Hey, Coach! I talked with him, and Jack J (author of the games per session post) would love to come on and chat with you further on this! Any way I could get the two of you in touch?
The topic of how many games is optimal is interesting. For me I struggle to get ranked games in, cause I often prioritize playing with friends. League is a hobby after all and even though I do want to play soloQ I find the company and companionship of playing with friends more rewarding recently.
Honestly playing a lot of norms with friends will probably lead to a more fun ranked play I at one point didn't play rank for a year and when I did I went from bronze I to plat IV with a 70% w/r because my mmr had become way higher than my rank and I had improved so it made those games way more fun than they would be if I tried to play non stop in rank
You can still improve while playing with friends. Heck, you can get huge jumps in improvements if you are utilizing that time to limit test and/or practice specific mechanics on a champ. Even just hard focusing on not missing CS, lane fundamentals (such as minion wave management) etc. can be practiced while playing with friends.
what I get from that graph is that the only significant change in win chance comes after 12 games. so burnout is DEFINITELY a big factor in win chance.
9:45 "A paper from October 2011 by Yaari and Eisenmann, a large dataset of more than 300,000 NBA free throws were found to show "strong evidence" for the "hot hand" phenomenon at the individual level. They analyzed all free throws taken during five regular NBA seasons from 2005 to 2010. They found that there was a significant increase in players' probabilities of hitting the second shot in a two-shot series compared to the first one. They also found that in a set of two consecutive shots, the probability of hitting the second shot is greater following a hit than following a miss on the previous one."
not sure how many variables are controlled for, but i agree with curtis’s biases. i imagine there’s a lot of players who lose their first ranked game then just stop lol
@@kingdomovies ngl i don’t believe you. i tend to join servers when ppl drop links in the chat even tho i wouldn’t call myself a member of those communities; i couldn’t tell you the amount of times there’s been @everyone ping for a duo, i hop in vc to play to watch, they lose and are immediately done for the day or queue a ton of aram/arena between their block. tbh i don’t imagine this study is controlled for ppl who have the habit of spamming other modes to de-tilt after a loss, so that could also be represented in the one and done blocks
I think another aspect of the 3 block method that is missed in this data is the learning/analysis part of the journey. I think 3 games is a reasonable amount to be able to remember back through all the games and their takeaways to ruminate on throughout the day. After 5 games, I don't know how people don't get mixed up, especially if their keeping a tight champ pool. TLDR: The data presented is very short-term focused with very little wr delta, while the 3 block method is about long-term improvement.
The blocking makes complete sense. I always do an unranked warmup and two more games. I completely agree that if you are streaking, there is no real reason to stop, but I do not do better continuing after two consecutive losses. I can feel myself getting emotionally attached to my or other players' bad calls after that. I also noticed that I do not go on losing streaks anymore once I implemented this rule.
7:59 to me the three block strategy feels like an investment principal. In the sense that almost all successful investors may not win the same way, but they all know how to lose. They do not compound their losses
at around 7 minutes, you are correct that there isn't enough to draw conclusions about the data. However, this includes the confidence-based gameplay you speak of. That's one possible solution, and maybe even a likely solution. But the data is not telling you anything about confidence-based gameplay. That's your own intuition / explanation for the data. Some additional data that could confirm or deny your explanation might be looking at the winrates by session length filtering for "previous game was a win" vs "previous game was a loss"
54:30 that whole thing about reframing the blocks to sort of loosen yourself up even if you're losing so you can get through them, I instinctively did that for myself so I could have an easier time dealing with the games I played. First block, two losses then a win. Second block 2 losses and as I'm in the loading screen for my third, I sort of randomly just thought it was really funny if this block ends in a win and how hilarious to myself it would be if the third block repeats. No longer thinking about the losses, especially after reviewing the games and I'm still excited to get into the third block. Ready to get that 2 losses and a win again. Get comfortable with failure, and master it, so you can then succeed of your own accord
Great episode guys, lots of interesting things covered on this one. It would be interesting to hear what evidence would convince Curtis and Nathan that another approach was superior to the 3-block system. I personally see the strength of 3-blocks, and think it is sensible to wait for consistency between many independent tests over time before adopting a new theory, particularly when the current one seems to otherwise have good fit with experience. However, as their ultimate goal appears to be to develop the most efficient system for learning the game and climbing in ranked, considering 'what would it take to change your mind' might help remain intellectually honest and not lose sight of the goal in order to preserve a cherished idea.
with the starting point of this data, they would probably want to see the different bars be differently sized to reflect the percent of the data it represents Then, they would be looking for evidence that the increased winrate is or isn't tied to winstreaks
I love the statement regarding LP volatility. I personally play in blocks of four because I find that I have the extra endurance consistently but I actually also need to play in blocks to curb the addiction and dopamine hit from playing longer sessions. Four games being “about a division also lets this work for me.
About the "Growth in League" section where Swagnaur talks about the mental importance of simply having a process. I wholeheartedly agree. But I think its less about focusing on learning and more about shifting the focus away from ranked goals. I've said this on Skool as well but when ranked goals are at the forefront of playing it makes you tilt easier and play worse.
I do not follow the 3 block method. Warm up depends on your own capability and awareness of your mental state, which most people are frankly very poor at doing (they tilt, they flame teammates etc. None of which would ever happen if they were self aware of their own mental state). I follow a 2-3 bad game stopping point method. Sometimes that means play 2 games and stopping because im not playing well at the moment, break for 1-2hrs minimum before playing again. Sometimes that means win 2 games, lose 1 game, win 1 more, lose 1 game, lose 1 more. Stop. Generally this means downward spirals are only limited to 2 games "negative" at a time, and take several days of consistent bad play to really accumulate, prevents spiralling. Sometimes a loss happens but you legitimately played well, that means you can keep queueing. If you lose a 2nd time in a row, start REALLY considering stopping. 3 almost always stop for the day entirely.
To me, winning the early game means exceeding expectations in terms of how the matchup is supposed to go. If your champion is stronger at early levels, that means gaining a significant lead on them in gold and xp. How much exactly depends on the rank you're at. The lower the rank, the larger the gap must be, because you're going to make more significant mistakes to squander that lead. But if your champion is weaker at early levels, you might already count an early game as won if you manage to stay even.
Just speaking from my experience, my best session I’ve had climbing was a 45 match spree, during this time went from emerald 1 to diamond 2. It seem d like the start was super slow but after I started winning it just didn’t stop
A few things I do is that I generally play 3 blocks 3 times a week and I do these games on stream meaning I have a set time in order to get this games done. I don't really have the luxury of overthinking a previous loss with how I play the game and how much time I am limited to to get these three games in, which is about three hours. I practice anywhere between 40 minutes to 1 hour and 30 minutes daily with zero VoD review so far and reach my best win / loss of 10 w's over l's. I am still low elo and the number of games so far aren't much of a sample size, but that's where I'm at generally speaking before my A/C died. Because of that I've been avoid high blood pressure games in this heat wave.
3 blocks have never worked for me. Sometimes it takes me 3 or 4 games to get into the groove. Some of my best winstreaks have come from a session that started 0-4 and ended 6-4
I hope people will understand that these insights need to be taken at their face value and nothing more. The TL:DR of the post is quite concerning in its verbiage, drawing such conclusions from data without any experimentation or controls in place. I agree with Coach and Nathan- trying to draw conclusions where context doesn’t exist is detrimental. It’s important to recognize the shortcomings of such an analysis and not simply fill in the blanks. I wish more drilling down was done, showing the number of people playing certain numbers of games per session, their average ranks, role differences, etc. I was left with more questions than answers, and hope people don’t blindly start altering their routines because “data says so”.
We need a true randomized control trial of league players playing a different of games per session to sort out this data. Nathan and Curtis are the doctors on the front line reporting clinical results and case studies while that reddit post was a weak powered study. As always, trust your doctor over a random study!
I believe the data about taking a break ultimately shows that as long as you’re playing another game within 24 hours (and therefore not rusty, and may still have the previous games mistakes in immediate thought) you are more prepared to win the next match. So, QUEUE UP
It would have been nice to see errorbars on the first graph since the higher amount of games likely have high variance due to low sample size, and how the distribution would look if you only took players who played 6-7 games (i.e you see their individual winrates on each game, so that the first win for example isnt washed by people who stopped after one game)
The games % won by session isn't telling the number of players at that point in the session. Meaning I'm pretty sure the "first game" is 100% of the data sample size. Because it has to be. However, what is the drop off after each game in a session and what was the result of the game before the drop off? This is what Curtis is alluding to with the bias. The data points are not biased, but the presentation is.
I’d like to start by saying that I’m loving the skool forum! I’ve already gathered soooo much valuable info about my lane (top) through the garen guide. Will you guys do anymore classrooms like that or was that like the try before you buy?
Nah, not really. It is actually a study that gets varying results. If you isolate it specifically, hot hand does seem to exist. However, there are other variables that may play a role. Maybe Bob makes 50% of his shots on average. Against Bill, who is taller and more athletic, Bob shoots at a lower %. However, when Jim is instead guarding Bob, Bob makes a higher % of his shots. It appears he has a hot hand, but the matchup actually plays a bigger role. Now, with that said, there are some people who are legit anomalies. They might start off shooting 20% against Bill, but then randomly start shooting 60%+ against Bill. I think we need to realize how important confidence is to player success. Hot hand, logically speaking, could very well just be 'Player X is feeling extremely confident right now, so they are making more shots.' Unfortunately, just due to how many variables there are, we can't really know if a single thing determines 'hot hand', if it's a multitude of things, or if it just varies player to player. Studies done to disprove 'hot hand' were riddled with data selection bias and just not all that accurate. With that said, hot hand isn't something we can truly tell just by purely looking at data. I say all this for a multitude of reasons, but I will try to translate to league terms how 'hot hand' might be real and how it might also be falsely used at times: Player Tom is extremely aggressive. He is great at punishing passive players. If Tom gets lucky that day and draws a bunch of passive lane opponents, Tom may appear to have 'hot hands'. In reality, Tom just got some pretty favorable odds. Now, let's say Tom is bad vs other aggressive players. However, on this specific day, he drew 50/50 in a 6 game set. 3 games passive opponents and 3 games aggressive opponents. He performed exceptional and won all those games. That scenario, where Tom performed better than normal against opponents he normally struggles against and just overall seems unstoppable, very well might be a case of him having hot hands. The only real argument to be made would be: what variables led to him having hot hands on that day? Did he get adequate sleep, eat healthy, exercise, etc.? How was his overall mental healthy and confidence? List goes on and on.
@@faveologytheres also the fact that there's certain champions your really good against and u played against those champs for like 5 to 6 games straight
16:04 but how do u know when not to make a big deal of a game or not? its like imma passionate person in general its almost like your asking me to not be passionate about the game at the same time i am tho ,i see about the not tilting thing because when people tilt they feel like a victim n blame everything but themselves which is good to not do, its just easier for u guys because u completed your journey i feel like so since u accomplished a big goal u can kinda just kick bacc n not feel bad about losing cause u already made it, but u see for me my problem i struggle with is losing a game and feeling like im behind , which in turn makes me feel like im behind on time , in time, which makes me rush my next game even after taking a break sometimes because internally i get impatient subconsciously on a deeper level cuz even when i repeat to myself in my head "chillout" i dont have all the time in the world to play league cuz i know one day ill end up having to stop playing league cuz ill just be too old , or what if something in my life happens to where i cant play league like i am now and what if i dont get to where i wanna be before that time might come, so it makes my gameplay become rushed at times and it shows in my performance, to disconnect myself from my emotions is hard for someone like me , imma fearless passionate person irl , imma man of action , i dont just talk that talk if i cant bacc it up , i demonstrate n show in conjunction with my words and can accomplish things off shear willpower alone , not being able to "just do" makes me impatient with my journey cuz i feel like time i running out cuz surely all things must come to an end at some point in time
Problem with the data set is that there is no control group - It creates a fundamentally dishonest argument. The author looked for trends and completed a write up this comes from not understanding / respecting statistics.
Hey Coach! I’ve been seriously considering the midlane academy but I am a Zed OTP, do you really think it’d be worth while for me to pay for coaching as a Zed OTP or should I pick up a few other champs and then join?
This is just a personal thing, I can't ever play 7 matches in a row without resting. Even playing 3 in a row is a bit much, I always take at least a few minutes between matches. Specially if I lose, that's just me, though.
Even if we were to take the findings at value without accounting for the first game bias, its less than 1% difference in win rate, actually quite tiny. If we were to say, "play 6 games in a row and increase youe chance of winning those games by less than 1% each" i dont see people being enthused by that. Typical case of a graph making it look significant when its not.
This article makes a too strong relation between statistics and probabilities. Statistics is an area of mathematics which aims at analysing data, probability an area concerned with likelihood of outcomes. The statistics showing a 49.6% winrate after a loss does not mean you have a 49.6% chance of winning your next game after a loss, what you can say is that there likely is some factor that produces such a distribution in the data. What I mean by this is that the statisical value of 49.6% wr does not imply a likelihood of 49.6% for that outcome, but that something or somethings are perturbing the data away from a clean 50/50, which manifests as this wr in the data, but where the underlaying cause is unknown.
I thought it's really bad these days to hurt your own tempo to cover early waves now due to the lack of gold and experience compared --- am I wrong, Nathan? I remember we used to have no problem with covering early waves if someone needed to base or had just died in the past, but the balancing changed and now it feels like you're sacrificing yourself too much for your teammate to get a little/to dampen their losses a little(they're most likely still going to be weaker than their lane opponent no matter what at this point so why make yourself weaker than the enemy jungler and give them free reign to steal camps on the other side of the map too?) Isn't it better to ensure you hit your item spike and/or instead of sacrificing tempo to continue to play around a lane that is losing so quickly, to sacrifice that tempo to swap what lane you're playing around? Isn't this what you reference when you say adapting?
I mean, to be honest, he does address his external and internal validities, sort of. However, these margins are so small, that its hard to claim anything from the results. He doesn't cite his sample size, p values, effect sizes or anything. I get that its just raw numbers, but there's so many confounds to someone playing a game that its hard to discern how the numbers actually matter. What is their emotional state, how much food and water have they had, what is their mouse/keyboard history. It was fun, but its nothing you can really draw a conclusion from.
They said that these are people who are likely to use software to analyze games, right? If you’re only going over the first two minutes of your game while you are in for your next game better than not looking at anything, but if you have a 10 minute gap in between games, then wouldn’t that mean that you are possibly looking at the clips from your last game and trying to really understand what happened during those first few bad plays? Watching your death and trying to figure out where the Ganger came from and why you might not have noticed on time. Having a 10 minute gap between games and that being bad seems very strange because 10 minutes is not very long.
Is this data from people who 3 block btw? I imagine the data would look different for people who review their games because it's mentally taxing to review a game. If people aren't reviewing then 6~7 games can make sense since ur suffering from less mental fatigue at ur 4th game with no review compared to 3 games with review in all likelihood.
I play league until i lost 2 consecutivr games then im done for the day only during the week. If during the weenend, ill just take a break and play other games.
Tilt Teemoing is one hell of an attention grabber. They already improved the bots btw. They're better than they used to be. Not quite player level, but definitely better than they used to be.
Curtis really wants to interpret the data in a way that confirm his apriori conclusions, instead of having a curious mindset. A bit disappointing, but I do think 3-block process is a good tool despite the data
As all these mathematical analisis' coming out of the league subs, this one too has a glaring oversight. The oversight being that momentary winrate is a lot less indicative of quality gameplay than assumed. League has this habit of making you win when you shouldn't, and making you lose when you play good, and lots of skills are completely invisible to any form of stats. Improvement is a very lenghty process and your learnings in some areas might not show for days, weeks or months. Anecdotal counterpoint: at the end of the split (before blackfire), I got 350 LP in high dia on karthus bot in 12 hours grinding. How does that fit into the data? I won a lot because i grinded and cheesed at a time where people stopped caring, and I got lucky. It's not because I suddenly became a much better player or because 12h is the optimal session size. And I certainly didnt learn anything. You would be surprised how much worse you can play compared to your peak performance, and still have similar winrate. You can be actively improving while on a 30% winrate, and you can play with 53% for 500 games and still learn nothing.
I'm a data analyst, my mom is a data architect and we both concluded that what Curtis said is actually correct. Obviously 100% of the sample size are a part of the "first game" category, but I guarantee that most people in that "control group" don't get to the 6 game mark, let alone the 12+ game mark. It matters a lot how many of the "6 games" people make up that winrate, because if it's a very small percentage of the people used for the study, it's not even close to a valid result. If most of the studied people don't make it to that point, then the result is obviously biased towards the people that can afford to play 6 games or more, or don't lose their minds in a loss streak when they get to their 6th game
Excited to hear there's an ADC academy coming! I feel like ADC has the worst educational resources out of the 5 roles right now.
I agree. However I do like this guy:
youtube.com/@bizyze?si=S71QUdR8Jzz4yYPr
3 dollar patreon xd
@@akirakato1293 not the jax coaching 💀
How
agreed
Hey, Coach!
I talked with him, and Jack J (author of the games per session post) would love to come on and chat with you further on this! Any way I could get the two of you in touch?
this would be sick
Did this ever happen?
The topic of how many games is optimal is interesting. For me I struggle to get ranked games in, cause I often prioritize playing with friends. League is a hobby after all and even though I do want to play soloQ I find the company and companionship of playing with friends more rewarding recently.
Wholesome, but for me the grind never stops
Honestly playing a lot of norms with friends will probably lead to a more fun ranked play
I at one point didn't play rank for a year and when I did I went from bronze I to plat IV with a 70% w/r because my mmr had become way higher than my rank and I had improved so it made those games way more fun than they would be if I tried to play non stop in rank
You can still improve while playing with friends. Heck, you can get huge jumps in improvements if you are utilizing that time to limit test and/or practice specific mechanics on a champ. Even just hard focusing on not missing CS, lane fundamentals (such as minion wave management) etc. can be practiced while playing with friends.
@@faveology Oh yeah absolytely thats what I meant by playing norms too, I just meant non rank play
Too bad for me my best friend grinds ranked so I do too 😭
what I get from that graph is that the only significant change in win chance comes after 12 games. so burnout is DEFINITELY a big factor in win chance.
9:45
"A paper from October 2011 by Yaari and Eisenmann, a large dataset of more than 300,000 NBA free throws were found to show "strong evidence" for the "hot hand" phenomenon at the individual level. They analyzed all free throws taken during five regular NBA seasons from 2005 to 2010. They found that there was a significant increase in players' probabilities of hitting the second shot in a two-shot series compared to the first one. They also found that in a set of two consecutive shots, the probability of hitting the second shot is greater following a hit than following a miss on the previous one."
not sure how many variables are controlled for, but i agree with curtis’s biases. i imagine there’s a lot of players who lose their first ranked game then just stop lol
Yeah if you are in a losing streak you a more likely to stop than if you've won 5 in a row
Like 0? Ive yet to see someone who loses one then just quits
Up to 3 I can imagine
@@kingdomovies ngl i don’t believe you. i tend to join servers when ppl drop links in the chat even tho i wouldn’t call myself a member of those communities; i couldn’t tell you the amount of times there’s been @everyone ping for a duo, i hop in vc to play to watch, they lose and are immediately done for the day or queue a ton of aram/arena between their block. tbh i don’t imagine this study is controlled for ppl who have the habit of spamming other modes to de-tilt after a loss, so that could also be represented in the one and done blocks
I think another aspect of the 3 block method that is missed in this data is the learning/analysis part of the journey. I think 3 games is a reasonable amount to be able to remember back through all the games and their takeaways to ruminate on throughout the day. After 5 games, I don't know how people don't get mixed up, especially if their keeping a tight champ pool.
TLDR: The data presented is very short-term focused with very little wr delta, while the 3 block method is about long-term improvement.
The blocking makes complete sense. I always do an unranked warmup and two more games. I completely agree that if you are streaking, there is no real reason to stop, but I do not do better continuing after two consecutive losses. I can feel myself getting emotionally attached to my or other players' bad calls after that. I also noticed that I do not go on losing streaks anymore once I implemented this rule.
hey dota player here I am enjoying dota much much better after listening to couple ep of urs thanks allot guys ur amazing!!
7:59 to me the three block strategy feels like an investment principal. In the sense that almost all successful investors may not win the same way, but they all know how to lose. They do not compound their losses
at around 7 minutes, you are correct that there isn't enough to draw conclusions about the data. However, this includes the confidence-based gameplay you speak of. That's one possible solution, and maybe even a likely solution. But the data is not telling you anything about confidence-based gameplay. That's your own intuition / explanation for the data.
Some additional data that could confirm or deny your explanation might be looking at the winrates by session length filtering for "previous game was a win" vs "previous game was a loss"
54:30 that whole thing about reframing the blocks to sort of loosen yourself up even if you're losing so you can get through them, I instinctively did that for myself so I could have an easier time dealing with the games I played. First block, two losses then a win. Second block 2 losses and as I'm in the loading screen for my third, I sort of randomly just thought it was really funny if this block ends in a win and how hilarious to myself it would be if the third block repeats. No longer thinking about the losses, especially after reviewing the games and I'm still excited to get into the third block. Ready to get that 2 losses and a win again. Get comfortable with failure, and master it, so you can then succeed of your own accord
Great episode guys, lots of interesting things covered on this one.
It would be interesting to hear what evidence would convince Curtis and Nathan that another approach was superior to the 3-block system.
I personally see the strength of 3-blocks, and think it is sensible to wait for consistency between many independent tests over time before adopting a new theory, particularly when the current one seems to otherwise have good fit with experience.
However, as their ultimate goal appears to be to develop the most efficient system for learning the game and climbing in ranked, considering 'what would it take to change your mind' might help remain intellectually honest and not lose sight of the goal in order to preserve a cherished idea.
with the starting point of this data, they would probably want to see the different bars be differently sized to reflect the percent of the data it represents
Then, they would be looking for evidence that the increased winrate is or isn't tied to winstreaks
I love the statement regarding LP volatility. I personally play in blocks of four because I find that I have the extra endurance consistently but I actually also need to play in blocks to curb the addiction and dopamine hit from playing longer sessions. Four games being “about a division also lets this work for me.
About the "Growth in League" section where Swagnaur talks about the mental importance of simply having a process. I wholeheartedly agree. But I think its less about focusing on learning and more about shifting the focus away from ranked goals. I've said this on Skool as well but when ranked goals are at the forefront of playing it makes you tilt easier and play worse.
I do not follow the 3 block method. Warm up depends on your own capability and awareness of your mental state, which most people are frankly very poor at doing (they tilt, they flame teammates etc. None of which would ever happen if they were self aware of their own mental state).
I follow a 2-3 bad game stopping point method. Sometimes that means play 2 games and stopping because im not playing well at the moment, break for 1-2hrs minimum before playing again. Sometimes that means win 2 games, lose 1 game, win 1 more, lose 1 game, lose 1 more. Stop. Generally this means downward spirals are only limited to 2 games "negative" at a time, and take several days of consistent bad play to really accumulate, prevents spiralling.
Sometimes a loss happens but you legitimately played well, that means you can keep queueing. If you lose a 2nd time in a row, start REALLY considering stopping. 3 almost always stop for the day entirely.
These two guys cook. They are doing a great job on data analysis.
To me, winning the early game means exceeding expectations in terms of how the matchup is supposed to go. If your champion is stronger at early levels, that means gaining a significant lead on them in gold and xp. How much exactly depends on the rank you're at. The lower the rank, the larger the gap must be, because you're going to make more significant mistakes to squander that lead. But if your champion is weaker at early levels, you might already count an early game as won if you manage to stay even.
Just speaking from my experience, my best session I’ve had climbing was a 45 match spree, during this time went from emerald 1 to diamond 2. It seem d like the start was super slow but after I started winning it just didn’t stop
Why no Curtis's Clip Corner that's my favorite part and look forward to weekly
A few things I do is that I generally play 3 blocks 3 times a week and I do these games on stream meaning I have a set time in order to get this games done. I don't really have the luxury of overthinking a previous loss with how I play the game and how much time I am limited to to get these three games in, which is about three hours. I practice anywhere between 40 minutes to 1 hour and 30 minutes daily with zero VoD review so far and reach my best win / loss of 10 w's over l's. I am still low elo and the number of games so far aren't much of a sample size, but that's where I'm at generally speaking before my A/C died. Because of that I've been avoid high blood pressure games in this heat wave.
3 blocks have never worked for me. Sometimes it takes me 3 or 4 games to get into the groove. Some of my best winstreaks have come from a session that started 0-4 and ended 6-4
I hope people will understand that these insights need to be taken at their face value and nothing more. The TL:DR of the post is quite concerning in its verbiage, drawing such conclusions from data without any experimentation or controls in place. I agree with Coach and Nathan- trying to draw conclusions where context doesn’t exist is detrimental. It’s important to recognize the shortcomings of such an analysis and not simply fill in the blanks. I wish more drilling down was done, showing the number of people playing certain numbers of games per session, their average ranks, role differences, etc. I was left with more questions than answers, and hope people don’t blindly start altering their routines because “data says so”.
i immediately thought of mustang too! it just feels right that garen drives a classic muscle car
We need a true randomized control trial of league players playing a different of games per session to sort out this data. Nathan and Curtis are the doctors on the front line reporting clinical results and case studies while that reddit post was a weak powered study. As always, trust your doctor over a random study!
wallahi i’m glad that adc main asked that question 🙏 legit not nearly as much as adc learning content compared to the other roles
I suffer so much from ranked anxiety yet never feel like it effects my gameplay just my ability to queu me
yup same exact feeling here
I believe the data about taking a break ultimately shows that as long as you’re playing another game within 24 hours (and therefore not rusty, and may still have the previous games mistakes in immediate thought) you are more prepared to win the next match. So, QUEUE UP
There is a video by DanteAxe who made league animations, displaying Garen driving a car to chase master Yi… so this question has already been answered
It would have been nice to see errorbars on the first graph since the higher amount of games likely have high variance due to low sample size, and how the distribution would look if you only took players who played 6-7 games (i.e you see their individual winrates on each game, so that the first win for example isnt washed by people who stopped after one game)
The games % won by session isn't telling the number of players at that point in the session. Meaning I'm pretty sure the "first game" is 100% of the data sample size. Because it has to be. However, what is the drop off after each game in a session and what was the result of the game before the drop off? This is what Curtis is alluding to with the bias. The data points are not biased, but the presentation is.
I’d like to start by saying that I’m loving the skool forum! I’ve already gathered soooo much valuable info about my lane (top) through the garen guide. Will you guys do anymore classrooms like that or was that like the try before you buy?
9:43 hot hand has actually been disproven
Nah, not really. It is actually a study that gets varying results. If you isolate it specifically, hot hand does seem to exist. However, there are other variables that may play a role. Maybe Bob makes 50% of his shots on average. Against Bill, who is taller and more athletic, Bob shoots at a lower %. However, when Jim is instead guarding Bob, Bob makes a higher % of his shots. It appears he has a hot hand, but the matchup actually plays a bigger role. Now, with that said, there are some people who are legit anomalies. They might start off shooting 20% against Bill, but then randomly start shooting 60%+ against Bill. I think we need to realize how important confidence is to player success. Hot hand, logically speaking, could very well just be 'Player X is feeling extremely confident right now, so they are making more shots.' Unfortunately, just due to how many variables there are, we can't really know if a single thing determines 'hot hand', if it's a multitude of things, or if it just varies player to player. Studies done to disprove 'hot hand' were riddled with data selection bias and just not all that accurate. With that said, hot hand isn't something we can truly tell just by purely looking at data.
I say all this for a multitude of reasons, but I will try to translate to league terms how 'hot hand' might be real and how it might also be falsely used at times:
Player Tom is extremely aggressive. He is great at punishing passive players. If Tom gets lucky that day and draws a bunch of passive lane opponents, Tom may appear to have 'hot hands'. In reality, Tom just got some pretty favorable odds. Now, let's say Tom is bad vs other aggressive players. However, on this specific day, he drew 50/50 in a 6 game set. 3 games passive opponents and 3 games aggressive opponents. He performed exceptional and won all those games. That scenario, where Tom performed better than normal against opponents he normally struggles against and just overall seems unstoppable, very well might be a case of him having hot hands. The only real argument to be made would be: what variables led to him having hot hands on that day? Did he get adequate sleep, eat healthy, exercise, etc.? How was his overall mental healthy and confidence? List goes on and on.
@@faveology its literally called the hot hand fallacy and was disproven, read up
@@faveologytheres also the fact that there's certain champions your really good against and u played against those champs for like 5 to 6 games straight
16:04 but how do u know when not to make a big deal of a game or not? its like imma passionate person in general its almost like your asking me to not be passionate about the game at the same time i am tho ,i see about the not tilting thing because when people tilt they feel like a victim n blame everything but themselves which is good to not do, its just easier for u guys because u completed your journey i feel like so since u accomplished a big goal u can kinda just kick bacc n not feel bad about losing cause u already made it, but u see for me my problem i struggle with is losing a game and feeling like im behind , which in turn makes me feel like im behind on time , in time, which makes me rush my next game even after taking a break sometimes because internally i get impatient subconsciously on a deeper level cuz even when i repeat to myself in my head "chillout" i dont have all the time in the world to play league cuz i know one day ill end up having to stop playing league cuz ill just be too old , or what if something in my life happens to where i cant play league like i am now and what if i dont get to where i wanna be before that time might come, so it makes my gameplay become rushed at times and it shows in my performance, to disconnect myself from my emotions is hard for someone like me , imma fearless passionate person irl , imma man of action , i dont just talk that talk if i cant bacc it up , i demonstrate n show in conjunction with my words and can accomplish things off shear willpower alone , not being able to "just do" makes me impatient with my journey cuz i feel like time i running out cuz surely all things must come to an end at some point in time
Problem with the data set is that there is no control group - It creates a fundamentally dishonest argument. The author looked for trends and completed a write up this comes from not understanding / respecting statistics.
Hey Coach! I’ve been seriously considering the midlane academy but I am a Zed OTP, do you really think it’d be worth while for me to pay for coaching as a Zed OTP or should I pick up a few other champs and then join?
This is just a personal thing, I can't ever play 7 matches in a row without resting. Even playing 3 in a row is a bit much, I always take at least a few minutes between matches. Specially if I lose, that's just me, though.
Even if we were to take the findings at value without accounting for the first game bias, its less than 1% difference in win rate, actually quite tiny. If we were to say, "play 6 games in a row and increase youe chance of winning those games by less than 1% each" i dont see people being enthused by that.
Typical case of a graph making it look significant when its not.
This article makes a too strong relation between statistics and probabilities. Statistics is an area of mathematics which aims at analysing data, probability an area concerned with likelihood of outcomes. The statistics showing a 49.6% winrate after a loss does not mean you have a 49.6% chance of winning your next game after a loss, what you can say is that there likely is some factor that produces such a distribution in the data. What I mean by this is that the statisical value of 49.6% wr does not imply a likelihood of 49.6% for that outcome, but that something or somethings are perturbing the data away from a clean 50/50, which manifests as this wr in the data, but where the underlaying cause is unknown.
I thought it's really bad these days to hurt your own tempo to cover early waves now due to the lack of gold and experience compared --- am I wrong, Nathan? I remember we used to have no problem with covering early waves if someone needed to base or had just died in the past, but the balancing changed and now it feels like you're sacrificing yourself too much for your teammate to get a little/to dampen their losses a little(they're most likely still going to be weaker than their lane opponent no matter what at this point so why make yourself weaker than the enemy jungler and give them free reign to steal camps on the other side of the map too?) Isn't it better to ensure you hit your item spike and/or instead of sacrificing tempo to continue to play around a lane that is losing so quickly, to sacrifice that tempo to swap what lane you're playing around? Isn't this what you reference when you say adapting?
I mean, to be honest, he does address his external and internal validities, sort of. However, these margins are so small, that its hard to claim anything from the results. He doesn't cite his sample size, p values, effect sizes or anything. I get that its just raw numbers, but there's so many confounds to someone playing a game that its hard to discern how the numbers actually matter. What is their emotional state, how much food and water have they had, what is their mouse/keyboard history. It was fun, but its nothing you can really draw a conclusion from.
They said that these are people who are likely to use software to analyze games, right? If you’re only going over the first two minutes of your game while you are in for your next game better than not looking at anything, but if you have a 10 minute gap in between games, then wouldn’t that mean that you are possibly looking at the clips from your last game and trying to really understand what happened during those first few bad plays? Watching your death and trying to figure out where the Ganger came from and why you might not have noticed on time. Having a 10 minute gap between games and that being bad seems very strange because 10 minutes is not very long.
Please make a special ADC episode! 😊😊😊
Is this data from people who 3 block btw? I imagine the data would look different for people who review their games because it's mentally taxing to review a game. If people aren't reviewing then 6~7 games can make sense since ur suffering from less mental fatigue at ur 4th game with no review compared to 3 games with review in all likelihood.
And here we see why I have a job: People don't understand statistics.
I play in 2:1 ratio so if I lose 2 games and have 2 wins it's time to hang it up.
I play league until i lost 2 consecutivr games then im done for the day only during the week. If during the weenend, ill just take a break and play other games.
Tilt Teemoing is one hell of an attention grabber.
They already improved the bots btw. They're better than they used to be. Not quite player level, but definitely better than they used to be.
The graph is sampled from 5 million people it says it on the graph you most likely missed it
It's very big sample of 100000 players and 5 million games
If you win your first few games in a session, you are more likely to keep playing
About the "optimal number of games" it's likely a highly personal statistic.
zero chance these guys don’t know how to pronounce guillermo and then called it italian literally after reading he’s from spain…?
@@smokelingers i am humbled to be the one to introduce you to the concept of a joke
the idea of taking a break doesn’t include reviewing. I think the benefits of reviewing bwteeen blocks is not represented at all here
Curtis really wants to interpret the data in a way that confirm his apriori conclusions, instead of having a curious mindset. A bit disappointing, but I do think 3-block process is a good tool despite the data
To be honest, most of the data points only have a difference of 0.1% which doesn't feel very significative.
what happend to all the talk about going to Korea? It's not early season anymore... 😅
Big talk, small factory.
As all these mathematical analisis' coming out of the league subs, this one too has a glaring oversight.
The oversight being that momentary winrate is a lot less indicative of quality gameplay than assumed. League has this habit of making you win when you shouldn't, and making you lose when you play good, and lots of skills are completely invisible to any form of stats. Improvement is a very lenghty process and your learnings in some areas might not show for days, weeks or months.
Anecdotal counterpoint: at the end of the split (before blackfire), I got 350 LP in high dia on karthus bot in 12 hours grinding. How does that fit into the data?
I won a lot because i grinded and cheesed at a time where people stopped caring, and I got lucky. It's not because I suddenly became a much better player or because 12h is the optimal session size. And I certainly didnt learn anything.
You would be surprised how much worse you can play compared to your peak performance, and still have similar winrate. You can be actively improving while on a 30% winrate, and you can play with 53% for 500 games and still learn nothing.
lol
New BBC lets goooo
😵💫
@@soul0172 lol
Kind of a lot of speculating on the data here to fit your narratives...
I'm a data analyst, my mom is a data architect and we both concluded that what Curtis said is actually correct. Obviously 100% of the sample size are a part of the "first game" category, but I guarantee that most people in that "control group" don't get to the 6 game mark, let alone the 12+ game mark. It matters a lot how many of the "6 games" people make up that winrate, because if it's a very small percentage of the people used for the study, it's not even close to a valid result. If most of the studied people don't make it to that point, then the result is obviously biased towards the people that can afford to play 6 games or more, or don't lose their minds in a loss streak when they get to their 6th game
Yessssss
fit data to your narrative, classic bbc
Aye early as f bois
You need to get a life if you're playing 6+ games of league a day.