Texas Holdem Hand Analyzer

Not exactly a coin toss

noreply@blogger.com (Jay) — Tue, 01 Sep 2009 17:36:00 +0000

Part of the reason I built the hand analyzer was that I didn't totally trust the odds I would see on the television. Whenever someone turns over Ace - King or Ace - Queen against a low pocket pair they always say that it is a coin toss. Well I decided to run 23,000 hands a see exactly how close it is to a coin toss.

You can see that after my 23000 hands are run, the low pocket pair (in this case 6's) wins decidedly by about 9.5 percent advantage over the Ace - Queen. I also let the two over cards have one of the same suit as the pair.

http://www.potnut.com/pa/screenshots/cointoss.jpg

In my second example, Ace - Queen versus pocket 8's, I decided to not overlap the suits. I ran the simulation again, this time for 50,000 hands. After it is all said and done, with no overlap of suits this time, the 8's are the decided winner, this time by 11.5%.

http://www.potnut.com/pa/screenshots/cointoss2.jpg

So the next time you flip over your small pair to someone's over cards, and they say 'coin toss' be sure to correct them. Take your 10% advantage and win yourself some money with it.

Recompiled latest version of Silverlight

noreply@blogger.com (Jay) — Wed, 25 Mar 2009 17:02:00 +0000

Hi everyone,
I was able to successfully recompile the project using the release version of silverlight, I am having some strange issues with one of my custom controls, where I can't play around with it in design mode, but I can compile it and it works.
I might have to just start from scratch on this control and build a new one if I can't figure it out.
Anyway, the app is working and I would appreciate some feedback. I have an idea of what I would like to see this app do, but it can only get better if you give me some input.
Jay

1.01 Bug fixes, found solution to card size

noreply@blogger.com (Jay) — Fri, 17 Oct 2008 20:52:00 +0000

Hi all,

I fixed a couple of bugs, first of all if two people both have a flush of the same suit, the person with the largest heart in his/her hand will have the higher score, thus will win the hand. Previously it wasn't taking into account other ranking cards in the flush.

Secondly I figured out a solution to allow the cards to be drawn smaller, my next step will be to allow 10 (or 11) players to all be drawn on the board at once.

Also I need to put a table in the background, and fix the summary of the kickers so it doesnt bleed over into another persons hand. This will take a lot of tweaking, but not necessarily a lot of coding.

Comments, as always are welcome.

Jay

Version 1.00 Released today!

noreply@blogger.com (Jay) — Wed, 01 Oct 2008 17:21:00 +0000

Hi everybody,

Well its been a long road to get here, well okay not that long only a couple of months, but a lot of code has been written to get to this point.

Version 1.00 is working and has all the originally envisioned functionality. I only go up to 5 players at the moment, the reason isn't because more players are harder or anything, its just that the size I picked for the cards is a bit too big. You can expect that in a future release the cards will be be better sized, and potentially there will be an actual table not just a green background. Also things like buttons and stuff will moved around, and maybe the interface will change, things will look different. I also hope to continue adding functionality on it.

Percentages are calculated on the fly, and the number of times a player wins a hand based upon a simulation is kept track of, until the refresh button is pressed, or until one of the redeal buttons is pressed.

Please post here, or contact me at getref@hotmail.com.

Added Some Player Tracking Ability

noreply@blogger.com (Jay) — Mon, 29 Sep 2008 22:04:00 +0000

Hi all,

It has been a while since I posted an update for everybody, so I figured I would compile the latest version and post it so you can see what is coming up.

I added a small grid on the left hand side that will be used to track how many times a player wins a particular hand. I'm not 100% sure how this part will work just yet, I think I will allow the user to reset the grid whenever they want, and it will automatically reset when you change the number of players on the right hand side.

Work continues, slowly but surely. I have much more in store for this application.

Work stalled

noreply@blogger.com (Jay) — Thu, 18 Sep 2008 14:48:00 +0000

Well work has temporarily stalled on the hand analyzer. I haven't forgot about it, but a new project just started at work and that leaves me less time to self-teach myself Silverlight.

I posted the latest build up, so everyone can play with it. It isn't much different than the last build, except it allows 5 players now. I think I need a smaller deck of cards if I am going to try and squeeze 10 players onto this table, either that or I need to increase the size of the table. I dont want to do this because it might cause problems for people with smaller resolutions than what I am running.

This is a fun project, and I will definitely work on it more, but for a little while work is going to be going at a slow pace. I need cards with smaller images (I can probably just resize these ones if I need to, in fact Silverlight might have an object that I can use that will resize the image for me.

As always, feedback would be much appreciated. Early feedback makes it a lot eaiser for me to change the application, and the earlier I fix / create something the easier it is to integrate. Late fixes always seem to be more difficult/sometimes impossible without a total rewrite.

Addition of Hand types / Declaring a Winner

noreply@blogger.com (Jay) — Fri, 29 Aug 2008 17:48:00 +0000

Well work is going well on this Silverlight project. There are only two hand types that the application does not properly check right now, a straight and a straight flush. Correction, it does check a straight, but it doesn't assign a score or a high card associated with it so the application would not be able to tell who won the hand if both hands had a straight.

I expect to have both of these algorythyms (straight flush and straight) completed soon, and have already tried adding a player to the board and my logic seems to work good for 3 players. I will try adding 4 - 10 hopefully soon as well.

I wanted to post a code update for everyone so that they can see the progress so far. Projects like this have a tendency to suck up a lot more time than they initially feel like they should take. This project is no exception to the rule.

If you find any places where the code seems to be acting in a way that it shouldn't then let me know. Please be very descriptive as to how to replicate the problem/error (step by step process).

Work Continues

noreply@blogger.com (Jay) — Mon, 25 Aug 2008 21:15:00 +0000

I have added all the basic types of hands now, but am missing high card, and there is only calculation for kicker done currently on two pairs.

The maximum value a given hand for score is now 141,414,141,414. I've gone over the logic many times and I think that this system for score will cover all the possiblities. If not I may have to rework the way score works at a later date.

Determining which hands wins the hand will be done as such. First, the types of hands are ranked from top to bottom, with straight flush being the highest valued hand, and high card being the lowest.

The logic works as such: All hands face off, the highest type hand wins outright, if there are two or more hands with the same hand type, that is the point that it goes to the score.

I was going to incorporate the hand type into the score, but looking at it I think I would require too long of a integer, also this sort of locks me into always combining those two values into score, and there may be some reason to split them in the future.

Each card is assigned a value equal to its Rank 2=2points, 3=3points, .. King=13points, Ace=14 points.

If two or more hands have the same hand type (ie: Two Pair) then we go to a showdown based upon score. Every hand has a top card which is always worth more than the bottom card. In the case of two pair, the top (highest) card is multiplied by 10,000,000,000 and is added to the score. So if someone had 4's versus someone else's 3's, they would have a score of 40,000,000,000 versus 30,000,000,000.

Lets say they both have 2's as their second card, since 2's is the low card in the two pair scenario, we only multiply it by 100,000,000.
This leaves us with 40,200,000,000 and 30,200,000,000 for scores.

Aces and Queens versus Aces and Kings. 141,200,000,000 < 141,300,000,000

When we go to add the first kicker, the first kicker will occupy the 5th, and 6th position of our 141,414,141,414 numbers.

Ace/King with a Jack Kicker versus Ace/King with a 10 Kicker.
141,311,000,000 > 141,310,000,000

I think this works for all scenarios of two pair.

Things are going well so far, many more things need to be done but the algorythyms for determining the score value of hands, which will be used to determine the winner of the hands are well underway.

Check for Quads, Full house

noreply@blogger.com (Jay) — Wed, 20 Aug 2008 20:50:00 +0000

I am re-working some of the code, and I created a couple new classes to help me determine if someone has some of the rank based hands. Checking for four of a kind was relatively easy, checking for a full house was a little tricky.
I expect checking for three of a kind to be easy, and two pair to be a bit tricky.

Regardless, four of a kind and full house are now being checked for. You will also notice a score that has been attached to each hand.

The score will be used, in combination with the type of hand to determine how good the hand is. So a four of a kind with a score of 13 would still beat a full house of score 1013. Four of a kind with a score of 13 should be 4 kings, full house of score 1013 shoudl be 10's over kings 1310 would be kings over 10's.

The app is coming along nicely, comments are welcome.

Turn and River

noreply@blogger.com (Jay) — Wed, 13 Aug 2008 21:52:00 +0000

I decided the next thing to get working would be to get the turn and river working. Also I put in some smart enabling/disabling of the buttons to help focus the user on what he/she should be allowed to do from that point on.

I have managed to write this application in such a way that it should (*fingers crossed) be relatively simple to start adding logic to test for the other hand types, also I made it so that it should (*) be easy to add more players as well.

My goal is to have it so that you can run a bunch of test hands to see how often one will will win over another hand. You can change any of the user's hands to whatever you want, and you can run simulations to see what the result will be.

First Pre-demo release

noreply@blogger.com (Jay) — Tue, 12 Aug 2008 16:46:00 +0000

Welcome to the Silverlight Texas Holdem Hand Analyzer project. Eventually I hope to have a better name for the project :).

I started this project as a way to learn the Silverlight architecture, and while playing around with Silverlight I thought that maybe it would be a good idea to build something that people could use.

It is a ways away from completion, but the basic engine is working, it currently checks for straights, and single pairs. And it will deal out hands to 2 players, and also deal the flop.

Clicking on a card will move up to the next ranking card, unless a player already has that card.

What is the purpose of this project?

I'd like to create a easy to use web application that will allow a user to quickly and easily determine their odds in a given Texas Holdem situation, either before it happens or retroactively when looking back at something that happened, eventually it would be nice to have this application available on a hand-held internet ready device.

I would love to read your comments, your comments will greatly influence which direction this project heads.

You can see the very first version of the Hand Analyzer HERE