[tytobyto]

Or would a different website entirely be more appropriate?

[barmar]

There's probably no perfect forum. Stack Overflow is good for help with programming problems in general, but you're not necessarily going to find lots of bridge-specific expertise (although I'm there). There are a number of programmers here, so this might be better. What specifically are you planning on programming? If it's anything that competes directly with a BBO service, we wouldn't be interested in facilitating it.

[Rossoneri]

I have recalled programming-related issues being discussed here, so fire away!

[tytobyto]

 tytobyto, on 2012-December-11, 18:56, said:

Or would a different website entirely be more appropriate?

Back in the early 80s I wrote some Basic programs to help my new girlfriend learn to bid Precision. Although I was a good programmer, I was not a very good bridge player, so I never tried to write a program to coach the play of the hand (other than defensive leads which were by formula).

In the later 80s, I became increasingly distressed by the fact that although there were good chess-playing programs, bridge-playing programs were an embarrassment. I bought a book on the subject (which I unfortunately no longer have) that described the early optimism that computers would soon be able to play bridge infallibly: just calculate the odds, weight the outcomes by the scoring, and you have your play. The book, while maintaining confidence that direct calculation would someday prevail, devoted it's bulk to examinations of heuristic approaches which they claimed were on the verge of competent play. I still remember standing in the shower as it occurred to me that generating random hands to fit the situation of original bidding and subsequent play, then examining all four hands to find best play and then calculating the most likely best play, was a strategy worthy of pursuit if sufficient hands could be generated and evaluated in a reasonable amount of time.

I had already developed some routines to generate random hands meeting certain specifications for my programs which tutored bidding. But I had no clue how to develop a fast and reliable double-dummy player so several years had to elapse before modern bridge programs were created by people more talented than I am.

Now, I'm still interested in programs to coach bidding, and I still don't know how to write a fast and efficient double-dummy player. I need a free one that I can link to from a C program (or source in C or C++). I want to employ the DD player in slam bidding. I'm also interested in thoughts from anyone else who has used DD players in bidding rather than in play.

[Rossoneri]

Have you tried looking at Ginsberg's paper on GIB (pdf file)?

Edit: Found this thread as well: http://www.bridgebas...new-simulation/

 tytobyto, on 2012-December-14, 19:39, said:

Back in the early 80s I wrote some Basic programs to help my new girlfriend learn to bid Precision. Although I was a good programmer, I was not a very good bridge player, so I never tried to write a program to coach the play of the hand (other than defensive leads which were by formula).

In the later 80s, I became increasingly distressed by the fact that although there were good chess-playing programs, bridge-playing programs were an embarrassment. I bought a book on the subject (which I unfortunately no longer have) that described the early optimism that computers would soon be able to play bridge infallibly: just calculate the odds, weight the outcomes by the scoring, and you have your play. The book, while maintaining confidence that direct calculation would someday prevail, devoted it's bulk to examinations of heuristic approaches which they claimed were on the verge of competent play. I still remember standing in the shower as it occurred to me that generating random hands to fit the situation of original bidding and subsequent play, then examining all four hands to find best play and then calculating the most likely best play, was a strategy worthy of pursuit if sufficient hands could be generated and evaluated in a reasonable amount of time.

I had already developed some routines to generate random hands meeting certain specifications for my programs which tutored bidding. But I had no clue how to develop a fast and reliable double-dummy player so several years had to elapse before modern bridge programs were created by people more talented than I am.

Now, I'm still interested in programs to coach bidding, and I still don't know how to write a fast and efficient double-dummy player. I need a free one that I can link to from a C program (or source in C or C++). I want to employ the DD player in slam bidding. I'm also interested in thoughts from anyone else who has used DD players in bidding rather than in play.

[FM75]

Bo Haglund DD link.
http://privat.bahnhof.se/wb758135/

I have gotten a version of this working on a Mac. I am NOT working on developing an AI game playing program, however.

Ginsberg paper is worthwhile if that is your interest.

[tytobyto]

Thanks for the links. Hopefully, I will soon know my way around this forum well enough that I won't have to start threads just become I'm lost. I would never have guessed that "simulation" was a word I should look for.

But I like this place so far. People seem very quick to help.