Heliotrope

    • grabek1985
      grabek1985
      Bronze
      Joined: 02.11.2008 Posts: 326
      This will be a blog showing how I approach poker hand analyses. I will try to participate to poker strategy related discussions, thus intesify the growth of my learning curve. Finally, I will share my own views regarding the strategy building process itself. I play nl1k+.
  • 3 replies
    • zumpar
      zumpar
      Bronze
      Joined: 05.02.2012 Posts: 1,185
      sounds awesome, GL sir!
    • grabek1985
      grabek1985
      Bronze
      Joined: 02.11.2008 Posts: 326
      Nl2 J5
      nl10 SB KQo
      NL10 AQo Overvaluing Top Pair
    • grabek1985
      grabek1985
      Bronze
      Joined: 02.11.2008 Posts: 326
      When analyzing a game of poker I am often bumping into kmeans related approaches to conduct a singular value decomposition in a different than a standard one manner. I am always thinking as how to decompose the strategy into matrix factorization problems.

      Again, born to look at the bots. I took a part in a competition where we had to prepare great bots for bots. It felt like you need more than 100m hands to be almost perfect. A simple decomposition is like intrinsic truculence, not even close to what really needs to be done.

      When playing with my bot felt deprived of strength, so made yet another convolution to defeat the former one with slightly skewed strategy for medium strength hands (in different types of spots). That worked. What a relief. The first bot won, so my "initial" ideas were not good enough to play with Nash.

      A bot is a pariah among the people who would never want to lose. It castigates its opponent with this sense of humour. A bot is not a misanthrope, he is just refractory enough to cluster big enough a data.