Poker Theory: One Minus Alpha 

This channel doesn’t get enough traction, but it’s literally the best poker content out there.

This is amazing Phil, coaches and content creators just hover over these basic theory concepts and just assume people know them ( and most people who think they know them, don't ) so thank you for taking the time and making this

Summary: Against an unbalanced opponent, there are two cases: (a) If your opponent doesn't bluff this node enough, fold every bluffcatcher. (b) If your opponent bluffs this node too often, call every bluffcatcher. If your opponent is fairly balanced, there are also two cases: (c) If there's lots of overlap between your range and your opponent's range, fold the bottom alpha of your range and call the top 1-alpha of your range. (d) If there's minimal overlap between your range and your opponent's range, fold alpha worth of hands with bad blockers and call 1-alpha worth of hands with good blockers. Finally, if your range is fairly condensed and you can't easily approximate alpha and 1 - alpha simply by choosing hands to call with and hands to fold, use some randomization to decide what to do.

Phil Galfond is insanely good at understanding things and explaining them, a rare quality...Thanks

I was aware of this poker toy game concept but Phil is the first one to explain it in a way a person with a double digit IQ could understand it, which is great because I finally got it

Thanks Phil, looking forward to more of your vids! You're a very chill guy and I feel your videos are very approachable and calm.

For what we know,solvers tend to get closer and closer to mdf as we get to the river when deciding whether or not to call a bet.In fact,if our opponent is playing balanced on the river,our bluffcatchers have in most cases 0 EV,so we are indifferent between callilng or folding them,but we call with those that have some sort of blockers,cause having them will infict a kind of unbalance in your opponent's bluff/value ratio thus making your play ev+. In the toy game,we are indifferent between calling or folding all the time with our K,the problem with it is that our opponent can adapt and changue his game strategy.In poker,if we overcall on the river our opponent will start valuebeting hands that he shouldn't ,NOW making them profitable. The nice part about mdf is that it reduces the risk of our play,if you don't know if your opponent is unbalanced or not.If villain overbluffs,we win less than we would have won if we bluffcatch with any bluffcatchers,but if our opponent underbluffs,we lose less than if we would have lose if we called with any bluffcatcher.

If only we could send YT videos back in time :) So cool this is just out here for players to get their fundamentals in such a well explained and concise way, keep 'em coming !

Enjoying your new content Phil. Any plans to do a video covering online hand reading/applying ranges?

Great episode. Love the education and story content. Thanks for sharing

Good video. Would be nice if you touch upon what happens when you allow a mixture of different bet sizes and why that does not improve EV in this case over single size pot bet.

I'm new to poker but the understanding I came to of GTO was this. "GTO" is a strategy which is unexploitable/unbeatable by the metric of statistics/probabilities. If we play rock paper scissors, and I employ the strategy of playing each option 33.3% of the time randomly, you cannot come up with a strategy that has a statistical edge over me. Conversely however, in this particular case, the GTO strategy will have no statistical edge over my opponent. Define GTO strategy as a strategy for which no one can gain a statistical edge over you, and also maybe add the caveat that in some cases (RPS is not one of those cases) a GTO strategy will beat every other strategy with the exception of an opponent who copies your strategy. GTO strategies in some example games I've seen seem to mean that your opponent has to copy your strategy or they lose (depends on the specific game being played whether GTO beats all other strategies or whether it breaks even). In games where GTO beats all non-GTO strategies (e.g. a poker game with no blinds at all, GTO is to fold all hands except pocket aces) GTO is truly the ultimate way to play. The only thing your opponent can do is copy you, or they lose. But whether GTO breaks even or is winning in poker, in the realm of pure theory, that I do not know. If you do know I would be interested to know. I am leaning towards the understanding that perfect GTO breaks even. Also, a statistical or probabilistic edge says nothing about what will actually occur. Probability exists because people don't know what will occur. So having an edge doesn't guarantee you'll win, it's just the best thing you can do most of the time. Anyone who's played enough poker knows they get destroyed if they play with bad odds and equity. This is a tangent though, main thing is idk if GTO breaks even or wins against non-GTO strategies, seems to me like it breaks even. You want to be unpredictable and unreadable by your opponent as well as knowing their habits, so you play the non-GTO way that exploits them. Play in a way which is a bit exploitable, so long as you exploit other people more, it becomes okay. We're playing against humans, not robots

probably one of my fav vids of yours. could very well have a number of titles.

Love your channel. Thanks for the great explanation.

I've heard of minimum defense frequency which is what is described here. It is a good concept to understand, but it is important not to use it as a crutch where you forget to hand read.

How was I just made aware you had a channel? I hope many more susbribers come your way as your knowledge and understanding of the game is nearly unmatched.

Great explanation. Thanks Phil! I played two 7 hour, totally spot and card dead sessions this past weekend. Instead of 1-alpha, I was thinking 1-WTF!?? LOL. Thanks for what you do!

Very well explained and simplified

This is an interesting concept thank you for the video.

I love this video, it also makes me think that Phil could do videos explaining economics concepts.

Love it! Thanks!

Great video. All the best Phil

I've heard "1-aplha" called "minimum defense frequency" which is a good name imo. But there may be some difference that I'm not aware of.

The last point you made Phil is so spot on lol. I remember when i first heard about MDF back in 2013-2014 i rushed to explain it to some poker friends of mine at the time. I thought i found the golden chalice and i acted like that too...None of my friends agreed with its significance and i thought they were idiots well I couldn't have been further from the truth lol. My friends were correct all along and i was wrong, apparently MDF is not the golden chalice of poker. It did end up making me ultra curious about poker and this in turn helped me become a successful professional poker player but its effectiveness is just overhyped noone can actually count down combos to the necessary precision other than an GTO bot, so one could say its just a myth as the golden chalice always were too... :)

Amazing!! 🤩

Dope vids!

Amazing!

After watching a few Weeks of your videos I found myself playing to many hands yesterday and I lost most of my money getting valued by other players . I never lost a big pot just I lost all the small ones and medium ones.

Mr. Philliam, as someone that has watched you play and watched your training videos for a decade or so, I'm just a little confused by a concept explained here. What is the "folding" that you mentioned that you can do on the river. I was not aware that that was an option.

The dulcet tones of Phil Galfond, poker therapist

"you need to have 2 value bets for every bluff" alright im with you this far, that makes perfect sense. But the point where this gets confusing is even you actually play the game, like, i cant force two value bets for every time i want/have to bluff to win a pot. The theory is perfect in a vacuum heads up scenario but 8 handed with 45bb (or however many) this becomes very confusing to me. Like i cant literally just value bet 2/3 hands and bluff one, i dont know how to make this perfectly crystal clear, but how do i keep track? Even if i keep track which i could, is it for ever opponent or for total hands played? This is where poker gets hard, Yeah range vs range and all of that, but then you have actual poker, it seems different from theory to say the least. That being said i enjoyed the video

One thing is for certain: Basically Tony G’s live cash game playing style from 2010-2011 Pokerstars Big Game seasons (all available free on RU-vid and his general overall livecash game play from the past was always more aligned with GTO than anyones obviously Your boy Durrr too, however Tony was perceived as maniac aggro loose (donkish) BUT he crushed those games $$$ he lost some big pots but he won MUCH more than he ever lost he was a consistent winner, and even to this day when he played in Triton Kings Cash Games he played GTO it was much more difficult to put him on a hand,because he could have anything from any position and people can’t range him postflop as accurately as his ranges were so wide (except one loose cannon who had studied tony g before he came on the show) he would steal so many pots post flop with air middle pair bottom pair. But fundamentally he was playing closest to GTO at the time, perhaps the only part of his game that lacked true gto was he didn’t have enough spots where he would check/call nutted hands post flop all the way down to the river he didn’t take passive approaches and lines but back THEN no one did because at that time you lose too much value trying to play passively: gto puts you in long run+ev but it sometimes means foregoing higher expected payoff’s in the short run where you once took value maximising lines with nutted hands- but since more players acquired knowledge information on the game and that has changed the game even down to mid to lower stakes those standard value maximising lines are/were too predictable for pros vs pros and even now amateurs mid stakes players because players don’t have equal combo’s of bluffs and it becomes easier for players to discard hands when being faced with value maximizing lines. Qualitatively The goal is to be in the middle, playing all of the tendencies sometimes, being all of the tendencies sometimes, You want to be perceived as, loose passive, loose aggressive, tight aggressive, tight passive simultaneously, so that hand reading you is impossible because you can have any two cards in any spot, ** actually you have just as many combinations of value hands as you do bluffs in every spot.

very cool vid. i'd take a top 5 of these any day. that said, i feel weird trying to implement pure theory based tactics against people who i don't believe are employing anything resembling a strategy. or maybe that's the best time? hmm.....

Feel tells me what balance is

Actually in the ace king queen game the more you adjust your frequency of bluffing with the queen the more likely they are to make mistakes.

1 - alpha = MDF

Well I'm definnitely not as good at poker as I thought I was at the start of this video lol

Phil says it means you have to defend 2/3rds of the time. Isn’t that incorrect? You need to defend the top 2/3rds of your range.

I tried your toy game example to understand it, but it doesn’t seem to add up. What am I missing? Yes, if I bluff 100% by betting $100 (into a $100 splashed pot) all 8 hands (randomly 4 Aces and 4 Queens vs your constant King), you simply call all 8 hands, and we’ll each be up $400 after. So then I bluff 0% by checking every Queen and betting every Ace, you then just fold to every bet and check back and win when I check, and we’ll still each be up $400 after. Finally I shift to the 2 to 1 value-to-bluff ratio by bluffing only every other Queen, $100 into $100 (Alpha = 1/2), and you shift by calling 1/2 the time (1 minus Alpha). You stated we've now completely solved this toy game, and achieved purely optimal strategy for both players. I really want to understand this part, because I don't see it. Across 8 hands, I’ll check twice and bet $100 6 times. You’ll call 3 times and fold 3 times. All 4 Aces I’ll bet $100. You’ll pay me off twice and fold twice. Result will be: My stack +$600 Your stack -$200 With the 4 Queens I’ll bet $100 twice and check twice. You’ll win $100 both times I check, and with my 2 bluffs you'll fold once and call once winning yourself another $200 there. The end result will be: My stack +$600 Your stack +$200 Instead of us both winning $400 after 8 splashed pots, I'm now up $600 and you're up only $200. It’s no longer an even scenario. True, by bluffing half the time and you calling half the time, the bluff spots are break-even for both of us, BUT in order to achieve that you also have to pay off my value bets half the time, and so I'm ahead in the end. And the more times we do this, the farther ahead I’ll be. This is easy enough to understand: I'm only bluffing 1/3 of the time, but you're paying me off 1/2 the time. So of course I'll end up ahead in the long run. There's probably some simple element here that I missed, but as explained I don’t see how One minus Alpha solves for indifference. In the end the aggressive bluffer always come out ahead... Thanks for your time and these interesting videos!

New poker players now love using all the cool new poker buzz words to sound smart 🤣

this should be remaked by greg goes all in

1150 views and on 96 likes???? I oughta................

Jon Snow knows his stuff.

This is the shit you're looking for.