im a nonbinary-ish trans woman with kind of an androgynous voice. im not going to waste my precious time tracking down every bozo who misgenders me, but if you care, i use she/they pronouns and prefer feminine-coded language where applicable (and also thank you ❤)
this is something that i more or less figured out intuitively after playing around 30 hours and it's nice to see the math behind it to confirm what was previously just a vibe
I've had a lot of fun using Abandoned Deck to make straight builds, but now that I'm all the way up to orange stake it seems impossible to build a consistently performing deck unless I go all-in on flushes or get a particularly juicy joker. Seeing this puts a lot of my own personal experiences with the game into perspective. Excellent video!
I agree with most of the points made in the video, flushes are the most consistent strategy for lower stakes, but after purple stake flush Strat falls off. once you get to gold stake its hard to get into the late game with enough strength to beat the final ante. Its just with other hand types there is far more potential to beat it rather then dyeing just before the finish line. By all means flush builds still are a viable strategy for gold ante but far from the best.
I wasn't going to subscribe until you promised more autistic game math. I'm all for that as an autistic queer mathematician myself. Lovely vid, thanks for making it.
Another thing to point out is over-flushes have the highest play around potential. For instance, if i have 6 cards to a flush, and lets say a 9 and a J, and i have either a 9 or a jack in my flush, I can play everything in the flush but the 9 or J to then create a pretty high odds of hitting a 3 pair or a fullhouse. You can also play everythin but say a 9 when holding an 8, to increase your odds of hitting a straight on the next hand. However, Fullhouses, and to a smaller extend straights, do not have this advantage nearly as often. That's why flushes are the best play around for a 5 suited hand, in the game, especially early game.
oh, this is really good info to know! i've been having a lot of fun with balatro and it's good to know about Flushes Being Optimal for future play also: there's one of those rare mult jokers for flushes, but in order to unlock it you have to beat a run without playing any flushes. i used Abandoned Deck for my unlock attempts and successful unlock -- i'm not sure of the MATH behind it, but i know it felt like i drew into a lot more straights (and later on, pairs/full houses/four-of-a-kinds) because of the lower card count in the deck, making it easy to focus on a hand type that wouldn't result in me defaulting to playing flushes.
Great video! One thing tho - the scaling factor for flush is lower compared to straights and full houses (even three of a kind scales faster!). I agree that flush is the most consistent and reliable hand in the early game (before ante 5), but i would anticipate the ECV of flushes to drop after a few planet cards. But anyways, great analysis and great effort! super clear reasoning and presentation ❤
great video -- only thing i could add for further discussion is more emphasis on what you alluded to with 5 of a kind hands. you do usually need to do better than mere flushes and going to straight flushes is a difficult task. however, going from 3-of-a-kind/full house build to four- and five-of-a kind is a little more straightforward. Now, this isn't necessary usually for ante 8, but of course, the flush five is the best hand in the game and it is half flush -- but that's the simpler half. Building a flush five deck is a lot easier when you have the five-of-a-kind half already done. But yes, at ante 1, flushes are absolutely the way to go. After that you really have to play the cards that are dealt to you IMHO.
agreed, i focused a lot on ante 1 in this video and i'm seeing that a lot of ppl are interested in cracking the higher antes --- will be covering that in the near future :)
Sure, but this is not taking into account all the jokers, and sure, that would become pretty difficult. But in my gold stake games going for four of a kind, three of a kind, or even pair is often very valuable. They have different strengths, for example, pairs are very easy to get and upgrade with burnt joker. Four of a kind has the highest mult card in the game, and 4 of a kind is hard to get for sure, but the spade spectral card helps. She makes a good point, and the math is probably pretty solid, but I just can't completely agree with this.
im trying the last stake all the time and i find myself that sstaring with flushes and trying to transition to flush houses works best for me in the late game, ofc you need some extra copies, but the extra value flush houses give is usually worth it
It would be interesting to see how the scaling of the poker hand levels correlates changes the EV. Omg just realized I could compute a derivative for it
love the video! I have one tidbit that seems wrong to me, although you might have just misspoken At 2:32 you mention that the partial hand for a straight is "four cards in a row" -- this is true for an open-ended straight draw, but not for a "gutshot" straight draw in which you have two sides of a possible straight, for example 2 3 4 6 (where you're drawing for a 5) Again, possible that you accounted for that and just misspoke in the video, but if not then I think the partial-draw probability for straights is underrepresented in your table I actually did a very similar monte carlo simulation a couple days ago to get a rough idea of how good each hand is at the very beginning of the game when you have no upgrades, and my (rough) expected value for straights were quite a bit higher than for flushes, although to be fair there were some limitations to my approach. If you're curious, I can share the results/code (I don't want to paste a link here in case I'm marked as spam :P) Anyways really good video, love to see affirmation of my instinct to go flush build every time
you're correct! i discounted gutshot straights on the logic that most of the time, ppl will go for a different hand if they see a gutshot vs. something else. in retrospect that does mean i undercounted for straights, but i think i undercounted for all the hands similarly so it evens out :P
I feel like most people will intuitively figure it out , i realised quite fast that flushes are just easier to get , and that i hated straights . Also it's easy to build a run around and a lot of jokers buff flushew so you are likely gonna feel encouraged to build for em anyway
I generally prefer to go for full house in the early game, because there are A LOT of jokers that give you bonuses for pairs, two pairs and threes of a kind (full houses contain all three).
I naturally found myself only go for flushes the more I played this game, but it makes sense now that you showed the math. I knew some of the math but it’s nice to see precise numbers to what I was thinking. Fun fact I also would go for full house as a 2nd option and pairs/two pair as a third. 3 of kind is the only hand I haven’t experimented with yet. But it seems to be decent on your list.
I was winning with mostly 5 of a kind builds until I reached higher stakes, but once I was in gold stake, I tended to lean into flush most of the time. 5 of a kind works sometimes if I get enough copy/delete effects, but flush is definitely the most consistent.
@@autumnchiualso remove the annoying problem of playing a "flush" and then it only scores as a pair and you squint your eyes and realize one of those spades is a club
@@verbfrombonsai8852turn on high contrast cards in the settings. It turns clubs blue and diamonds orange, so all the cards are their own unique color. Makes them very easy to separate
There were about 3 of 4 places in this video I would have given an additional like. What a great and insightful video! Crazy how a simple seeming game can draw out some complicated math!
Very helpful comment. I had a feeling that flushes were easier. Glad to hear the explanation. My first two wins were with flushes. My third win was with a deck where I had like 18 Kings, making three, four, and five of a kind and full houses, with some awesome Jokers
While I do agree with your conclusion, there are some other reasonable "partial straights" which are four cards within a run of 5 (2-3-5-6). This is harder to actually complete than a 4-card run where you can get either end, however. I don't think straights are quite as dire as the initial chart shows.
I feel so validated, it really felt like flushes were broken with discarding, coming from a real poker background I was a little confused why. Thanks for running the numbers for us!
I prefer Full House since it's more consistent vs suit bosses. But 100% agree flushes are easily one of the more consistent decks that can pair really easily with suit jokers.
Making flushes is the easiest and most consistent hand build to make in most decks, finding the jokers and tarot cards to provide optimal synergy, not so much
Thank you for explaining that when I constantly fail to get the fifth card I need for a flush, I am hitting a 38.5%. I am god's least favorite gambler. 😊
For odds of completion are you calculating based on one redraw or one multiple? While flush could be more attainable after one redraw, a full house could become more attainable given more opportunities to discard
i did this indirectly in the second half of the video! the math is for cards discarded and drawn as opposed to hand size, but i think you still get the results you're interested in
great question! the code for a full straight flush checked for *at least* 5 cards in a row, but the partial is defined as *exactly* 4 cards in a row. so any 6, 7, or 8 card SF is included in the former case
thanks for doing the math! I've been playing flushes early because i just sorta intuitioned that it was highest value, so i appreciate the validation with the math! im way too lazy to do it myself lmao but i was curious how the table would look if we counted 3 of a suit as a partial flush as well, considering we can actually discard 5 cards at a time with the starting deck?
i didn't include the numbers in the table, but the odds of a runner-runner out situation (drawing two cards) are generally very low. my quick estimate is: 9C2 suited cards * 42C2 other cards * 6 orientations = 9 * 8 * 42 * 41 * 6 combos with at least 2 suited cards 44C4 = 44 * 43 * 42 * 41 total combos eliminate to get 6 * 8 * 9 / (44 * 43) = 6 * 2 * 9 / (11 * 43) = 108 / 474 = ~22% now that I've done the math, it's higher than i thought... interesting!
the odds of pair on first draw includes all cases of two pair, three of a kind, etc. so it's a much bigger number. for the odds of a partial, you have to specifically exclude the chances of getting the hand immediately, so the number is smaller~ lmk if it's still confusing and i can make a diagram
@@combokings8346 high cards and pairs are easier to get and can become very high yield if you level them up them as well, you get more held in hand effects and you get to discard a couple extra cards when you play a hand; builds that play around pairs and high cards are therefore generally better.
there is way more math than the one part you're talking about. if you stop worrying about trying to be smarter than some fuck on the internet, you might learn something useful
while i used ChatGPT to help me write the code, all the numbers were generated using Python scripts verified by my actual programmer brain. i spent many many hours cross checking the results with different methods, and it all led to these results.
*Too bad you find this entire thing now suspect and are unwilling to use it. The answer was the video creator's and they mentioned that they put in reasonable effort to verify the information they gleaned from GPT was accurate. Many people likely can and will use the information this video provides. You're simply choosing not to be one of them.
im a nonbinary-ish trans woman with kind of an androgynous voice. im not going to waste my precious time tracking down every bozo who misgenders me, but if you care, i use she/they pronouns and prefer feminine-coded language where applicable (and also thank you ❤)
As a nonbinary-ish trans woman I feel this lol Also thanks for the great and informative video! ❤❤❤❤❤
You do you but your voice is not really androgynous
@@nevyteooleave me alone :(
Bro all respect here but its a bro / you angle for me (bro being a genderless word here)
@@autumnchiu felt that
Now I have a mathematical proof that I'm playing the optimal play not just "monkey brain same color click together" which I'm definitely not doing
this is something that i more or less figured out intuitively after playing around 30 hours and it's nice to see the math behind it to confirm what was previously just a vibe
my favorite part of doing math :))
SAME I just went for a ton of flushes 2 hours in
Same, but at 5 minutes :D
nl was cooking all this time. And we were Neo wondering why our penis hurt.
I've had a lot of fun using Abandoned Deck to make straight builds, but now that I'm all the way up to orange stake it seems impossible to build a consistently performing deck unless I go all-in on flushes or get a particularly juicy joker. Seeing this puts a lot of my own personal experiences with the game into perspective. Excellent video!
"$20 is $20" 😂
I agree with most of the points made in the video, flushes are the most consistent strategy for lower stakes, but after purple stake flush Strat falls off. once you get to gold stake its hard to get into the late game with enough strength to beat the final ante. Its just with other hand types there is far more potential to beat it rather then dyeing just before the finish line. By all means flush builds still are a viable strategy for gold ante but far from the best.
I'm planning to release a followup video for gold stake specifically!
@@autumnchiudid you ever get around to it! im excited to see more!
@@jadeamulet2339 getting there! i was cooking for a bit but then the new gold stake dropped and now i have to relearn it 😅
girl, i wasn't expecting such an engaging yet informative video. you have a new subscriber.
wow thank you anonymous stranger
I wasn't going to subscribe until you promised more autistic game math. I'm all for that as an autistic queer mathematician myself. Lovely vid, thanks for making it.
Another thing to point out is over-flushes have the highest play around potential. For instance, if i have 6 cards to a flush, and lets say a 9 and a J, and i have either a 9 or a jack in my flush, I can play everything in the flush but the 9 or J to then create a pretty high odds of hitting a 3 pair or a fullhouse. You can also play everythin but say a 9 when holding an 8, to increase your odds of hitting a straight on the next hand. However, Fullhouses, and to a smaller extend straights, do not have this advantage nearly as often. That's why flushes are the best play around for a 5 suited hand, in the game, especially early game.
oh, this is really good info to know! i've been having a lot of fun with balatro and it's good to know about Flushes Being Optimal for future play
also: there's one of those rare mult jokers for flushes, but in order to unlock it you have to beat a run without playing any flushes. i used Abandoned Deck for my unlock attempts and successful unlock -- i'm not sure of the MATH behind it, but i know it felt like i drew into a lot more straights (and later on, pairs/full houses/four-of-a-kinds) because of the lower card count in the deck, making it easy to focus on a hand type that wouldn't result in me defaulting to playing flushes.
That sequence of flipping through each partial hand at 2:30 was really cool, very well done.
Great video! One thing tho - the scaling factor for flush is lower compared to straights and full houses (even three of a kind scales faster!). I agree that flush is the most consistent and reliable hand in the early game (before ante 5), but i would anticipate the ECV of flushes to drop after a few planet cards. But anyways, great analysis and great effort! super clear reasoning and presentation ❤
great video -- only thing i could add for further discussion is more emphasis on what you alluded to with 5 of a kind hands. you do usually need to do better than mere flushes and going to straight flushes is a difficult task. however, going from 3-of-a-kind/full house build to four- and five-of-a kind is a little more straightforward.
Now, this isn't necessary usually for ante 8, but of course, the flush five is the best hand in the game and it is half flush -- but that's the simpler half. Building a flush five deck is a lot easier when you have the five-of-a-kind half already done.
But yes, at ante 1, flushes are absolutely the way to go. After that you really have to play the cards that are dealt to you IMHO.
agreed, i focused a lot on ante 1 in this video and i'm seeing that a lot of ppl are interested in cracking the higher antes --- will be covering that in the near future :)
i found this out when i started playing balatro and sorted by suit with flush related jokers and got way past ante 9 on my first go.
This is a certified Checkered Deck moment, thank you very much for validating my strategies
Excellent and very informative video. Flush always felt best to me but its so cool to see actual hard facts to back it up. Thank you 😊
Sure, but this is not taking into account all the jokers, and sure, that would become pretty difficult.
But in my gold stake games going for four of a kind, three of a kind, or even pair is often very valuable. They have different strengths, for example, pairs are very easy to get and upgrade with burnt joker. Four of a kind has the highest mult card in the game, and 4 of a kind is hard to get for sure, but the spade spectral card helps.
She makes a good point, and the math is probably pretty solid, but I just can't completely agree with this.
See, this is why you should pivot from flushes to straight flushes
damn shoulda thought of that
im trying the last stake all the time and i find myself that sstaring with flushes and trying to transition to flush houses works best for me in the late game, ofc you need some extra copies, but the extra value flush houses give is usually worth it
and trying to force color changin tarot for consistency
Thank you Balatro for introducing so many people to the wonderful world of combinatorics
It would be interesting to see how the scaling of the poker hand levels correlates changes the EV. Omg just realized I could compute a derivative for it
Me sitting here thinking this was just a balatro video, before laughing hard at "$20 is $20". Great vid!
i got some zingers in me
love the video! I have one tidbit that seems wrong to me, although you might have just misspoken
At 2:32 you mention that the partial hand for a straight is "four cards in a row" -- this is true for an open-ended straight draw, but not for a "gutshot" straight draw in which you have two sides of a possible straight, for example 2 3 4 6 (where you're drawing for a 5)
Again, possible that you accounted for that and just misspoke in the video, but if not then I think the partial-draw probability for straights is underrepresented in your table
I actually did a very similar monte carlo simulation a couple days ago to get a rough idea of how good each hand is at the very beginning of the game when you have no upgrades, and my (rough) expected value for straights were quite a bit higher than for flushes, although to be fair there were some limitations to my approach. If you're curious, I can share the results/code (I don't want to paste a link here in case I'm marked as spam :P)
Anyways really good video, love to see affirmation of my instinct to go flush build every time
you're correct! i discounted gutshot straights on the logic that most of the time, ppl will go for a different hand if they see a gutshot vs. something else. in retrospect that does mean i undercounted for straights, but i think i undercounted for all the hands similarly so it evens out :P
I feel like most people will intuitively figure it out , i realised quite fast that flushes are just easier to get , and that i hated straights . Also it's easy to build a run around and a lot of jokers buff flushew so you are likely gonna feel encouraged to build for em anyway
Going for flush seems to be the most intuitive way to play. Glad that there os math to back it up.
I generally prefer to go for full house in the early game, because there are A LOT of jokers that give you bonuses for pairs, two pairs and threes of a kind (full houses contain all three).
I naturally found myself only go for flushes the more I played this game, but it makes sense now that you showed the math. I knew some of the math but it’s nice to see precise numbers to what I was thinking. Fun fact I also would go for full house as a 2nd option and pairs/two pair as a third.
3 of kind is the only hand I haven’t experimented with yet.
But it seems to be decent on your list.
makes sense. There are only 4 suits, all else being equal each has a 1/4 chance of being drawn
I was winning with mostly 5 of a kind builds until I reached higher stakes, but once I was in gold stake, I tended to lean into flush most of the time. 5 of a kind works sometimes if I get enough copy/delete effects, but flush is definitely the most consistent.
This is why the checkered deck is my favorite
you get me
@@autumnchiualso remove the annoying problem of playing a "flush" and then it only scores as a pair and you squint your eyes and realize one of those spades is a club
@@verbfrombonsai8852turn on high contrast cards in the settings. It turns clubs blue and diamonds orange, so all the cards are their own unique color. Makes them very easy to separate
@@verbfrombonsai8852turn on hight contrast cards please. i literally saved my life cause im blind as hell
@@verbfrombonsai8852there’s high contrast cards in the settings! it makes the clubs light blue, much easier to see
There were about 3 of 4 places in this video I would have given an additional like. What a great and insightful video! Crazy how a simple seeming game can draw out some complicated math!
Very helpful comment. I had a feeling that flushes were easier. Glad to hear the explanation. My first two wins were with flushes. My third win was with a deck where I had like 18 Kings, making three, four, and five of a kind and full houses, with some awesome Jokers
very useful. thank you! pogging out of my gourd.
thank you! please don't explain to me what that means
While I do agree with your conclusion, there are some other reasonable "partial straights" which are four cards within a run of 5 (2-3-5-6). This is harder to actually complete than a 4-card run where you can get either end, however. I don't think straights are quite as dire as the initial chart shows.
I feel so validated, it really felt like flushes were broken with discarding, coming from a real poker background I was a little confused why. Thanks for running the numbers for us!
They fr tryna make “anomanously” a thing 🔥🔥🔥
Math proof that checkered deck is the best deck! I will still go for full house in the first blind because it's easier to oneshot...
Best video!!! For balatro! You’re very underrated the hard work shows!
very well edited, interesting video, hope you pop off king
thanks, me too!
Winning without using a flush was the most annoying.
It's the bread and butter of my runs.
No one tell Northern Lion
i noticed i was able to get a ton of flushes in balatro- so i was wondering abt this thanks
theydidthemath
they did the monster math
I prefer Full House since it's more consistent vs suit bosses.
But 100% agree flushes are easily one of the more consistent decks that can pair really easily with suit jokers.
something extra to factor in is that Saturn is much stronger than Jupiter
I'm waiting for - in a couple months - someone makes a generational learning AI that can calculate all of this to play optimally.
The easiest strat is to let the jokers do all the points and multy, an then proceed to just throw random shit with the cards xd
Making flushes is the easiest and most consistent hand build to make in most decks, finding the jokers and tarot cards to provide optimal synergy, not so much
based theory crafter
Thank you for explaining that when I constantly fail to get the fifth card I need for a flush, I am hitting a 38.5%. I am god's least favorite gambler. 😊
he gives you his strongest battles
For odds of completion are you calculating based on one redraw or one multiple? While flush could be more attainable after one redraw, a full house could become more attainable given more opportunities to discard
Great video! Would you be so kind to please make a table for a lower/higher hand size? Like how does hand size affect the odds?
i did this indirectly in the second half of the video! the math is for cards discarded and drawn as opposed to hand size, but i think you still get the results you're interested in
Why is straight flush partial lower than straight flush first draw?
great question! the code for a full straight flush checked for *at least* 5 cards in a row, but the partial is defined as *exactly* 4 cards in a row. so any 6, 7, or 8 card SF is included in the former case
Also i think flush is way easier to get then straight
thanks for doing the math! I've been playing flushes early because i just sorta intuitioned that it was highest value, so i appreciate the validation with the math! im way too lazy to do it myself lmao
but i was curious how the table would look if we counted 3 of a suit as a partial flush as well, considering we can actually discard 5 cards at a time with the starting deck?
i didn't include the numbers in the table, but the odds of a runner-runner out situation (drawing two cards) are generally very low. my quick estimate is:
9C2 suited cards * 42C2 other cards * 6 orientations = 9 * 8 * 42 * 41 * 6 combos with at least 2 suited cards
44C4 = 44 * 43 * 42 * 41 total combos
eliminate to get 6 * 8 * 9 / (44 * 43)
= 6 * 2 * 9 / (11 * 43)
= 108 / 474
= ~22%
now that I've done the math, it's higher than i thought... interesting!
I FUCKING LOVE GAMBLING!!!
I liked the statistical analysis but you forgot how important what jokers are good with what hands and how much better that also makes flushes.
And chat called Egg a madman.
We all know high card is the true best hand in this game.
when you go sicko mode with those steel cards? nothing better
why are odds of a pair on first draw, and odds of partial two pair or partial three of a kind all different? wouldn’t they all by the same hand?
the odds of pair on first draw includes all cases of two pair, three of a kind, etc. so it's a much bigger number. for the odds of a partial, you have to specifically exclude the chances of getting the hand immediately, so the number is smaller~ lmk if it's still confusing and i can make a diagram
Wow, some actually transformative use of AI
Awesome video. Awesome nails.
My dog is better now, thx
glad to hear it
I fucking knew it
Ah yes my favorite part in videogames, math.
And i dont even joking i really like it xD
you forgot like 3, in the real world you forgot royal flush and in the game there is flush five and flush house at 0:35
- royal flush is a straight flush
- the odds of getting a flush five or flush house in the starting deck are exactly 0.00%
- you forgot five of a kind
This is a great breakdown
your videos are phenomenal, thank you so much
Flushes aren't even close to being the best hand type in the game. Pairs are better than it in almost every way possible, followed by high card.
Better in terms of probability of getting a pair or high card or better in terms of gaining chips?
@@combokings8346 high cards and pairs are easier to get and can become very high yield if you level them up them as well, you get more held in hand effects and you get to discard a couple extra cards when you play a hand; builds that play around pairs and high cards are therefore generally better.
Nah, high card all the way. What's better than 100% chances?
super interesting video
good video but Poker is a much better name for four of a kind
i did not make the game
nice work!
thanks!!
Great video!!
Dawg not when I'm waiting for my cat to get an operation
I hope your cat recovers :((
awesome video
W
Comment deleter... Heaven forbid someone bring up that different priorities require different heuristics
i genuinely have no idea what this comment means
I figured that out literally 10 minutes into the game.
thanks for telling us
there is way more math than the one part you're talking about. if you stop worrying about trying to be smarter than some fuck on the internet, you might learn something useful
@@autumnchiu I don't actually worry whether I am smarter.
I can't trust any GPT answer,
Too bad this entire thing is now suspect and can't be used
while i used ChatGPT to help me write the code, all the numbers were generated using Python scripts verified by my actual programmer brain. i spent many many hours cross checking the results with different methods, and it all led to these results.
*Too bad you find this entire thing now suspect and are unwilling to use it.
The answer was the video creator's and they mentioned that they put in reasonable effort to verify the information they gleaned from GPT was accurate. Many people likely can and will use the information this video provides. You're simply choosing not to be one of them.
Nail polish 🤢🤢🤮🤮🤮🤮
tf? grow up