2+2=????

[DcifrThs]

Heres another easier brain teaser...

imagine for a moment that instead of writing books, S&M could list all the the permutations of events and what to do in every case and some idiot savant memorized and executed it...

i.e. 1a,2b,3c,4d,...Nn where the numbers (1,2...N) indicate known perfect information facts like david clark raised under the gun, barron 3bets and clark folds [img]/images/graemlins/wink.gif[/img] and the letters (a,b...n) indicate distributions of past actions and whatever else varies (they can be vectors) but are reasonably well known.

now, this idiot savant knows every single response to that entire set and never makes a "mistake" (never tilts, never deviates from the response set)...and let say we put him in a mid limit live game...do you think he will earn more or less than somebody like david, or mason, or any player like that? and by how much?

basically i'm asking, how much is off the cuff creativity and ability to think "outside the box" worth in terms of win rate?

lets assume our "pros" make 2.5bb/hour in that good mid limit game. what would our idiot savant make? or the other way around...what if our savant makes x/hour? %-wise, how much more or less would our "pros" make?

-Barron

[TStoneMBD]

i dont think this riddle is fairly thought through as there as some uncovered variables yet to be deciphered. for instance, he can remember how baren and clark played a specific hand, but was this hand correctly played? were there other ways of playing it that members of the forum argued over? how can this idiot shavant distinguish what is correct and what is incorrect? has the shavant come across the answer to every possible situation that can be presented to him in a holdem game including the rarest of circumstances?

clearly, if a player learns everything that every player in this forum has learned, his win rate would be far greater than all any player here. creativity is important when sitautions are presented that have not yet been answered, but out of every post ever created on these forums how many of them could there be? this player would be able to make his play deceptive as well, because he would clearly be able to memorize every possible way to play a particular hand, after reading every response to a topic. pros here as liable to human error, memory lapses and even tilt, so your machinistic shavant clearly is a great advantage.

[Duke]

Idiot savant breaks them all.

~D

[rtucker5]

Barron,

No matter how you try to theorize it, 2+2 still equals 4. [img]/images/graemlins/tongue.gif[/img]

[chris_a]

If you assume this strategy to be optimal in some long-term sense against a typical middle-limit group, then you will definitely win a heck of a lot of money. However, answering in terms of quantitative win rate is nearly impossible. You'd have to have someone like this idiot savant in the same game as one of the pros and have lots and lots of hours logged. I'd venture to guess this have never been even close to done.

I think, in a middle-limit game, the idiot savant will have a very close win rate to the pro for the following reasons:

1) I assume the pros aren't idiot savants and don't have the optimal strategy memorized and they make mistakes. This will cost them overall giving the idiot savants a slight edge in this department.

2) You shouldn't really adjust to different players that much unless they are very unconventional and you are certain of them making errors. The fact remains that if you make suboptimal (game theory-wise) decisions to capitalize on their mistakes, you have to be sure that they will make the mistakes, otherwise they can punish you for deviating from the optimal strategy.

3) How often to expert plays that separate pros from optimal idiot savants really come up in middle-limit games anyway?

In summary, in a middle or low limit game, a pro will probably have only a slightly higher win rate than the idiot savant, if at all.

In tough high-limit games, I think that the pros advantage will become accentuated as compared to the idiot savant's because typically most of the players are solid and a small number are weak. Therefore it becomes more important to adjust to players. Here the idiot savant will definitely earn from the weak players, but not nearly as much as the pro will.

No numbers though sorry...

[shummie]

"You shouldn't really adjust to different players that much unless..."

unless the other player is the Idiot Savant. I mean we're talking about a computer here, right? The only way Gary Kasparov could beat Deep Blue at chess was to do some meta-thinking about how Deep Blue was playing (chess players do this for human opponents too, but a lot of assumptions had to be changed since this opponent was a computer).

Since by definition this Idiot Savant plays the same way in the same situation, he will play very similarly in similar situations. Any good player would pick that up.

So... in order for your Savant to be good against a player who knew his modus operandi (or was just a good observant opponent) the Savant would have to be able to "change gears" just like we all do. This might be done if his "equation" involved some kind of "player remembers how I played this situation last time" or "player saw me play in a situation like this earlier" type of variable.

Still... the problem with computer controlled game players is usually that they are too perfect, meaning they become predictable to observant opponents. Especially ones that know how such a computer might "think".

On a similar note, I hear there is a pretty good heads up hold'em playing computer program out there. If I get unlazy, I might find a post a link. I believe that "changing up play" is very important in that programs coding.

Peace.

- Jason

[TStoneMBD]

neverlose is clearly a savant, and hes better than almost anyone here.