www.flopturnriver.com

Poker Forum

Over 1,246,000 Posts!

Already Registered? Username: Password: Remember Forgot Password
Tweet

Maniac Math

LinkBack

Thread Tools

Display Modes

Robb

10-12-2009, 04:28 AM Post subject: Maniac Math

#1 (permalink)

4-of-a-Kind

Join Date: Aug 2007

Posts: 3,072

Robb is an unknown quantity at this point

I've seen some threads in the BC lately that have me concerned. Someone will post a HH with villain's stats something like 50/30 over 30 hands but will discount that - could be just a TAGG on heater.

Not very likely. Using the binomial theorem, we can calculate the probability that a TAGG with 25/15 stats raises 30% of the time (or more) over 30 hands: about 2.75%. Over 50 hands, there's about 0.5% chance an actual 25/15 guy will raise 15 times.

My point is that when you see a couple dozen hands from a guy who's seein' the flop half the time, and raising a good bit, you don't need 1k HH's to assume he's pretty loose. It's time to mix up with him a bit before he donates his stack to someone else.

If anyone's interested in the math, I'll post something more detailed tomorrow.

My Operation and FTR Rethread: Stations are *sob* so hard to play against

Play for FREE and practice your game at...

Join the FTR Poker Forum to disable these banners and start posting!

dranger7070

10-12-2009, 04:32 AM

#2 (permalink)

4-of-a-Kind

Join Date: Oct 2008

Location: Wake up in the mornin feelin' like P. Diddy

Posts: 2,524

dranger7070 has a little shameless behaviour in the past

[x] interested

BooG690

10-12-2009, 04:39 AM

#3 (permalink)

Join Date: Mar 2009

Posts: 1,439

I smell some standard deviations a-comin'.

That's how winners play; we convince the other guy he's making all the right moves.

Ztech

10-12-2009, 04:45 AM

#4 (permalink)

Join Date: Jul 2008

Location: Staying out of the FTR Community threads and sticking to the pokerz

Posts: 58

Very interested in this. Run into this type of situation quite a bit.

JKDS

10-12-2009, 04:58 AM

#5 (permalink)

Full House

Join Date: Feb 2008

Location: Chandler, AZ

Posts: 1,024

http://en.wikipedia.org/wiki/Binomial_theorem

???

you mean bayes theorem?

Quote:

Originally Posted by OngBonga

But no, jkds is lolvillager and anyone who wants to string him up is sighbad.

Robb

10-12-2009, 11:58 AM

#6 (permalink)

4-of-a-Kind

Join Date: Aug 2007

Posts: 3,072

Quote:

Originally Posted by JKDS

http://en.wikipedia.org/wiki/Binomial_theorem

???

you mean bayes theorem?

Yeah, a lot of folks use Bayes Theorem here and treat this as conditional probability, but we can make a simple assumption and just compute the probability directly. The Binomial Theorem works for repeatedly "coin flip" situations, whether the coin is fair or not.

Suppose we flip a coin with a 20% probability of heads. Let x = .2 (probability of success) and y = 1 - .2 = .8. Now, this may seem trivial at first, but note that x + y = 1. So (x+y)^5 = 1 for instance. This happy fact allows us to simply expand the binomial power and calculate the probabilities we want. If you remember Pascal's Triangle in high school algebra, it's the simplified version of the Binomial Theorem. And the process works for any probability simulation we can create with fixed chance of success.

So, imagine a TAGG opening in the CO with a known chance of success (open-raising) of x = .2. And imagine that we see him in this situation 5 times (after learning the iron-clad fact he open-raises 20% in the CO). What is the probability he open-raises 4 times or more out of 5 trials?

The probability of 5 successes is a straightforward (.2)^5.

The probability of 4 successes is a bit more interesting:

(5 choose 4)(0.2^4)(0.8)

The 4 successes each have probability 0.2, the failure has probability 0.8. But there are 5 ways to choose one slot out of 5 for the failure to occur (or 5 choose 4 ways to choose the 4 slots for the successes to occur):

F S S S S
S F S S S
S S F S S
S S S F S
S S S S F

Recalling Pascal's Triangle:

1
1 1
1 2 1
1 3 3 1
1 4 6 4 1
1 5 10 10 5 1

The bottom row allows us to expand the binomial (x+y)^5 by telling us the coefficients in the pattern where powers of x ascend from 0 to 5 and powers of y descend from 5 to 0:

(x + y)^5 = y^5 + 5 x y^4 + 10 x^2 y^3 + 10 x^3 y^2 + 5 x^4 y + x^5.

Again, recall that this value is just = 1, since x + y = 1. In the expansion, the pattern relates to the probabilities:

y^5 term - probability of 0 successes, 5 failures
x y^4 term - probability of 1 success, 4 failures
x^2 y^3 term - probability of 2 successes, 3 failures, and so on

Example: For a fixed chance of success x = .2, what is the probability of 5 successes in 8 trials?

P(X = 5) = (8 choose 5)(0.2^5)(0.8^3)

The "8 choose 5" simply calculates all the possible ways to permute the 5 successes into 8 slots.

Awright, given all that introduction, the setup for my post is simple.

Hypothesis: let's assume we have a TAGG reg who plays <25/15 over some reasonable sample of hands. We can think of his PFR = .15 as a chance of success, so x = .15. Then the probability of this event occurring 15 times OR MORE out of 50 trials is the sum:

P(X >= 15) = (50 choose 15) (.15^15)(.85^35) + (50 choose 16) (.15^16)(.85^34) + (50 choose 17) (.15^17)(.85^33) + ...

I did the sum in an Excel spreadsheet. There's an extension of the binomial theorem called the multinomial theorem which would allow us to take VPiP into account simultaneously with PFR, which I may get into later if you're still interested after all that.

Just to check the viability of this math, you can see the same calculations for other probabilities.

For x = .2, there is about a 6% of him open-raising 15 times (or more) out 50.

For x = .25, there is about a 25% of him open-raising 15 times (or more) out 50.

For x = .3, there is about a 55% of him open-raising 15 times (or more) out 50.

For x = .4, there is about a 94% of him open-raising 15 times (or more) out 50.

For x = .5, there is more than a 99% of him open-raising 15 times (or more) out 50.

Let me know what you think. If you have more questions, I can try to answer them.

My Operation and FTR Rethread: Stations are *sob* so hard to play against

Micro2Macro

10-12-2009, 05:48 PM

#7 (permalink)

Moderator

Join Date: Jan 2009

Location: http://three-pair.com/

Posts: 4,463

After studying this theorem closely 1-2 months ago I can say my win rate has improved and my awareness has increased significantly.

"Once we reach a certain level of mastery, we see there are higher levels and challenges. If we are disciplined and patient, we proceed. At each higher level, new pleasures and insights await us--ones not even suspected when we started out. We can take this as far as we want--in any human activity there is always a higher level to which we can aspire."

Check out my blog here!

"You are a degenerate Gaam-balur"

http://www.philgalfond.com/lets-make-some-changes/

bjsaust

10-13-2009, 12:05 AM

#8 (permalink)

Straight Flush

Join Date: May 2007

Location: Ballarat, Australia

Posts: 5,842

I for one am prepared to accept your conclusion.

Just playing to improve.

dranger7070

10-13-2009, 12:29 AM

#9 (permalink)

4-of-a-Kind

Join Date: Oct 2008

Location: Wake up in the mornin feelin' like P. Diddy

Posts: 2,524

Quote:

Originally Posted by bjsaust

I for one am prepared to accept your conclusion.

lol same

I didn't quite understand the point of all the gibberish at first, but I just went and looked it over again, and I think I understand its usefulness now. A guy starting out 30/20 over 25 hands probably isnt going to be a 10/8 on a heater all that often, etc.

I don't know if thats the whole scope of the post, and maybe I'm missing the actual point of the thread, but thats what I got out of it. :/

Robb

10-13-2009, 12:46 AM

#10 (permalink)

4-of-a-Kind

Join Date: Aug 2007

Posts: 3,072

Quote:

Originally Posted by dranger7070

Quote:

Originally Posted by bjsaust

I for one am prepared to accept your conclusion.

lol same

I didn't quite understand the point of all the gibberish at first, but I just went and looked it over again, and I think I understand its usefulness now. A guy starting out 30/20 over 25 hands probably isnt going to be a 10/8 on a heater all that often, etc.

I don't know if thats the whole scope of the post, and maybe I'm missing the actual point of the thread, but thats what I got out of it. :/

That was it - some folks wanted to be convinced, I guess.

25 hands of 60/25 is enough to be pretty doggone certain he's no TAGG. If you've got that, you're gtg.

My Operation and FTR Rethread: Stations are *sob* so hard to play against

dranger7070

10-13-2009, 12:51 AM

#11 (permalink)

4-of-a-Kind

Join Date: Oct 2008

Location: Wake up in the mornin feelin' like P. Diddy

Posts: 2,524

Sweet, good post Robb. I've always known that to be the case, but makes you feel a lil better seeing the maffs that make it true.

mbiz

10-13-2009, 08:18 AM

#12 (permalink)

Straight

Join Date: Jun 2009

Location: Orstraya

Posts: 178

mbiz is an unknown quantity at this point

moral of the story: low sample size (30hands) stats can usually (but not always) be interpreted to be roughly the same as bigger sample size stats. for extra points use showdown reads

StillDeadMoney

10-13-2009, 10:42 AM

#13 (permalink)

Join Date: Sep 2006

Location: The Crazy Russian

Posts: 19

StillDeadMoney is on a distinguished road

Unless you're playing somebody who's randomizing their play. Then again, if you do, you're probably playing at a level where you don't need this forum's help.

Quote:

Originally Posted by bigred

So i was gone for the weekend but wow...you lynch sdm, you don't shoot ragnar orrrr stax. dumbest fucking village. i hope we lost cause we deserve it. dumb fucks. im also drunk and high but srsly fuck u

Quote:

Originally Posted by a500lbgorilla

No fears, people fuck up in this game all the time. Just make sure you deflect your failures by reminding people that SDM is somehow worse.

Robb

10-13-2009, 11:13 AM

#14 (permalink)

4-of-a-Kind

Join Date: Aug 2007

Posts: 3,072

Quote:

Originally Posted by StillDeadMoney

Unless you're playing somebody who's randomizing their play. Then again, if you do, you're probably playing at a level where you don't need this forum's help.

A lot TAGG-res randomize lots of things about their play but would rarely have 60/35 stats over 50 HH's. I guess I don't really get what you're sayin', here.

My Operation and FTR Rethread: Stations are *sob* so hard to play against

Extremophile

11-02-2009, 09:10 PM

#15 (permalink)

Flush

Join Date: Oct 2008

Location: Stackton

Posts: 451

Quote:

Originally Posted by Robb

25 hands of 60/25 is enough to be pretty doggone certain he's no TAGG. If you've got that, you're gtg.

Thanks for this valuable info Robb. I don't know if a similar math can be applied for a TAGG. How much of a sample do we need in order to accept someone TAGG?

Robb

11-02-2009, 11:01 PM

#16 (permalink)

4-of-a-Kind

Join Date: Aug 2007

Posts: 3,072

Quote:

Originally Posted by Extremophile

Quote:

Originally Posted by Robb

25 hands of 60/25 is enough to be pretty doggone certain he's no TAGG. If you've got that, you're gtg.

Thanks for this valuable info Robb. I don't know if a similar math can be applied for a TAGG. How much of a sample do we need in order to accept someone TAGG?

The problem w/ TAGG is that we're looking to narrow down the vpip to, say, 15 - 25. With LAGG's/maniacs, we have this huge range of vpip's > 35 or so.

But, usually ~60-75 HH's is enough to guess someone's got TAGG tendencies. You'll be surprised every now and then if a nit gets on heater or LAGG goes card dead.

But only the preflop stats are accurate early. Flop stats and things like AF take a lot longer to be reliable.

My Operation and FTR Rethread: Stations are *sob* so hard to play against

JoeHaw

12-11-2010, 05:05 PM

#17 (permalink)

Straight

Join Date: Mar 2009

Location: IL

Posts: 186

Robb, what function in excel did you use to sum

Quote:

P(X >= 15) = (50 choose 15) (.15^15)(.85^35) + (50 choose 16) (.15^16)(.85^34) + (50 choose 17) (.15^17)(.85^33) + ...

Imthenewfish

12-12-2010, 11:07 PM

#18 (permalink)

Full House

Join Date: Apr 2010

Location: sites im allowed to play on

Posts: 945

wow thanks for bumping

JoeHaw

12-13-2010, 03:32 AM

#19 (permalink)

Straight

Join Date: Mar 2009

Location: IL

Posts: 186

yeah no prob still trying to figure out that excel shit tho

Robb

12-13-2010, 05:13 PM

#20 (permalink)

4-of-a-Kind

Join Date: Aug 2007

Posts: 3,072

Quote:

Originally Posted by JoeHaw

Robb, what function in excel did you use to sum

OK, the binomial coefficient "4 choose 2" gives the total number of combinations possible when selecting, say, 2 Aces from a total of four.

Excel has the formula:

= combin(1st,2nd)

where 1st is the number of things being chosen from, and 2nd is the number of things being chosen. I used the binomial theorem to determine the probability of a unfair coin flip situation. Here's how it works with another, simpler example. Suppose I flip a coin 10 times, but it has a 60% chance of Heads.

Then, the probability of 10 Heads in 10 trials:

P(X=10) = .6^10

The probability of 9 Heads in 10 trials:

P(X=9) = (10 choose 9) * .6^9 * .4 ^1

This can be thought of as lining up 10 slots. The "10 choose 9" selects a space for the Tails (and the other nine for the Heads). The ".6^9" is the probability of Heads occuring raised the 9th power since it happened 9 times, and same for Tails.

You'll see the pattern if I list the next few:

P(X=8) = (10 choose 8) * .6^8 * .4 ^2
P(X=7) = (10 choose 7) * .6^7 * .4 ^3
P(X=6) = (10 choose 6) * .6^6 * .4 ^4 ...

Now, the cool thing here is that these probabilities are EXACT, not estimates. So we can ask the question: "What is the probability that this coin is flipped 10 times, and we see AT LEAST 6 Heads?"

To answer, we simply create a simple spreadsheet that indexes through the 11 outcomes, calculating the probability for each one, then sum up the five in question. Knowing that everywhere I wrote (n choose k), I just used the Excel formula "=combin(n,k)" makes it straightforward, if you're used to Excel.

If not, I created a spreadsheet in Google Docs which demonstrates the above example, and I will post the screenshot later today, with the formula visible. Google docs works in this instance exactly like Excel.

My Operation and FTR Rethread: Stations are *sob* so hard to play against

Robb

12-13-2010, 05:27 PM

#21 (permalink)

4-of-a-Kind

Join Date: Aug 2007

Posts: 3,072

Here's the spreadsheet:

Uploaded with ImageShack.us

My Operation and FTR Rethread: Stations are *sob* so hard to play against

Latest Poker News
KoRnholio	05-26-2012, 03:08 PM Australia Legalized Online Poker coming up in next 6 to 12 Months
According to an email sent out by Mark Bryan, a gaming analyst at Merrill Lynch, the Australian government plans to legalize online poker sometime in the next six to 12 months. This move will coincide ... more at Poker Blogs »

« Previous Thread | Next Thread »

Thread Tools
Show Printable Version Email this Page
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

Posting Rules
You may not post new threads You may not post replies You may not post attachments You may not edit your posts BB code is On Smilies are On [IMG] code is On HTML code is Off Trackbacks are On Pingbacks are On Refbacks are On Forum Rules

All times are GMT. The time now is 08:56 PM.

Play Texas Holdem Online, Online Texas Holdem Strategy, & Poker Forum

This is not a gambling website.

Poker Forum

Over 1,246,000 Posts!

Poker Forum

Beginners Circle

Maniac Math

Play for FREE and practice your game at...