Bill Chen - "The Mathematics Of Poker" Study Group

D

deuceswild

Enthusiast
Silver Level
Joined
Feb 24, 2015
Total posts
42
Chips
0
For the first part I suppose if you have the villains range defined to those specific hands then it makes sense if this is a toy game. However in real life I would put a whole bunch of other hands that the villain might have on that board that would have me beat. If we are assuming that range is correct then once again I can see it. I just don't see a real life application to that. Part of it is yes I do play only live. America, what a country..... But in this situation unless I am playing a totally bluffing, any two cards playing maniac, I'm not seeing that I have 33% equity with my pair of 5s. As for the second part I like your what is the worst hand they will play for stacks with as i still struggle with certain situations such as when to lay down top two/a flush with a paired board/ the nut straight when the board pairs. You online guys have all sorts of statistics to work with, i just have to go with what i can remember from the sessions i have played with them. And nobody EVER 3 bets unless its aces kings or a maniac. Yup there's no middle ground. Which makes me wonder though. Does 3 betting more often in this scenario make sense as this often precipitates folding at the games i play. So with a raise that is 3 times the bet they have to fold more than 2/3rds to make this a valid exploitative strategy.
 
Fknife

Fknife

Legend
Silver Level
Joined
Oct 26, 2013
Total posts
1,128
Chips
0
For the first part I suppose if you have the villains range defined to those specific hands then it makes sense if this is a toy game. However in real life I would put a whole bunch of other hands that the villain might have on that board that would have me beat. If we are assuming that range is correct then once again I can see it. I just don't see a real life application to that. Part of it is yes I do play only live. America, what a country..... But in this situation unless I am playing a totally bluffing, any two cards playing maniac, I'm not seeing that I have 33% equity with my pair of 5s.

Ok, maybe that was not the best example...

As far as the bolded stuff goes...take a look at the short discussion in this thread: Polished Poker Vol. I Study Group (starting from the #1780 post).


As for the second part I like your what is the worst hand they will play for stacks with as i still struggle with certain situations such as when to lay down top two/a flush with a paired board/ the nut straight when the board pairs. You online guys have all sorts of statistics to work with, i just have to go with what i can remember from the sessions i have played with them. And nobody EVER 3 bets unless its aces kings or a maniac. Yup there's no middle ground. Which makes me wonder though. Does 3 betting more often in this scenario make sense as this often precipitates folding at the games i play. So with a raise that is 3 times the bet they have to fold more than 2/3rds to make this a valid exploitative strategy.

I'm sorry but I have no idea whatsoever how to approach live poker. Online, you can estimate/approximate lots of things so keeping track of Villain's ranges at each spot is much easier (but "they" also can do the same to you since most people these days use HUDs), of course there is still a lot of guessing but I think you can make quite reasonable assumptions just from looking at key stats.

I'll ask a guy, who plays live, to give you some hints regarding those issues.

Btw: Have you tried playing online? I know there are sites which allow US players...
 
D

deuceswild

Enthusiast
Silver Level
Joined
Feb 24, 2015
Total posts
42
Chips
0
Ah I see ya I didn't realize you were referring to a clearly defined range. Ya I mean I'm guessing a better example might be 2nd pair on a non straight non flush board by the river. Like we have j 10 and the board is Ac10s6h3d2h. There may be sites which will allow me to gamble there but I would still be breaking I think both state and federal law by doing so. I like poker but not THAT much.
 
D

deuceswild

Enthusiast
Silver Level
Joined
Feb 24, 2015
Total posts
42
Chips
0
Ok so here's a question. Should we be using exploitive strategies? I have a friend who is very into math and says that players at the live table make enough mistakes that all one has to do is exploit those mistakes. He is aware of the concept of optimal play but says all the live players simply make enough mistakes to exploit. I hear the skill level of internet players is much higher even at lower stakes.
 
Fknife

Fknife

Legend
Silver Level
Joined
Oct 26, 2013
Total posts
1,128
Chips
0
[..]There may be sites which will allow me to gamble there but I would still be breaking I think both state and federal law by doing so. I like poker but not THAT much.

Uhm, I don't really know how it works but I'm pretty sure you can legally play on eg: Bovada and/or those sites. I know lots of US people do play but...tbh I would definitely ask somebody more informed (than me) if you decided to play online.

Ok so here's a question. Should we be using exploitive strategies? I have a friend who is very into math and says that players at the live table make enough mistakes that all one has to do is exploit those mistakes. He is aware of the concept of optimal play but says all the live players simply make enough mistakes to exploit. I hear the skill level of internet players is much higher even at lower stakes.

Take a look at the thread, I previously posted a link to (if you haven't already).

Basically, you will always play exploitively one way or another because nobody knows The GTO solution for No-Limit Holdem (for now). BUT, as we have previously discussed, you don't want to just pick +EV exploitive strategies - you want to pick the highest available +EV exploitive strategies.
 
Fknife

Fknife

Legend
Silver Level
Joined
Oct 26, 2013
Total posts
1,128
Chips
0
No questions about Chapter 4? Ok...coming week is going to be Chapter 5 I guess. It's mostly about Handreading - again nothing really that hard or revealing BUT after we are done with it, I would like to take a little break from the book and talk about some practical approaches to Handreading such as: using basic stats to construct ranges, various types of ranges (capped, polarized etc), range advantages, maybe analyse some "real life" hands together etc.

A little follow-up to read after (or before) Chapter 5: TDA's post about Hand Reading on 2+2

Btw, is everyone familiar with Equity Distribution graphs?
 
R

rhombus

Legend
Silver Level
Joined
Mar 1, 2012
Total posts
2,601
Chips
0
No questions about Chapter 4? Ok...coming week is going to be Chapter 5 I guess. It's mostly about Handreading - again nothing really that hard or revealing BUT after we are done with it, I would like to take a little break from the book and talk about some practical approaches to Handreading such as: using basic stats to construct ranges, various types of ranges (capped, polarized etc), range advantages, maybe analyse some "real life" hands together etc.

A little follow-up to read after (or before) Chapter 5: TDA's post about Hand Reading on 2+2

Btw, is everyone familiar with Equity Distribution graphs?

looking forward to the part about using stats to construct variosu ranges - One of my weaknesses

Btw, is everyone familiar with Equity Distribution graphs? Do you mean the one in chapter 4 I understodd that one
 
Fknife

Fknife

Legend
Silver Level
Joined
Oct 26, 2013
Total posts
1,128
Chips
0
looking forward to the part about using stats to construct variosu ranges - One of my weaknesses
In that case, I'm looking forward to your contribution :)
> Btw, is everyone familiar with Equity Distribution graphs?
Do you mean the one in chapter 4 I understodd that one

No that...that was something completely different :) I'm talking about those kind of monsters:

2015 03 23 20 00 18 Equity graph

For quite some time there was only one program (EDVis) that created such graphs; few days ago I discovered that newer versions of HoldEq (addition to Flopzilla) are now also capable of doing it. Those graphs are very useful, hence my original question.
 
R

rhombus

Legend
Silver Level
Joined
Mar 1, 2012
Total posts
2,601
Chips
0
In that case, I'm looking forward to your contribution :)


No that...that was something completely different :) I'm talking about those kind of monsters:

View attachment 78756

For quite some time there was only one program (EDVis) that created such graphs; few days ago I discovered that newer versions of HoldEq (addition to Flopzilla) are now also capable of doing it. Those graphs are very useful, hence my original question.

kind of get the idea just been on ProPokerTools and did graph for flop 88vAA, learned something at same time didnt realise 88 had 70% equity on 50% of flops vs AA
http://www.propokertools.com/simulations/graph?g=he&h1=88&h2=ak&s=generic


Whats the difference between the two coloured lines
 
Figaroo2

Figaroo2

Legend
Bronze Level
Joined
Sep 9, 2013
Total posts
7,363
Awards
16
Chips
13
No questions about Chapter 4? Btw, is everyone familiar with Equity Distribution graphs?
No
I've noted the graph above and yes what do the two lines represent?
Can't be preflop there are no hands with 90%+ equity pre...
post flop there will be a point with plenty of hands drawing dead with no equity.
What does the 40-60 cross over point mean?
I look forward to you explaining it!
 
Fknife

Fknife

Legend
Silver Level
Joined
Oct 26, 2013
Total posts
1,128
Chips
0
No
I've noted the graph above and yes what do the two lines represent?
Can't be preflop there are no hands with 90%+ equity pre...
post flop there will be a point with plenty of hands drawing dead with no equity.
What does the 40-60 cross over point mean?
I look forward to you explaining it!

Ok, I will be back with an explanation and some examples :)
 
R

rhombus

Legend
Silver Level
Joined
Mar 1, 2012
Total posts
2,601
Chips
0
No
I've noted the graph above and yes what do the two lines represent?
Can't be preflop there are no hands with 90%+ equity pre...
post flop there will be a point with plenty of hands drawing dead with no equity.
What does the 40-60 cross over point mean?
I look forward to you explaining it!

FFS look what you started now, Martin is gonna give us homework now lol :eek:
 
D

deuceswild

Enthusiast
Silver Level
Joined
Feb 24, 2015
Total posts
42
Chips
0
I'm fine with poker homework. There's a reason I'm following a thread about the most complex math book to date. To crunch them numbers son!
 
Fknife

Fknife

Legend
Silver Level
Joined
Oct 26, 2013
Total posts
1,128
Chips
0
I'm fine with poker homework. There's a reason I'm following a thread about the most complex math book to date. To crunch them numbers son!

Alright, there will be homework :) BUT as we go deeper and deeper and I will be the only one posting examples/assignments/notes etc, I will need more time to come up with this stuff. It might seem surprising but I don't know everything :)eek:) and very often I need to watch some videos/read a book/experiment on my own etc before I post anything here (which is good ofc because it gives me an extra motivation but it also requires time).

Anyway, going back to those Equity Distribution graphs (there is an entire chapter in "Expert Heads Up No Limit Holdem Volume 1" about this so if you find my explanation not sufficient, just grab that book). Basically, it's just a way of showing how equity is distributed in a given range. Obviously, we can't just take AA and ask how much equity does it have because...that's hard to say (just like that) :) Therefore, we need some point of reference - another range - that's why there are 2 lines on the graph. It works like this:

1. You specify two ranges (r1, r2); you can also pick a board.
2. For each hand combination in r1, you calculate its equity against range r2.
3. You sort all the hands in r1 by their calculated equities and than you just create a graph where:
Y-axis: Equity
X-axis: Hand Combinations (starting from the strongests)
4. You repeat steps: 2. and 3. for every hand in r2.

Those graphs are useful because:
a) they visualize your own range so you know how it looks like in any given spot (is it too strong/weak/capped/polarized/etc? if my strategy involves check-raising Top X% of my hands, what hands am I supposed to do this with etc).
b) they visualize opponent's ranges (can I thin-valuebet vs that range; how many bluffs does opponent have etc; it will be especially useful in River play).
c) they show range advantages: basically some boards/turns/rivers might favour different ranges (eg: A, K-high flops usually favour 3betting ranges, while Q,J-high flops favour 3bet calling ranges). We will look at that closely in another post (I already have an example prepared).


Graph 1: So the first cool graph is an Equity Distribution of ATC vs ATC ranges. It looks like this:
ATCvsATCDist

It might look familiar to some people. Looking at the graph, the strongest hand has about 85% equity against ATC and those are...AA! Next are:
[AA (85.2%), KK, QQ, JJ, TT, 99, 88, AKs, 77, AQs, AJs, AKo, ATs, AQo, AJo, KQs, 66, A9s, ATo, KJs, A8s (61.9%)] - those are Top 10% of all Holdem starting hands!

AKs actually has about 67% equity and after it, the equity line declines fairly slowly and there are lots of hands close in equity (remember that there are 1326 combos in an ATC range and every combo has its "place" on the X-axis). The worst hand is 32o - it has 32.3% equity.

Graph 2: We've already talked about made hands vs draws. Let's say we have a board: Ah9c6c and some 65% range. What is a difference between KsKd (made hand) and 8c7c (draw)? They both are similar in their equities (67% and 61% vs 65% range)... but:

This is an Equity Distribution of 65% range vs KcKd (67% equity):

KsKdonAh9c6c

And this is 65% range vs 8c7c (61% equity):

8c7conAh9c6c

Thoughts? :)

Btw, if you don't have a full version of Flopzilla, you can still get those graphs using a tool called EDVis. It's free and I think I've already mentioned it.
 
R

rhombus

Legend
Silver Level
Joined
Mar 1, 2012
Total posts
2,601
Chips
0
I'm fine with poker homework. There's a reason I'm following a thread about the most complex math book to date. To crunch them numbers son!
Only joking about the homework :)

Is it possible to use the graphs on ProPoker Tools or are they totally different
 
R

rhombus

Legend
Silver Level
Joined
Mar 1, 2012
Total posts
2,601
Chips
0
Thoughts? :)

Btw, if you don't have a full version of Flopzilla, you can still get those graphs using a tool called EDVis. It's free and I think I've already mentioned it.

Is the reason the 87 has a flatter line toward the end whereas the KK declines alot steeper because the 87 needs to benefit from seeing both Turn and River :confused::confused:
 
Figaroo2

Figaroo2

Legend
Bronze Level
Joined
Sep 9, 2013
Total posts
7,363
Awards
16
Chips
13
I suspect it is because some hands have to improve twice to beat the kings but only need to improve once to beat 87.
take any unsuited separated hand like Q3 J4 T2, these hands must have more equity against 87 as they only need to pair the top card to get in front whereas against the Kings the hand has to improve twice to 2pair or trips which is quite a long shot. Therefore there are many hands with very little equity against the kings thus a steeper switchover in the chart for the kings.
 
Fknife

Fknife

Legend
Silver Level
Joined
Oct 26, 2013
Total posts
1,128
Chips
0
Is it possible to use the graphs on ProPoker Tools or are they totally different
I don't know tbh :(

Is the reason the 87 has a flatter line toward the end whereas the KK declines alot steeper because the 87 needs to benefit from seeing both Turn and River :confused::confused:

I suspect it is because some hands have to improve twice to beat the kings but only need to improve once to beat 87.
take any unsuited separated hand like Q3 J4 T2, these hands must have more equity against 87 as they only need to pair the top card to get in front whereas against the Kings the hand has to improve twice to 2pair or trips which is quite a long shot. Therefore there are many hands with very little equity against the kings thus a steeper switchover in the chart for the kings.

Yea, it's basically what you guys said. :) KK on A-high board is actually a classic example of WA-WB situation where when you start to be too aggressive with it, you will only be called by better (in a 65% range, 20% of hands are actually beating KK). You still can get some value from worse (eg: FDs which contribute to only 5% of range, or weak 9x) but it probably would not be a 3-streets value (also remember we are talking about 65% range here). If you are behind, there is a very little chance of improving (same goes with hands which are behind in that range eg: QQ need 2 outs to come back from their 10% equity):
KsKdonAh9c6cEdited

As far as 8c7c goes...its equity will change drastically on later streets depending on Turn/River cards - that's for sure. Also notice that only about 1.5% of 65% range is currently slightly ahead of 8c7c (those are mostly A-high flush draws). Other than that, 8c7c is beating even sets and Ax not to mention other weak pairs. BUT notice also that trashy, unpaired hands still have about 30% equity (eg: K4o), what's even more important -> only the bottom 10% of the range has < 30% equity (looking back at KK: bottom 70% already has < 30%...). So yea, you can say that this is one of the "reasons" for semibluffing (which was on the intersection of Protection and bluff circles on my beautiful "Reasons for Betting" painting).

8c7conAh9c6cEdited

So that's a thing to keep in mind: there is a difference between draw equity and a made hand equity. Same goes with entire ranges: two ranges can have the same amount of equity against another range but that equity can be distributed differently (which is very important from a strategic point of view).
 
Fknife

Fknife

Legend
Silver Level
Joined
Oct 26, 2013
Total posts
1,128
Chips
0
Chapter 6 has only 2 pages and presents mostly an online approach to concepts from chapter 5 therefore I will just combine both chapters into a single summary post (following week). See ya.
 
D

deuceswild

Enthusiast
Silver Level
Joined
Feb 24, 2015
Total posts
42
Chips
0
I am eagerly awaiting the next post...
 
Fknife

Fknife

Legend
Silver Level
Joined
Oct 26, 2013
Total posts
1,128
Chips
0
Chapter 5/6 Summary.

1. Common hand reading mistakes.
a) "Ranging" a single hand ("I put you on AK" syndrome).
b) Projecting your own strategies into opponents ("he would never open 45o UTG" syndrome).
c) Too aggressive range narrowing ("he would never check-raise with that hand here" syndrome).

2. A general approach to handreading.
1. Assign a full distribution of random hands (starting Preflop), with each adjusted for Bayesian Inference and card-removal effects.
2. As hand progresses modify various probabilities based on revealed cards and action taken (process of reduction and elimination).
3. Don't add hands which you have just removed on earlier streets.

Additionally:
- Always include "lost his mind"/"had a bad day at work"/"feels lucky today"/tilt probabilities.
- How will opponents play their hand distributions assuming they will play "well" (eg: as PFRs what goes into their: cbet/x-c/x-r/x-f ranges; when are they capped etc) + how it relates to their stats.
- Gathering enough info to narrow to just a single hand generally impossible.
- TDA's post about Hand Reading on 2+2
- Ed Miller - "How To Read Hands At No-Limit Hold'em"

3. Strategy Reading Techniques.
a) Direct hand evidence: at showdown (eg: how he plays his TPs when draws get there (bet/bet/x-c ???), does he pot control the turn with TPTK on dry boards (bet/x/bet ???); does he always raise his two pairs+ on dry boards leaving himself capped when he just calls?).
b) Opponent controlled hand evidence: opponent shows you (or tells) his cards. Information there not as reliable, as he maybe wants to get a psychological reaction (this probably applies mostly to live play).
c) Indirect hand evidence: Frequencies of an action (in online poker those are the stats).
d) Player classification correlation + basic adjustments: passive-aggressive, loose-tight etc.
e) Population tendencies/reads: "How does an average player at this limit play" / what is he capable of?

Know the limitations:
- Very few hands get shown down.
- Some stats need a lot of hands to converge (especially late streets stats) and are not as accurate as people like to think they are (eg: Fold To Cbet does not take board texture, stack sizes etc into account although this might be probably "solved" (or at least approximated) with the use of software such as NoteCaddy).
- While we can quickly identify outlines within few hundreds hands, the regular stats can be difficult to pin down (eg: open BTN 32 or 38?).
- Stronger range gets to SD more often, so introducing an artificial bias.
- Players are not that stupid as Hero likes to think they are. They are also adjusting, not playing a single fixed strategy all the time against everyone.
- Players get better in the long run (they are reading books/forums and watching training videos too).
-
gif.latex
where: T - a tell and A - a certain event attributed to T

(the second term represents a false positive - how often we see a tell T and it has no correlation to event A whatsoever. As this term goes to 0, the value/reliability of a tell T increases)

4. Online Poker.
- Absence or limited presence of tells (mostly betting patterns).
- Hand Histories, HM2/PT4 as databases (actually those programs are using PostgreSQL as a database per se but... whatever :) ) + HUDs.
- Data gathering is cool but it's only a beginning. Interpretetion of that data is much more important and difficult.

Hand History biases:
- You have a player in your database and he's breakeven. Is he a good or a bad player? (hard to say because you collected mostly hands played with you sitting at the table. If you are also a winning player, he likely is also a winning player.)
- The only player for which you have an unbiased record is you (great way to analyse your own play/results).
- Beware of mixing together hands (and stats) from SH, FR and HU games.
- HUDs do have value! Understanding outliners, spotting deviations (cbetting 70%+ is definitely "too much" whereas cbetting only 20% definitely too little) and seeing how various stats influence each other is extremely important.

Online play tougher than live: (sorry, that's what the book says :))
- Breaking point for losers is reached faster (and lots of people have to lose so that others can win).
- Can fool one self easier live. Online you have hard data that you suck!
- Small net losers have time to monthly reload.
- More hands, faster, convenient, bringing out better players. Volume is a big key.


Do we want to go over some basic stats or just continue with the book or maybe talk about something else?
 
D

deuceswild

Enthusiast
Silver Level
Joined
Feb 24, 2015
Total posts
42
Chips
0
HM I'm trying to build a solid base myself I'd like basic stats but if everybody else wants to move on I will concede. I play live only, I did some research, online play is technically still illegal in WI, its not prosecuted but not a chance I wanna take. However, should it become legal, going over some basics would definitely help. Yes even I have heard time and time again that online play is harder than live. At first I was even a little offended and that online players were just full of themselves, BUT after hearing just about everybody say it, even Brunson has said it, some pros who play from just over the border in mexico have said it, even my amateur buddy who played before black Friday in America has said it. Ergo I may as well accept it, but nobody wants to think their game is somehow inferior, oh well boo boo.
in any case, this book recommends not excluding categories such as lost his mind and so forth, even the book recommends bluffing with terrible hands occasionally, how much of a factor should these things be. The book seems good at throwing concepts out there but sometimes fails to really suggest how important they are or in what ratio this may happen.
 
D

deuceswild

Enthusiast
Silver Level
Joined
Feb 24, 2015
Total posts
42
Chips
0
Don't tell me that I am the only one left paying attention to this post
 
Top