Guys and gals, first of all, welcome!

The following idea has always been on my mind, but recently after reading some finance related articles about collective intelligence being very powerful, I've decided to give it a go on this forum.

Suppose we design a pokering software where two accounts are playing heads-up. One account is being operated by, say durrrr, the other account by 2000 pimple-faced, greasy haired 2+2 members. Every action receives input from these 2000 players, and the most frequent one is than chosen to be played.

Say the firts hand we are OTB and have 73o. The software receives from 2000 players: 44 calls, 1699 folds and (2000 - 1699 - 44) raises as possible action for the hand. The software will then choose fold, because it is the most frequent one.

With numerical bets we will make it multiples of the pot, say 50%, 60% etc, so that it is easier to choose the most frequent one.

So in effect, durrr is playing vs 2000 players. Ofcourse, there has to be incentive for the 2000 players to actually play good. Perhaps by pooling bankrolls. In effect you are playing 50c big blind no limit at Railheaven vs durrr, PA and the likes.

It has been proved that people, en masse, predict stock prices very accurately. How would this translate into poker?

It would be awesome if someone programmed something like that in Visual Basic. I guess it's my best shot for playing durrr heads up.

Discuss.

Homework: read this article --> http://www.businessweek.com/technolo...013_033687.htm

you're forgetting that 99% of poker players are total idiots

this would be so easy for him to win since all you have to beat is a general level of skill, like you would do multitabling without a hud.

in fact even with 200.000 random online players this would be easy

if would be a different story against 2000 mid/high stakes player though, collective they would be playing near perfect poker

They already did this with chess. It took four months but judging from Kasparov's comment it turned out pretty well.

Gary Kasparov vs. The World

well since a lot of plays should be made infrequently, but almost randomly (like three betting with 27 offsuit preflop) to keep your opponent guessing, this would take all of these plays out of "the collective's" playbook. Play would be way too predictable since it would be unlikely that everyone decided to make one of these mix-it-up plays at the same time

it's an interesting concept. In short, tho, I think it would be huge advantage for Durrrr. Even if we're talking 2000 solid players, what will happen from the "collectives" side is you'll get very solid, math based decisions, but a lot of the subtle reads on Durrrr, those that really good players that play him regularly could pick up on, would be lost on the collective. Good play would become very watered down and mechanical.

From Durrr's perspective, the collective would become very predictable....like playing a solid, but nitty player. How often would the "collective" vote to bluff the whole stack to win a big hand?? Probably never. EV+ for Durrrr IMO

You guys are missing the difference between the collective voting for an action, and then doing whatever wins the vote, and the collective randomly choosing one of the outcomes out of all of its given players where more popular outcomes are more likely to be chosen based on their proportion of votes.

The first would never work, since it is relatively simple to beat someone who always makes the most obvious play.

The second is more interesting, and assuming the 2000 players are relatively decent, it would provide much better optimizations than a person could do alone (I think it's been proven that most humans are terrible at randomizing their actions without feedback on what they have done in the past), but I think that it would fall apart because of the continual nature of the hand.

For example, on the flop you have 40% who want to fold 40% who want to raise, and 20% who want to call. Randomly, you select a call. Now, on the turn, you have some portion who think calling was stupid and still want to fold, and some portion who have changed their mind, etc. and you will get hands where decisions are relatively disjointed and non optimal.

Depending on how good people are at saying, "well I would have folded the flop but as played, shove the turn" versus "fold here, you should have folded the flop and calling was dumb so get out while you can" or something, this could get interesting.

i don't know against 2000 good solid players with stats on durrr he has very little space to make moves, and you have to realize optimal game theory can't be beaten.

my guess is that in the short term he would beat the collective but in the long term the collective has the better odds.

the collective would play very passive poker indeed

Curtains did the same thing vs 2+2.

lol yeah, if they have durrrs stats theyre unbeatable!

in the long run math will take over and the collective would make perfect math based decisions, in the short run durrr has to much space to change gears and fool them.

lol, you don't get it

But what if someone were to teach Durrrr how to switch gears in the long run, or if he watched some videos or something to learn how to do this?

People would not take enought time/effort to analyze the situation because their decision is only 1/2000th of the real decision and they will feel their vote will not have an effect on the final decision (as it happens with non-voters in elections)

On every next street ignore all players' choices who opted to fold.

The further the action goes, the less players are in charge of the decision process.

Also, you may consider some special way of splitting the winning pot among the players.

If a pot is lost all players share the cost equally, regardless of their action.
However, If a pot is won, all players that chose to play further streets revieve a precenatge 'bonus' of the pricepool.
E.g.
For all "river players" 5% is taken out of the total pot and split amog them.
For all "turn and river players" another 10% is taken out of the total pot and split amog them.
For all "flop, turn, river" another 15% is taken out of the total pot and split amog them.
The remaining 70% of the pot is split equally among all the players, including the ones that chose to fold preflop.

This might be an incentive for players to play 'bigger pots' and stay in the pot longer. An collective agro-donk, if you want.