stronger next games. ☺
Yes best advice .. people stop care about elo. Care about having a good game . If your elo is higher just means more chance to lose. Your ELO will match your ability after time giving you a better experience winning 50% of your games means you are at your current skill level.
Agree with having more in-game communication systems. I find that I enjoy the post-game analysis with my friends and discussing what went well and what went wrong. I would like to do that after ranked games as well with my opponents at times. If the post game screen had an AI analysis of the game and tips to get better and ways to share/ communicate with opponent & friends, even your losses might end up teaching you something and you would feel more positive about the experience.
I'd always wanted to play ranked but didn't want to get destroyed. So I watched a bunch of build order tutorials and created my own spreadsheet showing what to do with every villager up to Castle Age (based on research I found online) and practiced this against the AI a few times to get the muscle memory for it (and watched the games back on CA to analyse what I did).
After that, I jumped into my first ever ranked game and made some silly mistakes but nothing too major with the build order. However, I lost because the opponent was simply more aggressive and spent their resources better, which I only learned through watching the game back on CA (also it was Arena and I stupidly let the enemy's troops into my base so I just GGd).
Then I went into my second game and played a lot better, making a note to actually spend my resources (especially building multiple military buildings for faster unit production). I got very lucky spotting a couple of the enemy's attempts at sneak attacks, and just managed to out-play them and eventually win. It was so satisfying getting the win, and it was only really possible because I'd done so much prep beforehand. If I didn't know anything about build orders then yeah, I'd have been crushed, because both my opponents had enough of a grasp of the early game that I'd have fallen behind significantly.
I agree it's absolutely wild to start at 1000 ELO. Like, I know if I got matched with a player who was playing their very first ranked game with zero prep or research, I'd likely destroy them. But equally, if I got matched with a 1000 ELO player with multiple games who knew what to do then yeah, it'd be a very difficult and different game. But equally, you only get better by playing opponents better than you - there's no point constantly playing against people below your level.
i had played the game in single player for almost 20 years before ever trying an online game and it was in a custom lobby with friends, we played quite a few matches over a couple months that way and i noticed that, while i had the game knowledge, i was really not used to playing against humans but over those months i suddenly got way better than i ever was at the game and eventually i tried playing ranked where i played like 6 matches, won 3 of those and then decided it was too stressful for me lol
point is, i think a lot of new ranked players had similar experiences, getting good at the game while playing with friends and then actually coming into ranked fairly competent, completely unfair towards truly new players
I play for 3 years now, watch every Hera video, check meta play on T90‘s Channel with Red Phosphorus but still can‘t exceed 1400 ELO. The Online Community is insanely strong!
People are obbssed about their score instead of understanding their score is their current ability and will give them the best matches .. do you want to be 1900 and lose every match ...
I took a break from the game about 6 months ago and came back to find kind of the reverse effect for me. I jumped from 1200ish elo to 1600 after a couple weeks. I would go against either extremely strong opponents or those that would feel more like beginners still. It might not happen as often but I feel like your thoughts on 1000 elo also apply in other areas too. :)
What if you actually start at 0 Elo and there is no reset until you reach 1k? Everybody above 1k gets the seasonal reset and maybe there are even leagues... like if you break 1300 you only get reseted to 1300 and if you reached 1600 you get reseted to that, never lower, so you don't have to grind your way back up to your current elo range?
just recently returned to the game after a very long absence (stoped playing when hunwars where a thing) because it seems I can enjoy playing vs a lot of civs again. the "art of war" challenges do help to get back into the game for sure but they don't go far enough. like getting into feudal/castle with x vills asap is nice and all but they don't teach anything what happens after that. so can my vills just keep sitting on wood/food for the most part or should I shift them because I want to add TC/castle/stables+units or whatever? if so, when do I need to shift them to get my desired result? that is for everyone to find out himself.
now making buildorders is not much of a problem for me since I used to do that back then as well but for realy new players, that may be to much. especialy considering that you can comfortable beat hard DE AI without much of a plan. I did so yesterday on arabia with only: wall your base + reach castle with a few archers for defence asap into boom.
as nice as it was, I doubt I would see much light when starting ranked.
I've never played AoE online and, thinking about it, the game came on disks way back when I first started playing and I didn't even have internet at home. Therefore, whenever I played AoE as a kid, it was offline and that's just how I've always remembered playing the game, so, even after getting it again on Steam, I continued playing like I always had without any interest in online play.
2:22 - Seems like your cat also wants to play some ranked AoE II : DE game.
I think i can relate to this. Was a new AoE player myself, literally brand new 2-3 months ago, but i sort of had an idea for the game from just enjoying watching high level play and having played the campaigns as a kid.
I learnt build orders and everything before going ranked and even then i still got my ass kicked most of the time. But i think the problem is that 1k elo is inconsistent. I would meet a guy that had no idea what he was doing, and next game i was playing terminator and got my ass kicked HARD.
after ~20 games i ended up clearing up my dark ageand naturally climbed back up to 1k elo. Took me arround 50 or 60 games to break past 1050 because it just was so wildly inconsistent. 50 games and a month later i'm sitting at ~1250 and i'm still having some games where i match with way higher skill players but overall its okay. Its just the 1000-1050 range is completly messed up.
Most games nowdays im still loosing due to experience gaps and knowing what units and strategies to go for rather than skill (which is where your discord is actually very nice, always learning), so i can only imagine how bad it can be for someone stuck at 1k elo with 1000+ games
I got the same experience of 1kelo being inconsistent. Wonder why that is.
Thanks for the video Hera! Im 1350+ elo since 2021 i just dont find the way to improve. Im spending less time playing now because work and life, but it wasnt a matter of grinding, i just coudnt got better.
I started 0-2. And quit. But I went to the quick play mode (no idea how that matchmaking works) and played like 100 games with good results (probably less than 50-50, but it felt like progress). That helped me get the confidence back to jump to the ranked matches again. I think clarifying that progression can make sense for some people.
Quick play felt stress free, and the vast majority o the community is helpful if you ask, so you actually learn from people that are better than you
I have some background in game theory, and math overall. One solution to try is just add + 200 rating points to everyone. The reason for this is that AOE2 doesn't have adequate incoming number of players every year so there is no "feeding". Normally new players are supposed to come in that give rating points to good players and their rating goes up. But due to shortage of new players, everyone just exchanges points among themselves, even though community is getting stronger, their sum of ELOs is constant. So add a community growth value every year to every player.This doesn't have to be constant for example +200 at 1000 rating, +50 at 2000 rating, +10 at 3000 rating. Once the incoming number of players are high enough to compensate for community skill growth this adjustment can be stopped. A stagnant game where new players are not coming but skill keep raising we have to keep doing this.
Exactly. There is hardly any new players coming into ranked and is exactly the problem. You could also give more for win than loss which would push better players higher than 1000 .. who cares if the highest ranked is 5000 or 20000 it doesn't matter as long as the score can reflect the skill difference.
I was waiting for this video... When I started about a year ago, I got steamrolled for about 20 games, before being able to adjust. I'm still sub 1000 and players at this level know how to wall, some even quick wall. They do follow build orders, and they are really fast in aging up. It's very competitive.
i totally agree with the pb of starting a rank account at 1000 elo. in few years it went from ppl playing casualy to ppl tryharding rank
We've all been there offline, practising for the ladder. It's some part of the improvement
I definitely agree with heras point in ELO. I was a 1000 Elo player back when DE was first released and I was doing ok with some win and some loss. I recently got back to the game after practicing against AI, using the same build order (fast castle boom), and got stomped for like 10 games in a row.
Been thinking that for a long time, glad to see a big name addressing this issue.
I think setting initial ELO to 500 solves it short/mid term. Now beginners will play against other beginners or "actual" 500s, while actual 1000s continue playing among themselves.
Long term perhaps more changes are needed. I think something like making the first 5 games worth triple the points, and the next 10 games worth double should do the trick.
This is very true, after playing SCBW for 20 years I thought i'd reach like 1300 ELO within a month or two... but the game is very complex, gotta balance between 4 different resources, build orders are quite complex, there are tons of things to know, what unit counter what unit in which numbers. so far most games are either weird rushes i have no clue to deal with and I lose (lmao sicilian dungeon rush is ridiculous), or we turtle till castle/imperial and I usually win because I got much better late game mechanics but my unit composition is usually awful and it's still a complex game to figure out. The 1000 ELO guy in AoE2 has some very very decent knowledge of how to play an RTS. I mean even at 900 ELO people know how to push deer properly in dark age, know interesting rush build orders etc. Respect to AoE2 players, seriously.
An easy to implement solution would be to regularly adjust all players ELOs so that the average of the distribution is 1000 ELO. After that put in place, you can give newcomers a free pass (ina certain range) on where they want to start on the ladder. A recommendation system based on single player history also makes sense.
You can do an ELO inflation correction, like FIDE and national chess federations did. And maybe have K factor work better, that you have smaller changes if you're an active player, and if inactive you'll drop quickly due to higher K factor.
Some good points. I haven't played ranked since DE release and I haven't played aoe2 for a couple of years now. I'm still learning all these recently added civs. I really would like to play some matches again, but I think my almost 1300 rating from back in the day is a bit inflated now and I will get my ass kicked. I'm not even able to do a proper Dark Age anymore ^^" ... so I'm staying in my comfort zone in SP and practice against the AI haha
I'll be honest. I regularly watch you and T90, and never played an online game. I really like playing with my friends or vs AI, even managed to beat extreme AI couple of times, but I just prefer not to play with strangers. When I watch and try to evaluate my hypothetical ELO based on the players I watch I would guess it's a little below or around 1000. The fun fact is that I tried twice to play on-line ranked and start this whole ELO adventure and both of the times the servers were undergoing maintenance, so I guess it's not meant to be
If you consistently beat hardest or extreme AI you are around 1k in my experience. It's just that AI usually does the same things and doesn't wall so you'll have to adjust your own game play a bit when going ranked.
Playing with strangers is how you learn the game.
You do not have to play against Extreme AI either. Just make a lobby “4 vs Hard AI” and play. You will eventually learn what other people do, adapt it into your gameplay, and become much better.
Watching videos means nothing because, until you can quantify how good that player really is, you cannot even begin to compare yourself to what you see in videos.
It’s like watching the English Premier League or La Liga - we have no concept of how good these footballers really are until we see them in person.
@@LeftJoystick watching videos is fun, and gaming is all about fun for no-pros :) but I agree, that I would learn much more by playing ranked.
I get my practice playing those good ol 4v4 or 4vAI games. It's pretty fun
This is just how competitive games go. Especially when the casuals leave. I was gold level in Starcraft 2 in 2010 and obviously I had no idea what I was doing in Age of Mythology back in like, '01.
I love your cat in the background
I'm a 20+ year single player only player, and viewer of the competitive scene. A few years ago I tried my hand at multiplayer to see how I would do. And I basically won streak until about 1300, some players thought I was smurfing because I only had a few games but played like I knew a thing or two lol Most of my play experience is 1v1 hard-extreme AI. Mulitplayer AOE2 is just too stressful for me, when I play a game I want to relax, not get white knuckle sweaty, which is essentially AOE2 competitive lol
The emergence of streaming and collaboration has allowed this. When everyone played on MSN Zone from 2000-2008, there was no streaming. You had recorded games but it was view lock at best, there was no Capture Age, no analytics, no APM, what you saw in a standard rec is what you saw - the closest to watching real time was playing co-op and the 'honor' system that those co-oping wouldn't interfere. RUclips was in its infancy, expert players for the most part were still rather unknown aside from what records of them you could download. I achieved a 2100 rating on MSN Zone, but with all of that said above and the emergence of that today, people applying what they know NOW about the game would easily crush most of the competition.
I'm a fan of dirt simple solutions. If someone never played ranked before, the game could just ask them: "How experienced are you in AoE2?" Let there be three or four broad categories, like: 1. I have never played AoE2 before. 2. I know the basics and have a little experience. 3. I have a lot of knowledge and experience. Depending on the answer, put them into Elo 500, 750, or 1000. Add some "placement games" logic on top of that (rapid Elo adjustment) and people should land in the correct spot pretty quickly, ideally not having to lose more than 2-3 games in a row at the start. For example, if the first few games give or take 50 Elo, then a completely new player (starting at 500) losing 3 games in a row would already put them down to 350 Elo.
I will say. I started playing 2 years ago and immediately dropped to like 400 elo. One of my friends who plays kept me going to help me through the first games. I’m 1200 now and still grinding. It’s hard. But I’m very competitive so I’ve had a lot of fun with it
I don't have a lot of time to play the last few years... I feel like my knowledge of the game and my mechanics are becoming slowly better after 20y of playing but my elo just keeps on averaging or dropping from 1100 to 900-1000. You are absolutely right about quality of 1000 ELO player.. one thing that I do notice is the amount of all-in strats or one trick civs is very high. I'm all for the reset proposition!
The ELO system has two main things:
→ Good players have something they can show off. You try to climb the ladder to get a higher ELO-number.
→ You get to play against people of a similar skill as you.
I'd argue that the second one is more important than the first one, and it's what we should focus on.
Currently we got the problem that "average new players" (e.g. around ELO 500) take a long series of losses until they are paired with other similarly-skilled players.
Giving new players a lower starting position will help for a while, but with enough new players joining at 500, that will become the new average, as Hera mentioned.
I'm not sure how a complete reset is supposed to help.
Just erasing all history will make it so _everyone_ gets very one-sided matches after every reset (or after joining again after a break). It might make it "exciting" if you just value the experience of climbing the ladder, but it won't help against having one-sided games for new players. (It does help against the "camping at a high elo", but I'm not sure how much of a problem that is.)
Gotta love how the B roll footage is just Hera getting destroyed by a bunch 1000 elo players.
I'm late to the discussion as always, but I still want to throw in my 2 cents:
Resetting the ladder is likely not a good idea, I'd even expect it to have the adverse effect. We just saw this in RBW, starting everyone at 1k means that new players will very likely just face players far above them. It won't just be a loss anymore, it will be a demolition that is far harder to stomach. They'll feel like they can't even compete. Some might not even know who this Hera guy is, lose to him, and then decide to never play again. That can still happen today, but it's a nigh impossibility. And doing challenges as the higher rated player or such is not really good either, if they realize you're trolling on them, they'll be even more likely to avoid the ladder. And not everyone is as masochistic as me, and don't get me wrong, if I didn't value Hera's time, I would dm him right now to "1v1 me bro". But it would likely turn away even more potential ranked players.
Starting them off at 500 or so wouldn't work either. As mentioned in the video, it would only lower the ELO of the average player, but not their rank. ELO is meant as a long term ranking system, where the majority will, after hundreds of games, end up at 50% win rate. A lower starting point would have the same effect as the inflated TG ELO before, just in the other direction.
What I would suggest is changing the algorithm for the first 10 placement games. During these games, the range of ELOs that is included is a lot wider. In a regular match, a 1k player will almost always play someone 900-1100, but in placement games, it could easily become 600-1400. How about we make it so that during those placement games, there's a higher chance to match with someone lower than your not yet confirmed rating? Like, a 80% chance you match with a lower rated player, 20% higher rated. Even if you lose your first few games, with the wide ELO range, it makes it so you will almost certainly play with players below 1k, and that barrier will go down with you. And if you happen to win, well, then there's really no change, you get the ELO, get up the Rankings, and match with those around your rank - just more likely to be a bit below you.
I do always feel bad when I realize I met a newbie, and sometimes I even got them on their first game. It's hard to make it a fair fight for them, and motivate them to go on. This game is about slow and steady improvement at every level, and I agree that those first games are a harsh teacher. But I also think 20 losses straight is an outlier. Maybe if the game recommended "beating the hard AI" before going on ranked would also be an option? That was when I joined the ranked ladder, and I ended up almost exactly 1k after my placement games.
So yeah, those are my thoughts. Now 1v1 me bro 11
I played AOE2 when it came out at the age of 10. I just got it back after watching you, t90, and Viper. I had the basics down and figures i would ROCK as soon as i started. I lost my first 12 games. lol That being said I am 850 ish elo, 41 wins and 34 losses.... I have had real good dark/feudal ages, maybe 40 seconds idle TC time into castle, trying to optimize Eco micros, following build orders, luring boars, pushing dear. Most of the time my opponent is doing the same. I wanted to quit those first 12 games, but I didn't. That initial elo placement is ROUGH and if i had to do that every so often I don't know if i would keep playing the game. just my 2 cents
I think New Player Lobby like in Voobly would be a even better idea. (In this case a separate ladder for new players, maybe you need to reach a certain rank or maybe 100 games to get to regular ladder)
i actually think the starting elo should go up rather than down. if the average elo is 1000 and that level is so far above the beginner level, i feel like just a range of 1000 is not enough to accomodate for all the skill levels inbetween 0 and 1000. no that does not fix the issue of having to lose 20 games first before reaching your normal elo, but Dawnrim's suggestion here in the comments would help a lot with that.
as for me, ive been fighting this battle for years. i used to be 1600 hitting 1700 a few times, but i easily burn out on any game i play actively and take months off, and every time i come back i drop 200 elo and have to learn tens of new things in order to get back to that 1600 level, be it new build orders that up a villager faster every year or simply just rules of thumb. its not just being rusty like reddit likes to claim. nowadays, im just getting too old and gave up actively trying to improve, and notice more and more people gaining advantage over me. even just yesterday, i was losing a feudal battle on teamgame flank against a 1300 (1v1 elo). luckily i'll keep enjoying this game no matter how far my elo drops.
having this problem rn, me and a bunch of friends who all played aoe2 as kids started playing team games against AI last year and eventually that got kinda boring. so i suggested we switch to ranked, and we just keep getting absolutely stomped over and over again, to the point where 2 of my friends don't really wanna play anymore at all.
it's strange because i know there's other guys like us out there who probably feel the same way, but by the time we'd reach each other's elo level where we could comfortably play against each other a lot of us will have dropped out due to the frustration.
the worst part is that the whole reason we went for ranked instead of just joining lobby games was that we didn't wanna get stomped by high level players, we just assumed that the matchmaking would figure out we're not competitive players but oh boy were we wrong.
it will eventually, but the problem is it takes some games, also problem is if the elo of your team is wildly different, also team game elo is sometimes a bit misleading
I have observed this myself. I dont really play AoE, but some 3 years ago I played for a little bit and also did some ranked games and won against 1200 ELO players with reasonable ease. I dropped back into some online play earlier this year and was beaten handsomly at the same rank. I wanna say Im at pretty much the same skill level as 3 years ago. Still easily beat the hard AI, which is really the only thing I play. So yeah the average skill definitely increased a lot.
An easy fix would be to start the ELOs lower. What you're saying is bang on. When me and my friends were playing ranked games years ago, we were steady at 1100-1200, now we're all sat at 900 when coming back to the game lol. If the game wasn't competitive, it wouldn't thrive. So starting people at maybe 700 ELO to find their level is maybe a good move
After much consideration I started playing ranked about a month ago, i knew I was gonna lose quite a few at the beginning but slowly started to settle around 800-850 elo. Suddenly in one week I just went down to 700 elo. A streak of 7 losses in a row and some of them just really bad losses. Don’t really know if I really hadn’t gone to my real Elo and now I am but those losses do make you consider if ranked is still worth the frustration that sometimes comes with it
I started 0 20 or something like that too. I remember saying to myself ''you just need to tank it until you get better bro''
I'm actually 100 loses 110 wins or something like that, around 1030 elo. At least I can put some fight now!
Something else to keep in mind, we have an "unranked" matchmaking system that uses a hidden ELO. It's basically more streamlined version of the ranked matchmaking we already have but simply because it doesn't have a visible ELO a lot of more causal players will just go there. I discovered the difference in skill/tryhard when I was experimenting with strategies and just wanted to mess around and I was surprised by how big of a difference there was. 1000 ELO in "unranked" is probably still newbie kind of players or at least it feels that way.
It's so much simpler than that, the first X number of matches have boosted elo influence. instead of 20 matches to find where people land, it should take 5-10. Additionally having UI that explains to people briefly about the elo system and that "we're starting you at the middle point, and it should take about 5 games to settle you in." so players know that there's an end in sight and something to work towards.
I've definitely been there. usually a teamgame player of 3v3 and 4v4s, have dabbled a bit into 1v1s. about 1/3rd of my games feel fair, or I have the feeling I did something obviously wrong that I can improve on. The other 2/3rd of games I'm getting completely sweeped by a player that feels way better than those other opponents, like its an entirely different beast. Ended up quitting ladder alltogether, it just wasn't fun with how big the skillgap was.
Recently you said that the average 1000ELO player today would beat Viper 5 years ago. It seems this is true to an extent, and it makes me sad. My micro is decent and my build orders need work but my dark age is at least reasonable - but I can't compete at 1000ELO.
I've been playing AoE2 since its original launch, I love multiplayer AoE and I've been trying to learn from watching pros since 2012 and have gotten a lot better, but my decision making is terrible because I have no ability to predict what my opponent will do. That's my weakness in literally everything. And I'm so competitive that I take losses really hard. I can only do like one ranked match at a time before I have to go mope around for a while at how terrible I am for something I've put thousands of hours into, and I don't want to know my actual ELO because it's going to take too many upsetting losses to get there.
Anyway. The barrier to entry into ranked is very difficult for AoE2.
I've played far more ranked AoE4 (even though the poor balance keeps me from taking that game seriously); it's easier on the psyche to do 10 placement matches and get categorized than just to start at a standard rank which is apparently too good, and make you lose indefinitely until you reach your true rank.
Ive always been a TGs scrub and I noticed this when they reset the TGs Elo. I used to be ~2200 before and I faced scrubs most of the time.
After the reset my peak has been 1500 and 90% of the times these 1400-1500 peps are also 1500+ in 1v1. Idk if I need to play some 1v1s at this point. I don't really feel overwhelmed by my opponents but I usually don't do any dmg to them in feudal and the lot.
I also feel like I dominate post-imp late game but it's all throughout castle when I lose the most.
I also think some sort of Beginner Sandbox would be good. Like a test you have to do before you start. (Match vs. AI where different metrics are measured like apm, timings, ressource-balance etc. together with the difficulty you can beat the AI. If you beat an extreme AI, you're probably no Low Elo Legend.)
This is why when anyone comes to the Steam forum asking about ranked, about five people say. Go watch Hera, TheViper, etc pro player, watch Spirit of the Law number crunching and theory, and T90/Memb for casted games. Go practice the Art of War and try to beat the extreme AI or at least the hard one. Because doing anything else than that is going to result in some crushing defeats and you'll have no idea why.
Also I think some people need watch Spirit of the Law's video on ELO and how it works, because seeing some comments that don't understand why lowering the starting point does not work.
In sc2 u have placement matches and seasons. If you havent played before you are going to start of playing vs the avg mmr, and for every loss during the placemnet matches the mmr loss/gain will be amplifyied. So if you lose all of you matches, every match would be vs a much weaker opponent.
Great video! I have a solution: when I start the game after not playing for a while, I quit games voluntarily to lose Elo on purpose. I do that until I am sure i won't lose many games in a row.
Hence I would advise beginners to quit 10 to 20 games in a row on purpose to have this Elo climbing satisfaction when they decide to start trying for real.
Everyone should start in 500 ELO. I felt like that at the beginning and that's the reason I only play team games now. Also agree with the seasonal ranked, but with long periods of time and not total restart. For example, 1000 ELO could restart in 500, 1500 restart in 1K and 2K in 1500. Thanks for this video. I hope Microsoft hears you.
Placement matches for new players would be the best solution I think. New players would play 5 games which would set their starting ELO. This would start new players at an appropriate ELO. On a side note having ranks displayed for ELO - bronze, iron, silver, gold, would be a fun way to incentivize more ranked play.
I have over 1000 hours combined age of empires to HD edition and definitive edition. I consider myself a new player when it comes to ranked. I consider myself below when it comes to rank. The reason why is because I had to practice so much learn build orders to be able to maintain and do something in a ranked game. When I play a one V one it’s very intense and I’m sweating, so hard to get a minor advantage over my opponent and one or two mistakes that I make will cost me the game hence why I play team ranked instead solo ranked
I am one of the people that went like 3-20 in ranked before finding my rank. AOE2 was my first RTS and even knowing a lot of what to do isn't enough to save my 6 APM ass. It's pretty hard to motivate oneself to play when you know you have a fair few losses ahead of you before you start finding people around your skill level. I think the best solution would just be making point losses steeper for new accounts with some system to detect if they're getting stomped, so instead of taking 10 games to drop to 700 elo it takes like 3-5. Cause I don't think this issue is wanting to be at a certain elo number, but instead wanting to find matches which are relatively even.
Hello Hera, I am from the Sc2 Community and played Age 2 only as a kid. After playing the sc2 ladder I learned a lot of things for RTS in general and tbh I think Sc2 are in a lot of ways harder (more micro and faster pacing) and after playing Age 2 ladder the first time I won because I had a faster play style and looked up some conters like pike -> horse -> skirms etc. And only this was enough to win. Now I think many player go to Age 2, because there is RTS left with a big community and if you were good in your rts you probably can learn the ways of age 2 very fast what leads to good players joining the ladder and destroing the match making. Like I was Diamond in Sc2 and than came to the 1000 ELO Age2. Sry for my english
Haha yeah my friend who plays Brood War doesn't consider AoE2 a real RTS.
Starcraft is a joke of an RTS. Its all tactics, zero actual strategy. Also SC2 is way easier than AOE2. And Hera disagrees about pikes countering knights ;)
those players today are insane. ive got achievement from HD edition defeat 1v7 hardest AI, ive got half of campaigns on hard difficulty and im stuck at 1000 elo
I've never played a campaign or AI
highest elo 1600 1v1 and TG, been playing 2.5 years
My point is only way to learn is playing higher elos than yourself then studying the replays, or just watching replays of higher elo players than yourself , For me I'd always watch about 200 elo above where I was at, it's more relatable. Now I'm finally at 1600 watching the Pros is more slightly relatable, but I've realised I'll never be higher than 1700, never ever ever, not getting younger, not getting faster, don't kid yourself, eapm is legit. Someone's who's max eapm is 60 will never beat a 200 eapm player with tbe and same knowledge.
The top % of players have something we don't, they are special and no amount of training can make up for God given gifts....
The problem with Ladder 1v1 is that at ranges 1000-1200 which is the starting area, are many smurfs. I would say around 30-40% at least. People with higher ranks that dont like tgs and wanna chill or build a new account. I consistently beat in tgs 1400-1500 elo players in 1v1 fights (zero pocket help) with ease and 2v1 sometimes, and lost most game in 1v1 with 1200-1300 players. Then you check their total games and its like 50 or 200. 200 games you dont even know the civs well yet. The second problem is that when a player advances from 1200 > 1300/1400 he meets really good players, will lose some games, drop back to 1200 as expected, then meet the smurfs and drop more. What I would do to minimise this is take the total games of players into account, adding BOTH 1v1 and TGs. That way you wont be matched at least with a good player with new account which will cut down the smurfed games a lot. Voobly had a similar mech where people with less than a number of games, never match against people with more games, regaldless of ELO.
If you do a reset, Hera could be matched against a player who was 200 elo before the reset. It would lead to more mismatched games until elos drifted back to where they started.