Studying two okay players and two fishes(long)

I created a simulation in VB to get some insight into some questions that I had and thought I would share the results. If anyone has any comments on how I could make this more useful it would be greatly appreciated.


I basically programmed a full ring game with 1.00/2.00 bets, complete with blinds(.50/1.00), moving button, rake, and different AI in each spot. I forgot to add the tip. Here is the setup.


Preflop


Players 1 and 7 are intelligent players, taking into account cards and position according to one of Abdul’s writings

Players 9 and 10 are fish and always see the flop

I set all other players up with a strange algorithm that I have seen on low stakes online poker a lot. They love all pocket pairs, and like flushes(two suited cards), and like two cards bigger than ten. (Will always raise loved cards, and will call with liked cards)


PostFlop – All pretty much play the same. Fishes will call one bet, top-pair/better players will bet, and four-flush/str8’s will call.


Turn – One fish will call a bet regardless. All other players will follow the same post-flop criteria.


River – Top pair or better bets, everyone else stays at this point (I should probably change this).


Results: I first wanted to look at the standard deviation of players 1 and 7(who I consider to be the best players). Here is how they fared


Player 1 :

Total over 2000 hands - +420.5 units (A big bet is 2.00) (10.51 units/50 hands)

Standard Deviation for each 50 hands : 28.70 units

Biggest 50-hand win – 71.7 units

Biggest 50-hand loss – 43.2 units

Standard Deviation for each 100 hands : 35.67

Biggest 100-hand win – 77.3

Biggest 100-hand loss- 43.1

Low point -43.1 units

High Point +420.5 units


Player 7:

Total over 2000 hands - -24units (A big bet is 2.00) (-.6 units/50 hands)

Standard Deviation for each 50 hands : 24.9 units

Biggest 50-hand win – 60.9 units

Biggest 50-hand loss – 47.8 units

Standard Deviation for each 100 hands : 24.62

Biggest 100-hand win – 54.7

Biggest 100-hand loss- -35.2

Low Point -91.7 units

High Point +16.7 units


Average Players Results After 2000 hands

2 +.5 Units

3 +359.7 Units

4 –128.1 Units

5 –37.6 Units

6 +223.8 Units

7

8 –14.1 Units

Could this be a result of position? This is very odd to me.


Fishes

9 –1612.3 Units

10 –2007.6 Units


Rake 2819.2 Units


Net/Gain Loss after 7,000 hands

1 +1296.7 units

7 +732.4 units


2 +588.8 units

3 +501.3 units

4 +437.3 units

5 +633.6 units

6 +1072.2 units

8 –597.9 units


9 –5365.3 units

10 –9111.9 units


Rake 9812.8 units


Findings: The standard deviation findings were somewhat interesting, but this is more of a testament to game selection than anything else. Amazing how much two fishes can pay off. I was also amazed out how well, depending on position, the standard strategy worked.


I also took some comfort in seeing that a player with a clear advantage was actually breaking even (losing 12 big bets)after 2000 hands(between 45 and 70 hours)

Any comments would be greatly appreciated.


Thanks much,

-Jim

I'm somewhat skeptical of the realism your simulation enjoys, but this is nevertheless educating.


-MD

Yes, there is a lot to be desired, but it is being improved. The original goal was to determine the swings of a superior player, especially with many flush and straight draws. I believe this has been shown. My "list" to make it more realistic includes the following.


-Having players throw away missed draws on the river.

-Tipping

-Adding a more aggresive player

-Having players raise draws for value

-Having players stay in with middle pair/an overcard

-having certain players try and steal the blind

-having certain players defend

-having certain players retain a "respect" value for all other players. This "respect" value will be used in the check/call/raise functions.


Just some ideas, I would be interested in any other additions people would think were useful.


Thanks again,

-James

Cool ! I think you can learn a lot from this kind of work if you know what you are doing.


But think about Delphi, it kicks VB's ass !


Andy.