How would Wilson turbo do in a real game?

Turbo Texas Hold'em software seems to be the ultimate repository for playing strategy, though it has the disadvantage of responding mechancially to situations.


What do you think of the quality of their strategy?


If the computer sat down at a low-stakes table in Vegas, how would it do?

Turbo Texas Hold'em software seems to be the ultimate repository for playing strategy


This is very inaccurate.


The Turbo program would almost certainly lose in any ring game. I've played with it a few times and was unimpressed. I own the Stud software and it plays very weakly.

TTHE's best profiles have a serious flaw against all lowlimit tables, which cause them not do as well as possible. There are many cases where it will do better against a tougher player lineup because of this flaw.


Anyone know what the flaw is?


mr big

It puts in too many bets before the flop instead of seeing the flop cheaply.

I've never been check-raised in TTH. At least not that I can think of.


Is this it?

TTH is ridicolously overaggressive in its "tough" lineup. It will almost constantly 3 bet the flop with overcards if heads up (when checkraised), and never take a free card on the turn. It is totally oblivious to slowplaying. It doesn't checkraise enough. It will never induce a bluff, and will almost never lay down a big pair.


It has its uses, but it would get killed even in a loose passive low limit live game.

Answer:


It doesn't put in too many bets preflop in my opinion, it is one of the things it does well, if you take Lash or Conan, they play about 17% preflop and they play aggressively and well.

Most low limit players would love to play as well preflop; in fact most good players don't play as well, since this is an area which computers can lookup very well, and Wilson's algorithm is excellent.


It DOES check raise, especially Lash and other strong players do, in fact there are 4-5 rules for this (CR, CX, CC, DR,...). It is very good in these checkraising kinds of situations, both to make more $$ and to thin the field. It is way better than most very good human players at this.


What it does very poorly when playing against a loose group of weak players is to go too far on the flop without excellent cards. It is a huge flaw. Here is an example (maybe newer versions have fixed this):


It has KTs in the SB. There are 6 preflop players, all relatively weak loose passive. Flop comes AT8r. It will bet and rebet it's seconds pair with 6-7 loosies, where it is beat 95% of the time cause someone has an A or T8 or 88 or something else. In this situation, it is getting killed when it could fold or even hope to get a free card on the flop and then bet when a T or K come out on the turn.


Mr. big

Right you are. My custom LASH+ doesn't go crazy heads up with overcards. Big mistake. Or not-heads up either. I slowed it down in the TTHE tables. It also CR's much more (although the latest version does this much more and varied than before). And the new rule for HU calling instead of blindly folding or calling irrespective of how many, makes it much better. It IS missing another rule which would help: I would call it Z2: it would bet if headsup and call if more than one. This is a moderate oversight.


Mr big

I agree, ridiculously overagressive. I seemed to remember that this also applied to pre-flop play with hands that do not play well multiway, but perhaps I misremember, I haven't played it for awhile. I know it's true after the flop with weak hands.

Wilson software are like tread mills. They are good for you, but you know you really didn't run three miles. Where were the birds and the rest of reality.