huge botring smashed at ftp

[thetank]

So where do you draw the "Artificial intelligence" line?

Russia

[doubleup]

Pokerstove does things that would take me hours. 

You use pokerstove in game?

no i don't but I could and afaik it isn't prohibited

anyway the point that I am making is that technology that makes your brain bigger isn't the issue.  The machine taking the decision is (eg stngowiz).

[RED-DOG]

So where do you draw the "Artificial intelligence" line?

Russia

http://www.dafk.net/what/

[thetank]

I don't disagree

I was attacking 0800's assertion that there is very little difference to the bot playing and the programmer playing himself.
 
By the looks of things, this is how he seems to justify botting. It's not allowed but he thinks it should be coz bot = him. I argue that's not the case, there's stuff the bot does better and stuff he does better than the bot.

[EvilPie]

So where do you draw the "Artificial intelligence" line?

Poker bots are not a form of artificial intelligence.

They have a set of rules which they follow, they will do whatever the previous action tells them is the right thing to do based on other factors including what their hole cards and the community cards are.

They make no decisions on their own and have no intelligence whatsoever whether it be artificial or otherwise. They simply follow a set of rules which have been pre programmed.

If a bot with true intelligence has been developed then it's GG online poker. A truly intelligent bot would adapt to other players strategies over time and as such would be undetectable as a bot. It would play perfectly and would only ever be beaten by the element of luck that creeps in to poker oh so often.

The bots that are about now are very simple. If I could be arsed I could write one myself. It's ridiculously simple and will always win a small amount as long as there are enough bad players about. Stick a truly intelligent bot in there and it would outplay everyone. Even a bot that can beat Kasparov at chess isn't intelligent. It just knows every possible outcome of every possible move it may make next. It isn't intelligence.

The biggest advantage bots have over us normal human beings is the ability to play multiple tables and keep hitting the fold button without getting bored. No boredom = no tilt = no varience. If a bot is a 0.0001% winner then it will remain so forever as long as it can keep playing.

[thetank]

 Even a bot that can beat Kasparov at chess isn't intelligent. It just knows every possible outcome of every possible move it may make next. It isn't intelligence.

Nor is it possible

A computer the size of the world dedicated to the task would take from now until after the time the sun has run out of juice to solve the game of chess in this manner.

Sufficiently sophisticated chess computers can beat any human player, but they do not do so by knowing every possible outcome, they do not play perfect chess the way a computer can now play perfect draughts.
They do know quite a lot of outcomes though, they look at well over a million a second, and pick the moves that look to be the most favourable from that limited information. With this brute force humans are no longer a match for computers in a game of chess. Computers are not better at thinking, but the sophistication of their programming has advanced to the point wherby, in chess, their brute force at calculation can overcome that disadvantage and kick our asses.

Interestingly, there's a boardgame played on the same board as chess or draughts where the best computers are not yet able to beat the best human players. It's called Arimaa

The rules of Arimaa were designed specifically with the man vs machine, strategic thinking vs brute force working out different combinations battle in mind, and to demonstrate that computers cannot think and plan better than humans (yet).

There is a $10,000 prize for the first programmer that can prove a computer can perform considerably better than the best human players in an official match. It's still up for grabs.

[EvilPie]

 Even a bot that can beat Kasparov at chess isn't intelligent. It just knows every possible outcome of every possible move it may make next. It isn't intelligence.

Nor is it possible

A computer the size of the world dedicated to the task would take from now until after the time the sun has run out of juice to solve the game of chess in this manner.

Sufficiently sophisticated chess computers can beat any human player, but they do not do so by knowing every possible outcome, they do not play perfect chess the way a computer can now play perfect draughts.
They do know quite a lot of outcomes though, they look at well over a million a second, and pick the moves that look to be the most favourable from that limited information. With this brute force humans are no longer a match for computers in a game of chess. Computers are not better at thinking, but the sophistication of their programming has advanced to the point wherby, in chess, their brute force at calculation can overcome that disadvantage and kick our asses.

Interestingly, there's a boardgame played on the same board as chess or draughts where the best computers are not yet able to beat the best human players. It's called Arimaa

The rules of Arimaa were designed specifically with the man vs machine, strategic thinking vs brute force working out different combinations battle in mind, and to demonstrate that computers cannot think and plan better than humans (yet).

There is a $10,000 prize for the first programmer that can prove a computer can perform considerably better than the best human players in an official match. It's still up for grabs.

Sorry Tank. I was banking on people not actually knowing that what I said was wrong.

Yeah computers can't possibly know every possible outcome in a game of chess in the same way humans can't.

When they play there's countless millions of combos being processed but they will never be able to do the lot.

Unfortunately for a computer playing against Kasparov it has to work out what will happen if it sacrifices it's queen in every possible way it can. This leads to millions of ways that it will obviously lose but it needs to work out the lot.

Kasparov just knows that sacrificing his queen is stupid so does those 427 billion calculations without even thinking about it. The computer doesn't have this luxury because if that was programmed in then it would no longer be calculating every possible outcome.

No matter how many calcs it did though it still wouldn't be intelligent. It would only be doing a brute force calculation as you so eloquently describe it.

I love you tank. xx

[thetank]

I love you too xx

What you say still tilts me a little as chess computers do "cut off branches" in a combination trees all the time, they would still not be able to beat human beings if they did not do this.
Also, Kasparov still thinks about sacrificing his Queen whenever he has one. Probs looks for that before anything else, the bugger.

[Longy]

Tank might just be the greatest.

[gatso]

Tank might just be the greatest.

did you cut off the branches in your combination trees before you got to all the 'tank sucks' options?

[thetank]

It's likely that to conquer the best humans at Arimaa a computer would need a completely new approach to artificial intelligence.

Looking at all the possibilities 5 moves into the future (5 moves for each player that is, so 10 'half moves') in a game such as chess where you have ~30 options each time. This would involve looking at 590 trillion combinations. Very dooable for a computer.

If instead of ~30 possible moves each time you've got ~20,000 (as in Arimaa) you've got to look at 1.54 x 10⁴⁰ combinations

That's a silly number.

If you want to just look 2 moves (4 half moves) into the future, you'd only have to consider 1.6 x 10¹⁷ combinations
Very dooable for a thousand computers.

Brute force goes out the window though, as 2 moves just isn't enough clairvoyance to overcome superior long term strategy and intuition.

$10,000 is a bit of a shit prize for this task imo. If anyone manages it I'll chuck in another tenner myself.

[wormster]

so how can anyone play online poker when it's clear it's not a level playing field?

Steve! How are you?

Do NOT worry about "bots", Steve, you would not be at a disadvantage playing against them. It's wrong, of course it is, but the suggestion that they can beat a half-competent player does not wash with me.

Hi Tony,

I'm well thanks, how are things with you?

The bots thing doesn't worry me as I very rarely play poker these days. For me, it's very boring & there is too much luck involved.

Would be good to catch up with you soon. Our office has a monthly poker game, which you're more than welcome to come down to. Structure & standard of play isn't too bad & the company is good. The next one is Monday Oct 25th. Would be good to see you if you can make it

[RED-DOG]

So where do you draw the "Artificial intelligence" line?

Poker bots are not a form of artificial intelligence.

They have a set of rules which they follow, they will do whatever the previous action tells them is the right thing to do based on other factors including what their hole cards and the community cards are.

Doesn't that apply to all so-called artificial intelligence?

Anyway, my question still stands. Stoves - hud's - trackers - bots....  Where do you draw the line?

[rex008]

So where do you draw the "Artificial intelligence" line?

Poker bots are not a form of artificial intelligence.

They have a set of rules which they follow, they will do whatever the previous action tells them is the right thing to do based on other factors including what their hole cards and the community cards are.

Doesn't that apply to all so-called artificial intelligence?

Anyway, my question still stands. Stoves - hud's - trackers - bots....  Where do you draw the line?

I guess a bot would be classed as an Expert System. Semantics, shmemantics.

[RED-DOG]

So where do you draw the "Artificial intelligence" line?

Poker bots are not a form of artificial intelligence.

They have a set of rules which they follow, they will do whatever the previous action tells them is the right thing to do based on other factors including what their hole cards and the community cards are.

Doesn't that apply to all so-called artificial intelligence?

Anyway, my question still stands. Stoves - hud's - trackers - bots....  Where do you draw the line?

I guess a bot would be classed as an Expert System. Semantics, shmemantics.

And, it still doesn't answer the question.