Comments on Senzee 5: 7

Wow, I'm writing a poker bot, and found myself in ...

2009-02-02T03:25:00.000-05:00

Wow, I'm writing a poker bot, and found myself in need of exactly this! I have a table that has a 4 byte rank for each possible 7 card hands. (so about 500MB)....in reality i only need 20 bits per rank, but i figure 32 will be quicker. I read up on combinadics and was bewildered, but your code works!!!!! I tested it by feeding it every possible input and confirmed it genereated every possible output exactly once. THANK YOU.

Fantastic post. There's a buildable version of thi...

2008-09-06T05:53:00.000-04:00

Fantastic post. There's a buildable version of this evaluator over at http://www.codingthewheel.com/archives/poker-hand-evaluator-roundup (code to build the 266MB lookup table from index52c7.h and so forth). Paul: any chance we'll see some more poker-hand-evaluation goodness out of you?

Hi Paul,thats a really good work!I'm really intere...

2008-05-25T13:10:00.000-04:00

Hi Paul,

thats a really good work!

I'm really interested on your 64-bits minimal perfect hash generator. Is it possible for you to publish it?

Thank you very much!!

Hi Paul,thanks for your hard work!your fasteval.c ...

2008-05-12T12:36:00.001-04:00

Hi Paul,
thanks for your hard work!
your fasteval.c was really nice!!

but i´ve got a problem with your index52c7.h, i just dont get how to use it, what should I do with the result from the index52c7 func??
could anyone mail me an expample of how-to use this func?? thanks for your time. mail: phazz@gmx.net
with best regards.

Hi Paul,thanks for your hard work!your fasteval.c ...

2008-05-12T12:36:00.000-04:00

Sorry, meant to say 4 bits for a suit in the last ...

2008-05-11T00:45:00.000-04:00

Sorry, meant to say 4 bits for a suit in the last section. So binary & with this number for suit: 1000.

Here's an idea which might be even faster than com...

2008-05-11T00:43:00.000-04:00

Here's an idea which might be even faster than computing index52h7, and also requires smaller arrays.
First, we will check for flush, straight flush or four of a kind (the method will be explained later). So assume that there is no flush, sf or four of a kind. Consider a representation for storing the counts of each denomination. Two bits per denomination, for a total of 26 bits.
Each card is now represented as a count of 1 for that denomination. For example, a four: 01 00 00.
Now take the sum of the seven denominations. The maximum index is 60,817,408.
Now, how to check for flush, sf, or four of a kind. Consider a somewhat different representation of each card. Each denomination will have 3 bits devoted to its count. Each suit will have 3 bits devoted to its count. So for example, king of clubs: 000 001 ...11 groups of 000... 000 000 000 001. (here clubs is the smallest suit, and denominations are ordered normally). Now add the seven cards, and this number: 011 for each suit, 000 for each denomination. Finally, bitwise & with 100 for suit and 100 for denomination. The result is non-zero if and only if there's a flush, sf or four of a kind.

Hi,First of all, thank you for your code.I am a bi...

2008-02-23T06:16:00.000-05:00

Hi,

First of all, thank you for your code.
I am a bit at a loss as to how you chop up the 52-bit number. I suppose you split it up in 4x16bit numbers and then do a sort of representation "packing" with _bitcount and indexing. I cannot see through the choosearrays. What do they represent exactly?
I thank you in advance for your time.
(stellaferox@hotmail.com)

Hi, I am writing a Texas Holdem hand evaluator for...

2008-02-03T11:50:00.000-05:00

Hi, I am writing a Texas Holdem hand evaluator for my pocket pc, just for learning pusposes. I want to make a table from the different possibilities and want to know how you (using Cactus Kevs' code) evaluate your hand (rank) for 7 cards. Thnx in advance.
Marc

Hi, I don't know about using such a large table fo...

2007-10-22T05:12:00.000-04:00

Hi, I don't know about using such a large table for determining hands, but when I tackled this problem I wasn't limiting myself to a possible number of cards. I wrote a peice of code that takes any number of cards and spits out an integer representing the highest possible hand. When you are doing this hand calculations what do you include for benchmarking? I found that the slow part of my algorithm is loading the cards, not calculating the hand value, so do I include that time in the evaluation? I'm new to this and my code was originally ment to run in a game which didn't need speed as a high priority, any advice on how I can benchmark this? Thanks much. (hillplace@gmail.com)

Is the source code for your ranking system availab...

2007-09-08T04:28:00.000-04:00

Is the source code for your ranking system available anywhere? I'm trying to play with a few things involving hand ranks and having trouble finding decent ranking code. =)

The Paul above is not the Paul who wrote that code...

2007-08-15T09:57:00.000-04:00

The Paul above is not the Paul who wrote that code, because *I* am the Paul that wrote that code. :) You will also find a 5 and 7 card poker evaluator on my site as well:

http://www.azillionmonkeys.com/qed/poker.zip

I haven't bothered to benchmark it against anything, so I don't really know how good it is, but I thought you might be interested in checking it out.

P.S. Glad you like my optimization page.

Hi all,I have written a 5, 6 and 7 card evaluator ...

2007-08-03T12:02:00.000-04:00

Hi all,

I have written a 5, 6 and 7 card evaluator in PHP (a language certainly not known for it's speed). It is the simplest evaluator I have ever seen and is made up of about 10 lines of very basic code. It takes 16 microseconds (not milliseconds) to score a 7 card hand.

I have used it in my online Texas Hold'em probability calculator which can be seen at http://texas.holdem.poker.probability.cal.culator.org/

My guess is in C++ this evaluator would beat any other for speed, but I don't know C++.

I'd love to get in touch with anyone interested in the code or interested in porting it to a CLI in C++.

Anyone interested can email me at http://bokehman.com/email

Paul, thanks for the link!Assuming that you're the...

2007-07-27T16:51:00.000-04:00

Paul, thanks for the link!

Assuming that you're the same Paul that wrote the SuperFastHash article, let me tell you that I love your site - especially the optimization section (http://www.azillionmonkeys.com/qed/optimize.html). You've clearly put an enormous amount of work into it and it's a phenomenal resource!

If speed is the goal, you might want to give Super...

2007-07-27T15:49:00.000-04:00

If speed is the goal, you might want to give Super Fast Hash a shot; it's supposed to be faster than Bob Jenkin's algorithm.

For reference -Combinadicshttp://en.wikipedia.org/...

2007-07-12T01:52:00.000-04:00

For reference -

Combinadics

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

Are you still working on this?Is Jing's java avail...

2007-06-29T17:24:00.000-04:00

Are you still working on this?
Is Jing's java available?
I will check back for comments since I don't have an account.

Hi Jing, Thanks for the head's up. I'm checking ...

2007-02-05T12:36:00.000-05:00

Hi Jing,

Thanks for the head's up. I'm checking it out!

There is a discussion for a super fast algorithm a...

2007-02-04T01:05:00.000-05:00

There is a discussion for a super fast algorithm at 2+2 forum:

evaluator algorithm

I implemented it in Java. 1 sec for 133 M hands. C or C++ only 200 or 300 msec.