Open Side Menu Go to the Top
Register
COTW: The Mathematics of Poker COTW: The Mathematics of Poker

09-02-2010 , 12:43 PM
nice post need to reread it a few times. I am amazed on how much of the stuff has become second nature to me, but 3 years ago I sat down with matlab and did all this stuff about 1000 times.

One thing I want to not, is that stacks sizes are the foundation of how to calculate implied odds and ranges and effect our EV decisions, but what we really need to get into our thought is what does the villain do with his stack. If some nit opens UTG with JJ+, with 200BB, but wont put in more than 70BB without topset, then playing 22-TT for set value is not good. At the same time if we have a 40BB player who play on the surface is to open raise 99+, AJ+, but in squeeze situations they are only shoving A2+ and flatting everthing else, then with dead money involved we could be +EV to call the shove with JTs+
COTW: The Mathematics of Poker Quote
09-02-2010 , 12:54 PM
That's what you call a brick to the wall of my poker education!!!!!!!
Gotta go back to rea now!!!!!!!!!
Thx LX12
COTW: The Mathematics of Poker Quote
09-02-2010 , 01:48 PM
Quote:
Originally Posted by kmd107
Wow, solid post Lx12, lots of great info
No prob man. Glad someone is finding it useful
COTW: The Mathematics of Poker Quote
09-02-2010 , 01:50 PM
Quote:
Originally Posted by probaballistics
whoa - huge post dude. great job and thanks for the efforts!
Yeah it ended up being around 12000 words lol. I didn't know what I was getting myself in to at the beginning of this! Thank you and you're welcome
COTW: The Mathematics of Poker Quote
09-02-2010 , 01:53 PM
Quote:
Originally Posted by wmermus
graphomaniac....

just to imagine that I have read it all.....
Remember though, reading is not the same as understanding or not the same as being able to do! So go and create your own scenarios and do them yourself!
COTW: The Mathematics of Poker Quote
09-02-2010 , 01:53 PM
Quote:
Originally Posted by SammyG-SD
nice post need to reread it a few times. I am amazed on how much of the stuff has become second nature to me, but 3 years ago I sat down with matlab and did all this stuff about 1000 times.

One thing I want to not, is that stacks sizes are the foundation of how to calculate implied odds and ranges and effect our EV decisions, but what we really need to get into our thought is what does the villain do with his stack. If some nit opens UTG with JJ+, with 200BB, but wont put in more than 70BB without topset, then playing 22-TT for set value is not good. At the same time if we have a 40BB player who play on the surface is to open raise 99+, AJ+, but in squeeze situations they are only shoving A2+ and flatting everthing else, then with dead money involved we could be +EV to call the shove with JTs+
Very good point as always sammy
COTW: The Mathematics of Poker Quote
09-02-2010 , 02:53 PM
holy ****
COTW: The Mathematics of Poker Quote
09-02-2010 , 03:19 PM
Wow, tldr. A for effort. The overall approach is very good. However I have quite a few problems with the contents, which I'll point out while I'm working through it all. (Bear in mind, I'm a Maths teacher, so I might come across as overly pedantic.)

Quote:
Every time you play a hand differently from the way you would have played it...
There are so many issues with this so-called theorem... but let's not get into that. Let's just point out the following:

Quote:
The idea upon which it is based is that if we could see an opponent’s cards, we would be able to determine an action which maximizes our expected value (EV) i.e. make a decision which could be mathematically proven to be optimal.
Our EV does not only depend on the holecards. Our opponent's strategy has a big impact on the expected outcome, in particular when he plays far from optimally.

Quote:
What this means is that if one or more players have a +EV proposition, one or more players have a –EV proposition. This is important because every time your opponent makes a –EV decision, you make a +EV decision (assuming only two players).
When we are making a decision our opponent doesn't decide anything. He has to wait his turn. If our opponent's strategy is fixed we can't actually improve our EV; the best we can do is not letting it deteriorate. The whole "+EV" lingo sounds pseudo-scientific, but it is terribly colloquial and inexact.

Quote:
a range is simply the possible holdings an opponent (or you yourself) could have.
More accurately, a range is any set of possible hole cards.

Quote:
EV=1.5(1-c-r)+ c(6e-2.5)+2.5(b-r)+ 10(r-s)+ f(200v-99.5)+35(f-b) – Equation 1
I think there are a few flaws in the derivation of this formula, and hence it is inaccurate. This affects the remainder of the text. I'll suggest a correction in a later post. Edit: My bad, it looks like I didn't read it correctly.

Last edited by Cangurino; 09-02-2010 at 03:30 PM.
COTW: The Mathematics of Poker Quote
09-02-2010 , 04:38 PM
Quote:
Originally Posted by Lx12
Theory:

Say hypothetically that in the SB we raise to 3BB (Effective stacks of 100BB). This means we are investing 2.5BB (in addition to the 0.5BB that we have already posted) to win 1.5BB (because while it may be our small blind it belongs to the eventual winner of the pot and not us).

Assumptions:
  1. We open q% hands.
  2. Our opponent in the small blind will call with c% hands
  3. Our opponent will 3-bet (raise) to 10BB with r% of hands
  4. This means he will fold (100-c-r)% of the time.
  5. When we are called by the BB we have an equity of e (as a decimal)
  6. We cannot profitably call a 3-bet out of position i.e. we must raise or fold
  7. We will 4-bet to 35BB with b% of hands
  8. Our opponent will either fold or 5-bet shove with s% of hands in response to a 4-bet.
  9. We will call a 5-bet shove with a frequency of f%
  10. When we call a 5-bet shove, we will have an equity of v%

What happens now is the following (assuming we raise in the first place):
  • We have a hand that calls a shove with probability f/q.
  • We have a 4-bet-fold hand with probability (b-f)/q.
  • We have a bet-fold hand with probability (q-b)/q.
  • Villain has a shoving hand with probability s.
  • Villain has a 3-bet-fold hand with probability r-s.
  • Villain calls with probability c.
  • Villain folds with probability 1-c-r.

Since our hand is independent from Villain's hand (ignoring negligible card removal effects) we can multiply the probabilities of events concerning us and him respectively. The possible outcomes are as follows:
  • We win uncontested, P=1-c-r, result +1.5.
  • We get called and win, P=ce, result +3.5.
  • We get called and lose, P=c(1-e), result -2.5
  • We fold to a 3bet, P=r(q-b)/q, result -2.5
  • Villain folds to a 4bet, P= (r-s)b/q, result +10.5
  • We fold to a shove, P=s(b-f)/q, result -34.5.
  • We call a shove and win, P=vsf/q, result +100.5
  • We call a shove and lose, P= (1-v)sf/q, result -99.5.

If we sum up all these probabilities we'd better get 1:
1-c-r + ce + c(1-e) + r(q-b)/q + (r-s)b/q + s(b-f)/q + vsf/q + (1-v)sf/q
= 1-c-r + ce + c - ce + r -rb/q + rb/q - sb/q + sb/q - sf/q + sf/q
= 1

So the overall EV is
1.5 (1-c-r) + 3.5 ce +2.5(ce-1) + 2.5r(b-q)/q + 10.5 (r-s)b/q + 34.5s(f-b)/q + 100.5 vsf/q + 99.5 (v-1)sf/q
= -1 - 1.5(c+r) + 6ce + 2.5 rb/q - 2.5 r + 10.5 rb/q - 10.5 sb/q + 34.5sf/q - 34.5 sb/q + 200vsf/q-99.5sf/q
= -1 -1.5 c - 4r + 6ce +13rb/q - 45sb/q-65sf/q +200 vsf/q

So we actually need to consider our opening range in the EV calculations.
COTW: The Mathematics of Poker Quote
09-02-2010 , 04:46 PM
Quote:
Originally Posted by CrazyNL
holy ****
+1

tl;dr doesn't even apply to this...wat

Looks detailed and insightful though, will bookmark and read later.
COTW: The Mathematics of Poker Quote
09-02-2010 , 05:06 PM
I'm also unsure about the accuracy of this formula. We don't necessarily need to consider our opening range if we consider b, f etc. as percentages of our opening range that we would 4bet, call a shove with etc., respectively.

There are a few issues I have... for example, if we raise, get 3bet, and fold, shouldn't the EV be -2.5r(1-b)? It doesn't seem like the expression r-b should appear anywhere in the formula, since they are independent (r is part of his range, b is part of our range).

So far, it looks like Cangurino has things wrapped up pretty well, so... +1 for that. Still, mad props on an excellent job Lx.

Last edited by soupersong; 09-02-2010 at 05:11 PM.
COTW: The Mathematics of Poker Quote
09-02-2010 , 05:28 PM
Quote:
Originally Posted by Lx12
  1. Knowledge of how to use PokerStove (See COTW 1 : Equity Exploration)
  2. Knowledge of how to use StoxCombo (See COTW 1 : Equity Exploration)
Is this app still available anywhere?
COTW: The Mathematics of Poker Quote
09-02-2010 , 05:30 PM
I don't know if it is still available or not, but in some thread split said that he uses flopzilla now.
COTW: The Mathematics of Poker Quote
09-02-2010 , 07:17 PM
This is excellent. My math skills are weak and the concepts do not come easily to me, so I appreciate the step-by-step explanations and "homework" as well as the discussion and input from others (like math-whiz Cangurino ). I will work through this COTW slowly and most likely multiple times to completely gain a proper understanding of the concepts.

Thank you so much for this COTW.
COTW: The Mathematics of Poker Quote
09-02-2010 , 08:30 PM
Quote:
Originally Posted by CrazyNL
holy ****
+2

I got started with the third section and realized that I'm going to need a couple of days of study and playing with it just to begin to get value from it. So far, if I get 4bet with 35BB, I'll just mark the player as a 2+2 lurker.
COTW: The Mathematics of Poker Quote
09-02-2010 , 09:23 PM
Great stuff! I actually printed this as I'm way better calculating math on actual paper. Will be working through it, although I don't expect to be finished in the nearest days.
COTW: The Mathematics of Poker Quote
09-02-2010 , 09:39 PM
omfg nerdgasm

you put alot of work into this. only part way through as I want to do my homework and am too tired right now. very good so far.
COTW: The Mathematics of Poker Quote
09-02-2010 , 11:00 PM
I've always thought that this kind of thing is better described graphically.



First, we are only concerned with the smaller rectangle comprised of colored boxes (and not including the large white box on the left), since we are looking to find the EV of raising.

Straight up profitable boxes (villain folds) are shaded green; straight up unprofitable boxes (we fold) are in red; and boxes where there are no folds preflop are shaded yellow.

As an example, if villain has a hand that he will 3b with but fold to a 4b, and we have a hand that will 4b with but fold to a 5b, then what will happen?

1. We raise
2. Villain 3bets
3. We 4bet
4. Villain folds,

so that particular small box (villain = 3b but fold to 4b, us = 4b but fold to 5b) should be colored green.

Finally, the total EV is the sum of the net gains and losses in each of the boxes (weighted by the area/probability of each box).


*****
One possibly confusing final note: the q I use is the same as the q Cangurino uses, but the b I use is actually the b/q Cangurino uses - I thought it detracted from the main point, and looked cleaner/less confusing written down.

The difference is this - are you considering your 4bet range as a fraction of your raising range? If so, then your b is the same as Cangurino's. If you are considering your 4bet range as a fraction of all possible hands, then your b is the same as mine.


edit: ack, the bottom right corner of the graphic should say "raise, 4b to 3b, call 5b" instead of "raise, 4b to 5b, call 5b".
COTW: The Mathematics of Poker Quote
09-02-2010 , 11:10 PM
Quote:
Originally Posted by thejuggernaut
omfg nerdgasm

you put alot of work into this. only part way through as I want to do my homework and am too tired right now. very good so far.
+1.

Really looking forward to diving into this tomorrow.
COTW: The Mathematics of Poker Quote
09-02-2010 , 11:44 PM
but math is sooo tedious why did u have to post this.... here we goooo see you guys in 2 years when I get thru this.
COTW: The Mathematics of Poker Quote
09-03-2010 , 05:11 AM
Lx: In computing the hands we can 4bet if we get 3bet, we win 10.5bb if he folds (with probability 1-s/r, and if he shoves and we call, our equity is (s/r)(200v-99.5). (We win 101.5bb if we win, and 99.5bb if we lose). We need the sum of these two to be positive.

Now in your example, you have r = 0.042, s = 0.026, so s/r = 0.62, so we need

0 < 10.5*(1-s/r) + (s/r)(200v-99.5)
= 10.5(0.38) + 0.62(200v-99.5)
= 124v-57.7

so we need v > 57.7/124 = 0.465 (instead of the 0.395 that you got).

Thus it's unprofitable to 4b and stack off against this villain with QQ or even AK (which makes sense, since his 3b and 5b shove ranges are pretty close to each other). So yeah, b = f and our stackoff range against this villain is exactly KK, AA.
COTW: The Mathematics of Poker Quote
09-03-2010 , 05:48 AM
Quote:
Originally Posted by soupersong
Lx: In computing the hands we can 4bet if we get 3bet, we win 10.5bb if he folds (with probability 1-s/r, and if he shoves and we call, our equity is (s/r)(200v-99.5). (We win 101.5bb if we win, and 99.5bb if we lose). We need the sum of these two to be positive.

Now in your example, you have r = 0.042, s = 0.026, so s/r = 0.62, so we need

0 < 10.5*(1-s/r) + (s/r)(200v-99.5)
= 10.5(0.38) + 0.62(200v-99.5)
= 124v-57.7

so we need v > 57.7/124 = 0.465 (instead of the 0.395 that you got).

Thus it's unprofitable to 4b and stack off against this villain with QQ or even AK (which makes sense, since his 3b and 5b shove ranges are pretty close to each other). So yeah, b = f and our stackoff range against this villain is exactly KK, AA.
Ugh, I'm sorry, please disregard the above nonsense. Lx is totally correct, and I just realized the problem- when we determine our 4b range against this villain, we are, in fact, winning 13bb if he folds, and have equity 113v-87(1-v) = 200v-87. I got confused because (and I imagine this is a source of confusion to many who try to work through this) in the formula for the overall EV, 200v-99.5 is correct (I think), but in calculating your EV after you have already raised and are facing a 3bet.... in that spot, your 200v-87 is now correct.
COTW: The Mathematics of Poker Quote
09-03-2010 , 05:59 AM
Very very solid post.
It will takes me lot of time to digest this great post but I think it is so profitable!
Thank you very much
COTW: The Mathematics of Poker Quote
09-03-2010 , 08:07 AM
Thank you for this.
I send this thread to the printer and will study it over the weekend.
COTW: The Mathematics of Poker Quote
09-03-2010 , 09:58 AM
Quote:
Originally Posted by Lx12
Homework related to this chapter:
I was told there'd be no math

Ok, here's my solution. It does not coincide with OP's, but that's to be expected since I don't agree with his formula.



Spoiler:
We will work backwards. Villain shoves TT+,AQs+,AQo+ (4.7%). It costs us 65 to call in order to win a total pot of 200. Hence we need 32.5% equity to call the shove. Against his range we can call with 22+,AQ+.

He raises 9.4% of the time. This means that he folds to our 4-bet exactly half of the time. So when is 4-betting profitable?

When we 4-bet, we win the pot (13BB) half of the time. The other 50% of the time we lose 32BB if we don't have a calling hand. Since 0.5*13-0.5*32=-9.5, 4-bet-folding is a losing proposition.

So when can we 4-bet-call? We need to invest 97 BB in order to win a 200BB pot, so we need 48.5% equity against his shoving range. AKo is 49.175%, AK is over 50%, so is QQ. JJ is only 46.9%, so we should fold it.

So, our 4-betting range is just QQ+,AK. When we 4-bet we win 13BB 50% of the time, need to invest 97BB the rest of time and expect to win 119BB for a profit 22BB. So when we have a 4-betting hand and get 3-bet we expect a profit of .5*13+.5*22=17.5BB.

Now, let's see what happens when we raise initially. He raises 9.4%, calls 45.7%, and hence folds 44.9%. When he folds we win 1.5BB. When he raises and we don't have a monster, we lose 3BB. When he raises and we do have a monster we win 15BB (17.5BB from the last paragraph, minus our initial raise). When he calls the pot is 6BB and we invest 2.5 to win according to our equity e against his calling range.

All our 4-betting hands do well against his calling range, so we will certainly raise with them. For any other hand that has equity e against his calling range, the expected value is

E=.449*1.5 -.094*3 -2.5 + 6*e
= .6735 - .282 - 2.5 + 6*e
= -2.1085 + 6*e

If raising is better than folding then this needs to be positive
6*e -2.1085 > 0
6*e > 2.1085
e > .3514

So we can profitably raise with any hand that is at least 35.14% against his calling range. This means:
Any pair, any hand containing a jack or better, any suited hand except for 82s and 72s, T5o+, 96o+, 86o+, 76o, 65o, 54o.

In summary: We bet 79% of the hands ( 22+, A2s+, K2s+, Q2s+, J2s+, T2s+, 92s+, 83s+, 73s+, 62s+, 52s+, 42s+, 32s, A2o+, K2o+, Q2o+, J2o+, T5o+, 96o+, 86o+, 76o, 65o, 54o), but we fold everything to a 3-bet except for 2.6% (QQ+, AKs, AKo); these hands we 4-bet-call.
COTW: The Mathematics of Poker Quote

      
m