|
|
#1
11-13-2005, 06:14 PM
|
|||
|
|||
top 50 holdem hands - simulation
I've been trying to strengthen my preflop abilities, because i feel that I recently have been slacking in this area. I began to do some reading, and then I thought I'd like to do simulation, since I would get a better idea of relative hand strenght and equity values.
This simulation was performed 2.5 million times with various "player types" including Solid, Fish, Semi-Loose, and Rock. I did the runs at 0.50/1.00 limits with a 5% rake up to $3.   There are a few things I found interesting. 1. Most of the profitable hands are worth *less* than the small and big blinds combined. Raising in last place with hands worth less than the blinds and then only winning the blinds is a good thing. Therefore, open raising with any 2 cards seems to pay off (if the blinds are likely to fold a good chunk of the time). 2. Of the suited Aces, A3s and under seem to be the least profitable, losing about 0.01BB/Hand 3. KT and AT around about breakeven hands. In the "wrong hands" I would imagine these hands can lose a bit of money when they flop top pair, but are outkicked. 4. This one surprised me the most, since I'm still learning about this complex game. 99 makes more money and win's more often that KTs, QJs, A9s, hands that I almost always raise on the button after a bunch of limpers. I'm beginning to raise 99 much more now. I still haven't the courage to raise from the BB (am I losing money?) 5. AJo is the 20th starting hand, and sits below JTs. 6. hands like A8s and T9s have equities of about 22-23 percent. Could that mean a raise in last position against 4 limpers show a long term profit? 7. Since T9s is profitable, I'm surprised 98s seems to lose money. The both make 4 straights (with both hole cards) and T9s makes a flush only one pip higher, which i don't think can make *that* much of a difference. 8. I would think that hands 34 to 50 are playable, (and of course profitable) to play for 1/2 bet in the Small Blind. 
|
|
#2
11-13-2005, 06:18 PM
|
|||
|
|||
|
Re: top 50 holdem hands - simulation
this is kind of neat, however i think it doesn't take into account betting or folding, right? so you can't raise pf, bet the flop and take the pot here. instead you see all 5 cards, correct? i think that makes a difference.
|
|
#3
11-13-2005, 06:28 PM
|
|||
|
|||
|
Re: top 50 holdem hands - simulation
I can appreciate these kinds of simulations, as 1) they are intersting, and 2) can be enlightening if you interpret what they mean properly. Thanks for the post.
That being said, there is no way, for example, that 77 is a losing hand if you play it well. I think nearly every pocket pair should be in the black, but I don't even see anything below 77 on this chart. Easy criticism to make though, admitedly.
|
|
#4
11-13-2005, 06:42 PM
|
|||
|
|||
|
Re: top 50 holdem hands - simulation
It takes folding/betting/raising into consideration. Player profiles were used to simulate a more real game. "solids" won the most money, followed by "rocks", "typical" lost money, "fish" lost the most.
|
|
#5
11-13-2005, 06:46 PM
|
|||
|
|||
|
Re: top 50 holdem hands - simulation
[ QUOTE ]
It takes folding/betting/raising into consideration. Player profiles were used to simulate a more real game. "solids" won the most money, followed by "rocks", "typical" lost money, "fish" lost the most. [/ QUOTE ] i dont see how that's true if a hand like 77 or 66 isn't even on it/ playable, like greg said. I'm a winner w/ all my pocket pairs and I think all/almost all good winner players are.
|
|
#6
11-13-2005, 06:49 PM
|
|||
|
|||
|
Re: top 50 holdem hands - simulation
coheed = +EV
|
|
#7
11-13-2005, 06:50 PM
|
|||
|
|||
|
Re: top 50 holdem hands - simulation
[ QUOTE ]
[ QUOTE ] It takes folding/betting/raising into consideration. Player profiles were used to simulate a more real game. "solids" won the most money, followed by "rocks", "typical" lost money, "fish" lost the most. [/ QUOTE ] i dont see how that's true if a hand like 77 or 66 isn't even on it/ playable, like greg said. I'm a winner w/ all my pocket pairs and I think all/almost all good winner players are. [/ QUOTE ] So am I. I've made tons of cash with my little pairs. I also found it strange to see 66 and 55 missing. They've got to be in among the top 40 hands or so.
|
|
#9
11-13-2005, 07:39 PM
|
|||
|
|||
|
Re: top 50 holdem hands - simulation
Can you say more about your simulation methodology?
Being a software developer I've thought about this sort of thing a lot. I've read the papers by the Alberta guys (who now sell Poker Academy Pro) and written my own simple bots to plug into Poker Academy and simulate various things. I think the key point is that once you get past the highly profitable hands, the EV is highly dependent on the quality of the postflop play. In a simulation, this means the quality of the bots (simulation role players, whatever). And the assumption that if all the players in the simulation are equally bad, the results will sort of "scale" equally across all hands is probably naive and false. Thoughts? See http://www.cs.ualberta.ca/~darse/Papers/AIJ02.pdf Especially section 5.1
|
|
#10
11-14-2005, 12:09 AM
|
|||
|
|||
|
Re: top 50 holdem hands - simulation
[ QUOTE ]
Can you say more about your simulation methodology?... ...EV is highly dependent on the quality of the postflop play... [/ QUOTE ] Good point. However... [ QUOTE ] And the assumption that if all the players in the simulation are equally bad, the results will sort of "scale" equally across all hands is probably naive and false. Thoughts? See http://www.cs.ualberta.ca/~darse/Papers/AIJ02.pdf Especially section 5.1 [/ QUOTE ] Would you mind expanding on this point? I don't think I get what you're trying to say. (I didn't click the link --- yet)
|
 |
|
|
Linear Mode
