Which game is more skillful: chess, backgammon or poker?
Tad Perry
>
> Complexity Numbers
>
> Go 40
> Chess 14
> Scrabble 10
> Poker 10
> Backgammon 8
> Draughts 8
> Blackjack 2
> Craps 0.001
> Lotteries 0.0000001
> Roulette 0
>
>The good feature of this table is that it resolves at one stroke all the
>muddled thinking about luck, skill, games of skill and games of chance. Any
>game with no skill falls automatically to zero. For all other games, the
>relevant issue is not luck versus skill but rather the interplay of skill,
>chance and complexity.
>
Cool! This is exactly what I was trying to get at in my article. Had
I read this first maybe I could have organized my thoughts better.
tvp
helmet
>I do not know whether this question can indeed be answered
>however as I am not UP in backgammon - could someone post here
>and tell we ignorant chess players what the state of affairs is
>with backgammon and computers?
pun intended.
>Gilbert Palmer
>A Chess fan and a bad but keen backgammon player.
The top backgammon programs have not yet reached
the point where they are the best players in the world, however
they are very close. The situation is similar to chess in that
the backgammon progs can beat 99% of players but havnt
been able to solve the game. The good backgammon programs
are neural nets such as snowie and jellyfish. Jellyfish
can be downloaded free, perhaps some of the chess/poker players
would like to give it a try. It can be downloaded at
It is a very advanced program, at least the equivelent of fritz
in chess probably better.
helmet
helmet
of the most skillful game better than any other and is probably as near to a
scientific answer as we will have. I have moved the article from the stress to
the skill thread
helmet
Bill Robertie produced an interesting piece on this topic some years ago in
Inside Backgammon. It was so good that I sought Bill's permission to
reproduce the bulk of the article for my newspaper column as it was a
question often asked by readers. The article is reproduced below and should
provide food for thought.
How much of backgammon is luck and how much is skill? If I said it was 60%
skill and 40% luck what would I mean? If you said it was 50% skill and 50%
luck how could we prove who was right and who was wrong? The answer is we
couldn't. Let's look at the problem from a different perspective and compare
the skill and chance factors involved in different games.
We shall start with chess where ratings rank from a high of 2800 to a
theoretical low of zero. Chess ratings are also designed so that a 200 point
rating difference between two players anywhere on the scale means that the
higher-rated player has a 70-75% chance of defeating the lower-rated player
(discounting draws which are possible in chess but not in most of the other
games we will consider).
Now consider the following experiment:
1. Take the best player in the world (Gary Kasparov). Call him player 1
2. Find someone that the best player beats 70-75% of the time. Call him
player 2.
3. Call the difference between players 1 and 2 one skill differential.
4. Find someone that player 2 can beat 70-75% of the time. The difference
between player 2 and player 3 is another skill differential.
5. Continue this process until you have taken the chain down to an absolute
beginner.
6. Count the number of skill differentials involved. This is the complexity
number of the game.
We can now apply this process to any game, although we may have to give some
thought as to what constitutes a meaningful contest. In chess, a single
tournament game of four to five hours seems reasonable. In backgammon it
would probably be a 25-point match, in scrabble perhaps a best of five
series, and so on.
Here is a rough chart of how various games rank on the complexity number
scale:
Complexity Numbers
Go 40
Chess 14
Scrabble 10
Poker 10
Backgammon 8
Draughts 8
Blackjack 2
Craps 0.001
Lotteries 0.0000001
Roulette 0
The good feature of this table is that it resolves at one stroke all the
muddled thinking about luck, skill, games of skill and games of chance. Any
game with no skill falls automatically to zero. For all other games, the
relevant issue is not luck versus skill but rather the interplay of skill,
chance and complexity.
I am indebted to Bill Robertie, two time world backgammon champion for the
research for this article. I do not have a figure for bridge - perhaps a
reader may care to submit an opinion?
Henri H. Arsenault
> >I do not know whether this question can indeed be answered
> >however as I am not UP in backgammon - could someone post here
> >and tell we ignorant chess players what the state of affairs is
> >with backgammon and computers?
> pun intended.
>
quality of a good poker player (one who knows all the probabilities and so
on), is the ability to tell the opponent's state of mind from subtle
signs, factors that computers are unable to deal with. Most people (not to
mention most experts) are unable to control their reflexes and the
giveaway signs that tell an expert poker player whether or not he is
bluffing, whether or not he has picked up a good hand, etc. (For example,
when they get a good hand, most players will look away from the table to
feign indifference).Poker-playing dogs will wag their tails... ;)
A player who is bluffing will make an involuntary reflex start when he is
bluffing and an opponent moves his hands towards his chips as if to call
the bluff. Even champion-level poker players cannot totally control such
reflexes, but only experts are able to spot them infallibly (and some can
feign them...). Why do you think that the same guy won the world poker
championship five times? Luck? My ass...
Anybody for a little game of poker?... ;)
Henri
PMG
helmet wrote:
> <snipped>
>
> Nice post.
> Although the computer will have difficulty
> telling when an opponent is bluffing it will also work in reverse.
> The computer would have the ultimate poker face, if it was programmed
> to bluff occasionally it would be very hard to read.
> Kasparov had a similar problem playing deep blue, he normally
> glares at opponents and psyches them out. Against deep blue
> however it was he who was out psyched
> helmet
I've punched my computer monitor several times when loosing Chess games to it,
and it never even flinches!
Pete
Gary Philips
Henri H. Arsenault wrote among other things:
Why do you think that the same guy won the world poker
championship five times? Luck? My ass...
You're not one of Kenny Houston's psychics, are you? If you are referring to
the WSOP, nobody alive has won it more than twice. Only a couple who have
passed on have won it three times.
Gary (my lucky ass) Philips
helmet
>Poker does not fit well into this scheme of things, because the main
>quality of a good poker player (one who knows all the probabilities and so
>on), is the ability to tell the opponent's state of mind from subtle
>signs, factors that computers are unable to deal with. Most people (not to
>mention most experts) are unable to control their reflexes and the
>giveaway signs that tell an expert poker player whether or not he is
>bluffing, whether or not he has picked up a good hand, etc. (For example,
>when they get a good hand, most players will look away from the table to
>feign indifference).Poker-playing dogs will wag their tails... ;)
>A player who is bluffing will make an involuntary reflex start when he is
>bluffing and an opponent moves his hands towards his chips as if to call
>the bluff. Even champion-level poker players cannot totally control such
>reflexes, but only experts are able to spot them infallibly (and some can
>feign them...). Why do you think that the same guy won the world poker
>championship five times? Luck? My ass...
>Henri
helmet
>>
>>I'd guess that to produce a near-optimal program, heads-up limit holdem
>>would be the easiest, backgammon next, and chess the most difficult.
>>There just aren't that many truly different situations in poker
>>relative to the other two games, and the number of game-positions is
>>more-or-less what defines how difficult it is to write an optimal
>>player. Backgammon is clearly easier than chess; even now, if you want
>>a practically optimal program just let Jellyfish or Snowie pick its
>>plays by rolling out the position at every move. Although this will
>>mean it takes several days to complete a game, it's not hard to imagine
>>that in a few years computers will be fast enough to make this a
>>practical method.
>>
>>But, if you want a truly optimal opponent, probably poker is the
>>hardest because although many situations are fairly similar, they're
>>not *exactly* the same, and although the state-space is smaller than
>>that of backgammon or chess, the presence of hidden information (which
>>is known to one player, but not the other) makes the optimal strategy
>>calculation very difficult.
>Programming a computer to make near perfect poker (most forms)
>decisions against conscious human opponents is hugely more complicated
>than programming a computer to make near perfect chess or backgammon
>decisions against conscious human opponents.
>Others will argue differently, but they will be wrong, and they should
>stop arguing because they have now been provided with the answer.
>Straight Flushes,
>Mike Caro
Henri H. Arsenault
<lvd...@earthlink.net> wrote:
> Henri H. Arsenault wrote among other things:
> Why do you think that the same guy won the world poker
> championship five times? Luck? My ass...
>
> You're not one of Kenny Houston's psychics, are you? If you are referring to
> the WSOP, nobody alive has won it more than twice. Only a couple who have
> passed on have won it three times.
>
I don't really know, I thought I had heard that on a program on Discovery
Channel the other night, but I could be wrong, or the commentator could be
wrong...
Henri