I get by playing poker from approximating my pot equity, fold equity, and the action vs the opponent and board texture.

I was wondering if you guys are more precise. Anyone do EV calcs (weighted/unweighted) while playing? Seems it would cause much better mathematical decisions, but I'm not sure if it would be practical or even possible to do at the table.

I don't really think calculating EV is something you can reasonably do at the table...I've always been under the impression that it's more of something you do away from the table....I generally only do this when reviewing hands or situations

this ..... studying and then recognizing situations as they come up in play.

Nobody.

Bots do.

Yeah pretty much nobody. However... you run the sims long enough and stare at hands long enough, and you get pretty good at estimating.

While it is unlikely one does an actual EV computation, you are in effect doing that when when you compare pot equity or win probability to pot odds. Current pot odds is relatively easy; implied pot odds is often an educated guess.

Pot equity or win probability is harder but often doable -- e.g., an all-in call with TT vs an overpair (55/45 approx.) or valuing your nine outs for a flush before the river card as 37 to 9 against for a 20% win rate approx. etc.

So if your pot equity converted to odds is better than the pot odds, you call and that by definition is a +EV decision.

I beg to differ. I personally dont. But I am sure it is possible given you can do quick calculations(speed math)and have a great understanding of the game. I can do speed math and see how a great player can do ev calculations for simple decisions like all ins against one opponent possibly and probably omore IF you understand the game greatly(I dont I am struggling to learn and apply new concepts). I wouldnt put this past anyone especially at the higher levels and games. You can never say someone doesnt do it because its very well possible. I will learn with time how to do these and practice.

Look how easy is it to do multiplication in your head. Say you have 96*96
using 100 as a reference number. They are both below 4 and cross subtracting both ways 96-4=92. Now 92*100(reference number) =9200. Hold that number. Now remember the 4's? multiply them to get 16. Add 16+9200=9216 which is the answer. Equations like this can be done in seconds with practice. Square roots, addition, division, multiplication. Speed math also increases memory drastically.

Sq roots can be done as well. Imagine calculating the sq root of a play that you are about to make in seconds. for ex the sq root of 2809 can be done like this take 28 and seperate it from 09(first pair the digits back from the decimal). You will have a two digit answer. Estimate the sq root of the 1st digit pair. Which is approv 5*5=25(28). 5 is the first digit of the answer. Double the 1st digit of the answer 2*5=10 and write it to the left of the # which will be out divisor. Write 5, the 1st digit of our answer above the 8 in the 1st digit pair, 28.

We are done with the 1st digit of the answer. To find the 2nd digit of the answer square the first digit of your answer and subtract the answerr from your first digit pair 5*5 than 28-25=3. 3 is the reaminder and carry the 3 tot he next digit of the number being square which gives you a new working #30. Now divide 30 by 10 and get 3. The next digit for our answer. 10 divides evenly into 30 so there is no remainder to carry and 9 is our new working number. Finally cross multiply the alst digit of the answer 3^2=9. subtract 9-9=0. No remainder. 2809 is a perfect square. Its sq root is 53. This can literally be done in seconds with practice.

Since noticing that most players DONT do this. Think about the edge you would gain if you COULD do this. I am pretty positive people are doing it(higher level thinking players). If interested pick up a speed math book by bill handley. I use to do quick math in school and kids would look at me like i was a freak. And all the meanwhile, I was doing simple s*** like 3*3 hahaha. Its all aboutt he strategy you use with math. People who are better at math use better strategies

You'd have to do an insane number of calculations to figure your hand's equity vs. your opponent's range. It's not feasible nor worth it for people to work on being able to do that.

I can't calculate EV, but I try to do approximations in some simple situations. I stick to deep stack games, because you can be lazy and justify calls for "implied odds" :-)

Well if that's all you're doing you are shortchanging yourself .... because just cause the amount of $ you need for implied odds exists at the table doesn't mean you have the correct implied odds, especially at deep tables.

I agree. My last comment was a joke (well the part you're referring to anyway).

David Sklansky

EV calulcations are done away from the table in order to be able to recognize the good or bad spots when at the table fast enough to avoid at least significant errors.

I mean once you do them at home at your own pace you can be generally safe in applying certain playing styles and choices that fit those calculations.


Occasionally when playing online you may get into a critical hand and spend 2 min to decide which way during which you can practice simulating .

Example;

You have K9s on the small blind and only 4bb left. The pot is 2.25bb and then someone on middle position that you kind of respect goes all in for 10bb . They all fold. You call or fold?

In 2 min actually there is time to assume the guy is on a 10% hand ,calculate the pot if you call and if the bb also calls and see what is your equity using pokerstove for example. You have 29% 3way so you get out 3.7bb vs 3.5 if you fold. If the bb doesnt call you get 35% 2way for 3.4bb. If the all is a 5% hand you clearly get much worse results . That fast calc can take less than 1 min (if you have the application running mutiple instances in advance - eg one with a 10% range already in place waiting for your hand to put in , another with 5% etc) and help you decide to fold and get 10-20 more hands to see and possibly win even more money (prizes). Now once you do that next time you know and have an idea that clearly somewhere near 4-5 bb is the turning point for such decisions and as the quality of the original all in improves over 10% (better than i mean) or your stack is over 5bb the more inclined you should be to fold with hands of the K9s neighbourhood of strength. (even knowing how to group hands like that from simulations you have done earlier proves valuable eg A8o is like K9s in this spot etc nearby hands tend to behave the same) . You see where i am going ? You take one example and then you start toying around it with other parameters and see how it changes. That builds intuition seriously fast.

You then remembering what happened live go and study this case in general and see what kind of hands are a call in this spot and next time it happens you have a multidimensional parametric understanding of the situation (all in quality, stack size, your hand quality etc ) .

Another one could be that having AQ from utg and 10bb or less is marginally plus EV as all in . Knowing this everytime you get AQ or better nearby UTG you have a very clear understanding of the situation.

Another one can be calculating what kind of raises to call when you have small pocket pairs at sb or bb . You do the calc a few times at your own pace and then live you have a very good understanding .


So you are not repeating those calculations live but you are recalling the prior examples and trying to extrapolate around them . Having a scientific background with intuitive understanding of the math for a variety of problems and perturbations around them helps too. You have a sense of the important or not . Most of those decisions are stable and in the sense that they change slightly as you change the parameters of the problem, so knowing the rough avg of a core example helps you avoid big mistakes when nearby it. In the long run tiny mistakes of the 1% or 2% level will not matter. If a decision is marginal its also marginal +-1,2% away fom it.



Now in a situation where you have to call a huge bet and you have the luxury of a 5 min clock (some appropriate live situation i guess that they have this time available) i can imagine actually taking a notebook and doing some calcs for the hell of it even if all laugh at you . But put yourself in such tough spots 10 -20 times in your life and you no longer need to be that nitty as long as you learn from each one before. It becomes more amusing then to take out the notebook and less effective or practically important . Usually decisions that take this kind of precision are marginal anyway and the mistake either way is tiny statistically. The more examples you have done away from he table the better your ability to decide which cases are indeed marginal. If a decision is so close that it then requires a refined calculation there is no reason to even do it since you know it will prove irrelevant on a practical level., the answer is do whatever of the 2 choices, its all the same .

The value of EV calculations is that doing a lot of them at your own luxury builds immense intuition that live comes massively handy in dealing with strength and determination under a variety of situations without feeling the least puzzled or worried. And thats the beuaty of it really , its a self confidence mental boost for your overall game. I have no doubt we all do minor mistakes all the time . The probably more important and less easy to find out mistakes we do are in bet sizing and trapping or betting strong choices etc. They are mistakes in the sense they a re far from optimal choices not the kind of negative EV decisions , the kind of wasting profit decisions instead. How you bet a set post flop for example optimally is an important problem . You may know very well when to call with a flush draw in general so that its a plus EV decision (making the live calculation rather easy within 20-30 sec or faster) but the other one regarding the set which is less clear what it requires will probably prove more important for your overall profit long term.

So the conclusion i guess is we use EV calcs away from the table to better understand our game, find problems in it, improve our effectiveness when in positions of strength and generally examine out of the box ideas etc as we investigate the game from new angles. Live you cant do it at such depth and precision but it no longer matters because we have modified behavior to be such that the most hard decisions we face are already marginal (since we have fine tuned our game to be confident in all other instances) so we know when we are worried about something its less important statistically than it feels at the time due to volatility of the decision.

I'll repeat.

Every time you compare pot odds to win odds to decide to call or make a bet, you are essentially doing an EV calculatoin.

But this is a *really* simplified case because you're typically not figuring in secondary hands you might have to call with, extra bets you might make when you hit, how often you're going to bluff and how often you'll get called when you hit, etc, etc. It's like level 1 and most post-game analysis is at least like level 3-4

I agree that for many cases it is simplified and I'm all for doing post game analysis. But, whenever you have have an all-in call decision or a call decision after the river card, then pot odds vs win odds is a most useful tool, as simple as it might seem. However, even then, estimating your winning probability often does involve 2nd and 3rd level analysis, which I would think most expert players do.

I've wanted to start attempting real-time EV calculations for a while now. I don't think it's actually unfeasible to do a fairly thorough EV calculation in about 20 seconds using speed math techniques and mnemonic devices. I've been looking into the "imaginary abacus" method of speed calculation; basically, you practice doing lots of arithmetic on an abacus, maybe practicing half an hour to an hour a day, and you get really fast with the abacus. Once you get to the point where you can do an EV calculation on the abacus in under a minute, you start using some simplifying assumptions, estimations, etc. (but you're still doing a fairly thorough analysis), and you work your way up to doing the arithmetic on an imaginary abacus that you visualize and "manipulate" with your fingers. Think of how really good pianists can practice without an actual piano because they have good ears and good muscle memory.

The next part is being able to store several important numbers at a time in your short term memory. For this, I'm looking into the phonetic system, in which different digits are represented by different consonant sounds. You use the consonant sounds that correspond to a number you want to store by creating a memorable phrase using those consonant sounds in order. So, for instance, say that 4 corresponds to the "z" sound, and 8 corresponds to the "br" sound. If you do some speed arithmetic to figure out that you're 48% against a certain hand in your opponent's range on the river, you store the 48% as "zebra" and quickly calculate your equity against the next hand. Once you've done all the hands in his range you retrieve the numbers from their phonetic encodings, quickly sum them, and then divide them by the total number of combinations to figure out your net equity. I believe that it's feasible to do something like this in 20 seconds or so with a lot of practice.

I feel that too much emphasis is placed on the other aspects of poker, and raw calculation and deduction are neglected. My suspicion is that a lot of pros actually do something like the above, but they don't talk about it in their books because they know that people don't want to hear that to become a great player they need to learn some complicated mnemonic system, some complicated mental arithmetic system, and practice using them religiously. People want to hear that they can just casually go over the hands that they play and build up their "intuition" about what to do in certain situations.

Now, there are two parts to basic poker: calculation/deduction, and observation/induction. I also think that it would be interesting to develop a mnemonic system for storing player actions for the duration of the session. There are well-known mnemonic systems for remembering these sorts of things. I think that with enough practice, you could use your memory as a sort of crude poker tracker--but, more importantly, you could also remember the important mistakes that you've witnessed by encoding different sorts of mistakes into your mnemonic system.

The final thing that I think is necessary for any player who seriously wants to become a great player is to flat-out memorize hand matchups in all of the common situations (underpair vs overpair on junk flop, OESD vs TPTK, etc., etc.). If you really look at the different nuances, there are probably a hundred or so equity figures to memorize. I'd approach this like learning the multiplication tables. Just know them by heart.

Maybe I'm just obsessive? I think some of this comes from my experience with chess (I am an expert strength player), because there are a ton of books about strategic ideas, the psychology of the game, and rote opening lines, but a lot of chess boils down to raw calculative ability. Can you calculate forced mates of arbitrary length without getting fuzzy? If not, you will never be a great player. Sure, a lot of books emphasize tactics, but they tend to emphasize doing a lot of tactical problems away from the board to build up your "pattern recognition". While this is great up to a point, pattern recognition can only tell you when to calculate; it rarely gives you the exact right move. I think that "recognizing situations" in poker should increase your ability to know when to calculate, not replace the need altogether.

Maybe I'm just obsessive? Well, maybe, but I think that a lot of players are plateauing because they're happy playing a lazy game that doesn't strive to accurately remember the actions of other players and accurately calculate the correct decision in tough spots.