Introduction

I have been playing poker for quite some time, most of the time just for fun and a few eurocents... I especially love playing MultiTable SitNGo's (i.e. small tournaments which start when a certain amount of players has registered).
I am quite successfull playing this kind of games at lowstakes, at least I usually manage to get into the top 5% around halfway the tournament. But then the following usually happens (fictiuous situation):

- I get a premium hand (AKs for example) and raise 2-3 BB...
- I get called by a bigstack on the (BB)
- Flop, low card rainbow
- Bigstack checks
- I raise 2/3 pot
- Bigstack calls
- Turn: K
- Bigstack checks
- I raise POT
- Bigstack calls
- Some lowcard falls
- Bigstack goes all-in
- I call (I know I shouldn't, but I'm pot-commited)
- bigstack show something like 2-7os, but hits a double pair, where he hit the 2 on the flop, and the 7 on the river.
- I get busted out of the tournament (usually just before the bubble)

It's a leak !!!

Yes, now that I am reading it, I know what I should do: fold after the all-in.
I also should not have raised after the flop, maybe just a check-call if the investment is not to big.
I know I can't go all-in directly, since it's a bigstack that calls anything with rags (i.e. donk). He usually only folds on the river if he hits nothing at all (and maybe after value betting first)....

In the long term/purely based on statistics, I would probably beat him with AK. The problem is that in this phase of the tourney, in a comfortable position, I should play a whole other post-flop game, since staying in the tournament is simply more valuable then gaining a few chips
However, most of the time this happens just before the bubble, and I played a whole game just to get busted before the bubble...

And this happens to me a lot !!

So what should I do ?? Fold if the bigstack plays ?? Maybe I should, maybe I shouldn't.... but this got me thinking....

Fix the leak !!!

What if I could write an app that analyses these things, and gives me (based on statistics) the best possible solution for this situation..

I know these allready do exist, but since I am a developer and I love playing poker (even though I am really bad at it), I am going to write my own app for this... And while doing so, I will also blog about it, in order to get the discussion going and find the best possible solution for the problem.

Where to start ?

Since this is a whole new ballgame for me, I will probably need to delve deeply into statistics, fuzzy logic, some neural networking, and probably some datamining tools as well...
I will blog about this every now and then, and also add a link to my task list:

Tasklist

• Write the initial specs for the app
• Get a huge MTT SNG hand history for training purposes
• Find out how the winners got ITM
• Build some kind of system which matches the ITM players as much as possible using genetics
• Refine, refine, refine...
• Can I create a marketable product out of it ? How about open sourcing it ?? Which platform should I target (i.e. mobile, web, PC, ...)
• Do something which will result in "World Piece™"

Finally

This is supposed to be the first post of a long series, and since it is a hobby project, I have no idea whether and when I will finish it...

I hope you enjoy it while it lasts !!!