Saturday, December 27, 2008

Game Selection Response

My friend Rory emailed me a good response to my last post:

I think this might be a little bit backwards. Lets say you are sitting in a 6-max game with you, two good players and three bad players. The good players are playing let's say 30% of their hands each, as are you. The two bad players are playing 70% of their hands each. For each hand you play, what is the chance you will be in the hand with at least one good player if you decide to play? It should be 1 - the probability neither of them play. 1 - (0.7)^2= 51% so half of your hands will be vs one of the two other good players. What is the chance that you will be playing against one of the bad players? It should be 1 - the probability none of the three play: 1 - (0.3)^3 = 97% that means the majority of your hands will have at least one bad player in them, and only a very small percentage of the hands you play will be just you vs the good player. What if there are only two bad players? You then have a 66% chance of playing against a good player in a hand. What about the bad players? You have a 91% chance of having a bad player in the hand. So again, only a very small percentage of the hands you play will be just you vs one or more good players only. What about 4 good players? Then it starts to be even, 76% of the time a good player is in the and and 70% of the time the bad player is in the hand. The other problem with the GR-avoidance method of game selection is that just because you identify a no-GR table doesn't necessarily mean there is anyone that plays bad in the game. If you do the reverse and follow around bad players, you guarantee that no matter what you are going to be about 50-50% seeing flops with a really bad player even if all the other players in the game are good. Also just because there are two or more GRs in the game doesn't meant it's a bad game, in fact a table might be especially juicy if there are two or more GRs in the game; maybe they know something about the other players that you don't!


As you always do, you've made a bunch of good points. I think you're looking at this from a different perspective than me. You're trying to play as many hands against bad players as possible (which makes a lot of sense). Essentially what my method is accomplishing is finding tables where everyone minus the GR at the table is at least 1 BB/100 (my win rate) worse than me. If it does this then I should almost always do better than my 1 BB/100 average.

I need to think more about what you're saying and make some changes based on it.

I actually started looking at PTR's own table ratings yesterday. I noticed that I was sitting at almost all of their top rated tables every time I checked which made me wonder if I might just be better off using their ratings since it is less time consuming than my method. I'm trying to figure out how their table ratings are calculated. I got this from their FAQ's:

" uses a massive database of compiled player results to assess the quality of play at online poker tables. The players at each table are evaluated based upon their VP$IP, long term losses vs. long term winnings, and the total amount of data recorded for the players. Each player’s data is then compiled to provide a composite rating for that table. If the rating is high, you can be assured that you will be playing at a soft table with huge potential profits. If the rating is low, you can avoid that table and spend your time in a more profitable game"

Which really didn't help me. I emailed them yesterday and asked exactly what goes into their calculations and unfortunately received a canned response back that didn't answer my question. So, I sent them another email today. I'll post their answer if I get one.

Thanks again for the ideas Rory.

No comments:

Post a Comment