[Zelandakh]

 jogs, on 2015-April-08, 10:47, said:

The math guys use the term perfect information for deterministic games.

I do not believe these 2 terms are equivalent.

[jogs]

 Zelandakh, on 2015-April-08, 12:02, said:

I do not believe these 2 terms are equivalent.

Look at the second matrix on page 2 on Helene's link on post #74.
Perfect information is the definition of a deterministic game.

[helene_t]

 jogs, on 2015-April-08, 14:51, said:

Look at the second matrix on page 2 on Helene's link on post #74.
Perfect information is the definition of a deterministic game.

Not according to that document. It says "battleships" is deterministic with imperfect information while backgammon and monopoly are "chance" (nondeterministic) with perfect information.

[hrothgar]

 helene_t, on 2015-April-08, 16:08, said:

Not according to that document. It says "battleships" is deterministic with imperfect information while backgammon and monopoly are "chance" (nondeterministic) with perfect information.

I think that the claim about Monopoly having perfect information is incorrect is you include the order of the Cards

[helene_t]

 hrothgar, on 2015-April-08, 17:02, said:

I think that the claim about Monopoly having perfect information is incorrect is you include the order of the Cards

I am not sure about this.

You could replace the dice with a pile of cards numbered 1-6. So if dice-throowing counts as indeterministic (as opposed to imperfect information) then the cards in Monopoly must do the same.

"Imperfect information" presumably refers to information available to one player but not to another.

[alok c]

Bridge&Chess are entirely different types of games.Chess is a game of complete informations(though enormous) from the starting of the game-which is very suitable for a computer from the point of view of processing.On the otherhand bridge involves lots of conjectures in every steps(bidding,defence,handplay)-as guessing invariably involves some sort of emotions,computers are still not that good at Bridge.

[Richardrls]

As a student of logic, you will find that logic formulae are either well formed or not well formed. In Bridge this translates to the issue of definitions. In the game of Chess, all definitions are in fact complete or well formed and can therefore be programmed into a computer. In point of fact, there are in the game of bridge a number of definitions that are not complete, and yet the human mind can deal with such incomplete definitions. A computer or robot can not deal with such since they can not be programmed. Some examples of incomplete definitions: Cohen's concept of the Freak Deal or freak hand. A much much older example? the concept of good or bad hand texture. There are many others in bridge...none of which can be programmed into robot software. Hope this helps.

[Zelandakh]

The last 2 posts show a complete lack of understanding of this area. It is just as easy, probably easier, to program a computer with concepts such as hand texture than positional concepts such as space and tempo. You simply have to get world class players in that can explain how these things work, just as happened for chess computers. There is no emotion required, only the establishment of a suitable knowledge-based system. Once the basic rules are in place and start to be optimized there is no reason why the computer cannot exceed the performance of "emotionally-driven guesses", just as has happened for chess in positions that were once thought of as impossible for a computer to evaluate.

[alok c]

The last post completely lacks the understanding of how a computer logic works.

[steve2005]

 lackeman, on 2015-March-31, 03:09, said:

i think more humans play against them than vs chesscomputers. What do you think? And Why if so? Because human preforms better comparired w Computers!

Not sure but a lot of people play chess against computers, if more play against bridge computers it's a reflection on the popularity of bridge vs chess.

Chess is a deterministic finite game. The number of possibilities is very large. In late 70's or early 80's someone came up with an algorithm which trimmed the possibilities down to a workable level. Then they just had to wait for an improvement in computer hardware. Checkers which has a much smaller finite number of moves and has had computers being unbeatable much earlier than chess computers were even expert level.

[hotShot]

Being good at chess is/was associated with being intelligent. Building a chess playing computer that could beat the best was good advertising for Soft- and Hardware.

A human bridge player uses several methods to gain information:
1) Bids and plays an opponent/partner actually made
2) Bids and plays an opponent/partner did not make
3) Experience on former hands played (bidding style aggressive or cautious, playing strength )
4) Body Language / Timing issues (probably less in WC games)

One can only make one bid, but that leaves a lot more bids not made, so there is a lot information in 2).
It is hard enough to write down a system description defining what the bid and various sequences show,
it's almost impossible to write down all the hands that are totally unlikly once a bid/bidding sequence is made.

Algorithms rely mostly on 1) only a little on 2) and consider 3) by not trusting opponents and partner to much.
So there is a chance that playing WC opps is easier for the software since they are more reliable.

I see 3 ways that could improve bridge programs:
1) Finding a better way to "explain" biddingsystems, lead conventions and signals to the program
2) finding a different algorithmic approach
3) Introduce real AI-elements like neuronal nets or machine learning etc. to the software

I don't think that computational power is the problem, it's money. Take a few experts in probabilistic evaluation, artificial intelligence and software development and some 5-10 years of time and I'm pretty sure they will come up with something that is (much) better than what we have now.

[Antrax]

For the record, poker-playing software heavily relies on your #3 above.

[kontoleon]

Real sorry from my answer, but why said that Bridge is a game of skill and luck?

I just thing that is clear a game of SKILL only.

in fact at all the table have just the some odd to win, (all the players have exacly the some cards and start at the some possision!) sorry guys...


So, why is a game of skill AND luck? if you miss the finess here , miss the finess and the other table, if the correct call is 6s here the correct call is 6s and on other table.

ok maybe if the other table open other first card, but from my to oppen the correct card is SKILL...,
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------

An exaple 11 card fit miss the K. the correct plan is the cash (52 vs 48), maybe some one try the finess, and win by finess. From my is 4% error,

One time personally i win a hand by misclick! click by error the A, and no the finess, (9 card fit, breaks 3-1 with single K!) but is still a HUGE ERROR,

I thing that is the only change to win a beginer an expert on bridge...

[kontoleon]

Real sorry from my answer, but why said that Bridge is a game of skill and luck?

I just thing that is clear a game of SKILL only.

in fact at all the table have just the some odd to win, (all the players have exacly the some cards and start at the some possision!) sorry guys...


So, why is a game of skill AND luck? if you miss the finess here , miss the finess and the other table, if the correct call is 6s here the correct call is 6s and on other table.

ok maybe if the other table open other first card, but from my to oppen the correct card is SKILL...,
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------

An exaple 11 card fit miss the K. the correct plan is the cash (52 vs 48), maybe some one try the finess, and win by finess. From my is 4% error,

One time personally i win a hand by misclick! click by error the A, and no the finess, (9 card fit, breaks 3-1 with single K!) but is still a HUGE ERROR,

I thing that is the only change to win a beginer an expert on bridge...

(of corse the new player have 0% change of win, just because the complex beat sestem) PLEASE DELETE THIS MESSEGE DOUBLE POST

This post has been edited by kontoleon: 2015-May-11, 02:12

[mycroft]

Why is it a game of skill and luck? Because it is not a perfect information game, and because the starting position is variable.

There's a lot of skill that goes into system-building; there's a lot of skill that goes into exercising that system; there's a lot of skill that goes into minimizing the luck in the play; there's a lot of skill in playing the metagame/field. Almost all of it involves managing the luck in the game (note, not necessarily minimizing it; like "shortstack play" in tournament poker, "well behind in the match" play involves choosing when to take your anti-field chances hoping that this time it wins, or choosing plays that increase the luck factor of the game, hoping it will hit in your favour this time).

[nige1]

IMO Bridge mixes skill and luck in much the same proportions as Backgammon. Where Bridge differs from most sports and games is the importance of the partnership aspect. I think bridge-programs already play better than most humans. The best programs would be world-class, If they

• used efficient methods (they could cope with complex systems beyond practical human capacity).
• had access to a database of opponent's methods.
• were allowed to keep asking questions until they had a reasonable grasp of the systemic-meaning of an opponents' bid or play.
• recorded opponents' propensities and idiosyncrasies.
• hived off simulations to a massively parallel subsystem.

If Gib could complain about humans, then BBF would need Gb more storage