[RossSCann]

Any one here know of open source projects for C/C++ for single or double dummy play (not bidding).
I doubt there is any hope for single-dummy because if it's any good it is worth a lot of money.
I have written one and it is making money.
Double dummy however, must be available somewhere because it has no significant commercial value.
I don't want to write my own if I can avoid it.
I need this for my project to build a good program to analyze alternative bid panel questions which
I have discussed in another thread in this forum.
I also now realize I can use it to improve my existing single-dummy logic.

[Stephen Tu]

http://privat.bahnhof.se/wb758135/

[RossSCann]

 Stephen Tu, on 2012-October-19, 17:16, said:

http://privat.bahnhof.se/wb758135/

Thanks Stephen,
I really did not want to do the work myself.
I have done a single dummy function which is limited to NT contracts of a certain class
and it was hard to do. It finds a can't be defeated line of play to make the contract against any lie of the cards.
It is in my program but very seldom comes up in random hands. I really don't like it because the help text for the
user is not informative. "Play this card now or the defense may defeat your contract." Especially when it is the card to be played from dummy on the opening lead. Really especially when that card is not a likely choice.

Have you ever evaluated Haglund's executable? I assume you think it is good by your reference to it.
What other source code projects, if any, do you know of in this domain? i.e. Line of play % evaluation using simulation
against all possible lies of the cards, based on bidding, for example? i.e. Brute force Monte Carlo simulation analysis.
If this isn't available I will have to do it since we now have massively parallel multi-threading available to deal with
the combinatorial explosion problem.
Best,
Ross

[antonylee]

There is also bcalc (http://bcalc.w8.pl/) which claims to be twice faster than Bo Haglund's DDS, but the source code is not available.

I (and others) have used Bo Haglund's DDS, which is more than enough e.g. for my redeal program (https://github.com/anntzer/redeal).

On the other hand DDS is GPL-licensed so think twice before bundling it with a program you want to make money with.

[RossSCann]

 antonylee, on 2012-October-19, 19:53, said:

There is also bcalc (http://bcalc.w8.pl/) which claims to be twice faster than Bo Haglund's DDS, but the source code is not available.

I (and others) have used Bo Haglund's DDS, which is more than enough e.g. for my redeal program (https://github.com/anntzer/redeal).

On the other hand DDS is GPL-licensed so think twice before bundling it with a program you want to make money with.

Thanks. No I don't have any intentions to sell my program if I can get it working in a useful fashion to provide really good analysis of bid questios with a really easy to use GUI. I would post it for free download on my web site if there is interest.

I downloaded his dll and was pleased to have it compile immediately with only one name space conflict,
which is amazing since I have more tha 3000 variable/function names, all bridge related.

[bluecalm]

I used bcalc a lot about a year ago and it's very fast and easy to use. I used it to calculate minimaxes on 150k hands vugraph database and it only took one night on an i7.
It also has an API now (I wrote Python wrapper myself for it which worked by scrapping human readable output) so it should be very good choice for a project.

[Scarabin]

bcalc also has a single dummy solver, although still experimental

[RossSCann]

 bluecalm, on 2012-October-20, 17:59, said:

I used bcalc a lot about a year ago and it's very fast and easy to use. I used it to calculate minimaxes on 150k hands vugraph database and it only took one night on an i7.
It also has an API now (I wrote Python wrapper myself for it which worked by scrapping human readable output) so it should be very good choice for a project.

Thanks,
How did you direct output away from the screen to file(s). That is what I would need, plus file input of the hands, of course.
Ross

No longer a problem. I now have the Haglund double doummy dll working great. Nice code.
I will use standard opening leads immediately follwed by his analysis. I expect to prune
his search a lot for speed, eliminating outcomes of no interest for my purposes.

[antonylee]

command > file

Yes I'm being sarcastic.

[Zelandakh]

 RossSCann, on 2012-October-19, 16:35, said:

I have written one and it is making money.

Is this really true? I put in some Google searches (eg 'bridge "single dummy" solver') to try and find your product but to no avail. I did find a 2002 student project. I would suggest that if you are already making money with your current level of exposure it should be a real money spinner when you release it for (paid) download...

[PhilKing]

 Zelandakh, on 2012-October-22, 02:46, said:

Is this really true? I put in some Google searches (eg 'bridge "single dummy" solver') to try and find your product but to no avail. I did find a 2002 student project. I would suggest that if you are already making money with your current level of exposure it should be a real money spinner when you release it for (paid) download...

It plays a bit of zoom poker on Stars after it has finished analysing the Frank Stewart column.

[RossSCann]

 Zelandakh, on 2012-October-22, 02:46, said:

Is this really true? I put in some Google searches (eg 'bridge "single dummy" solver') to try and find your product but to no avail. I did find a 2002 student project. I would suggest that if you are already making money with your current level of exposure it should be a real money spinner when you release it for (paid) download...

Pin money. My program is BridgeMentor free download trial from my site www.bridgegame.com
Teaching program for beginners. Several hundred thousand lines of bid and play help txt.
The single dummy solver is a function called for a certain class of NT hands. It was created
to deal with a Frank Stewart newspaper column hand. My logic uses hundreds of Frank Stewart columns
with changes to avoid any copyright issues. He is a good source of stuff for beginners.
Lots of Eddie Kantar hands as well froom his books.
This is a retirement hobby work in progress. However, thousands of copies sold over the years from
the free trial. Fred Gittleman once emailed me that he liked the program. But it is just a hobby project.
The problem of dealing with any random hand with good play and help text is a lot of work.
I have looked at and coded for more than 10,000 hands, a drop in the bucket of the 10 to the 27th possible.