Large HandHistory for microlimit FL neede

    • Laegooose
      Laegooose
      Bronze
      Joined: 01.09.2008 Posts: 5,496
      Hello!
      I've implemented an algorithm based on white FL artictles, and I'd like to improve it.
      If I have handHistory, I parse it and compare decision of my algorithms and hero's decision, so I find situations in which algorithm works badly and change it.

      Could someone send me large well-played HH for microlimit FL at PartyPoker (in PartyPoker format)? Coaches' histories would be very nice!

      Also, did anybody tried to implement an algorithm or are there any articles/books in this field?
  • 4 replies
    • ciRith
      ciRith
      Bronze
      Joined: 25.03.2005 Posts: 18,556
      Hi Laegoose,

      are you trying to programm a bot?
    • opal99
      opal99
      Black
      Joined: 05.02.2008 Posts: 8,270
      Originally posted by ciRith
      Hi Laegoose,

      are you trying to programm a bot?
      it looks like... :rolleyes:

      Anyway; you should try SSS.. It's simple and can be profitable on micro stakes even with non-perfect bot - I wouldn't do that, because grinding 30/60 SH is definetely more profitable than microstakes bot (no matter if it's FL or SSS) ;) And getting banned on big site is always bad....
    • Laegooose
      Laegooose
      Bronze
      Joined: 01.09.2008 Posts: 5,496
      Well, at the moment I have no idea how to make an interface; also I dont want to be banned at PartyPoker. Maybe it will be useful at analyzing my own game. Anyway, I think it is intresting challenge to write a "bot" who will play not awfully :)

      NLH is much more difficult, because (at least basic) hand-reading is very important there. Also, there are strange situations like mini-raises and overbets. Also, mistakes in NLH can be very costly - you can lose whole stack because of wrong decision.
    • Alafoe
      Alafoe
      Bronze
      Joined: 28.08.2008 Posts: 202
      To test some algorithm on large sample of data is not creating the bot. In the articles are anyway clear algorithms. IE SSS is pure algorithm not strategy at all. So to know wheather given algorithm should work is good idea. You can also evaluate profit when using another aproaches. So this is good idea and i definitely want to test some algorithms and prove their ability to make someone winning player. This area is not largely investigated...