[grahamg]

The BBO hand record and review system is excellent.
One addition I would like to see is a display of all making contracts, given perfect knowledge of the layout and best play and defense.
Yes it is not reality, but still a useful tool in my opinion.

[Vampyr]

grahamg, on 2020-September-05, 04:10, said:

The BBO hand record and review system is excellent.
One addition I would like to see is a display of all making contracts, given perfect knowledge of the layout and best play and defense.
Yes it is not reality, but still a useful tool in my opinion.

BBO do not make any improvements, as I am sure you are aware, but perhaps there is an easy way to export the hands to deep finesse.

[grahamg]

Vampyr, on 2020-September-05, 04:58, said:

BBO do not make any improvements, as I am sure you are aware, but perhaps there is an easy way to export the hands to deep finesse.

Thank you for suggesting the export idea. Dealmaster Pro format would be good as it can handle batches of hands.

I am a bit bemused by your comment on no improvements. Currently on BBO is a topic on upgrades to the servers which shows various improvements.

[paulg]

It is easier to download your BBO hands into Double Dummy Solver as you can do this in batch mode (for example, get all my hands from BBO for the last week) - http://bridgecaptain...ownloadDD.html.

This will also show you how you've performed against par (the score that neither pair can improve on with perfect play) on each hand.

BBO has made lots of improvements although, perhaps, not those that Vampyr considers most important. In my opinion, quite reasonably, most seem focused on stabilising the platform with the increased workload and support for virtual clubs rather than more visible new features.

[Vampyr]

paulg, on 2020-September-05, 10:09, said:

It is easier to download your BBO hands into Double Dummy Solver as you can do this in batch mode (for example, get all my hands from BBO for the last week) - http://bridgecaptain...ownloadDD.html.

This will also show you how you've performed against par (the score that neither pair can improve on with perfect play) on each hand.

BBO has made lots of improvements although, perhaps, not those that Vampyr considers most important. In my opinion, quite reasonably, most seem focused on stabilising the platform with the increased workload and support for virtual clubs rather than more visible new features.

Yeah, I guess I am more concerned with complying with the Laws of Bridge; it is ludicrous that BBO even includes “Bridge” in their title; IMO it should have more resemblance to the actual game.

The other thing I care about is that games be a valid contest; maybe there is another platform that will have that.

Yes, it is fun poring over the hand records trying to figure out why West can make an overtrick while East can’t, but I consider things like this to be peripheral to the actual game being played.

[Huibertus]

This would be a pointless overview. It would show for instance you can make 12 tricks in ♥, and if you examine declarer play that will lead to 12 tricks in hearts it might turn out to be a play that is unlikely to succeed, ie dropping a bare K in a side 4-4 fit, where any decent player would just take the finess.

[pigpenz]

yes double dummy solver from birdge captain will show all the makeable contracts, its a good tool but not exact science because it doesnt take into account human interacton on the play or defense

[paulg]

pigpenz, on 2020-September-06, 11:14, said:

yes double dummy solver from birdge captain will show all the makeable contracts, its a good tool but not exact science because it doesnt take into account human interacton on the play or defense

It is an exact science because it is a double dummy solver. GIB's implementation on BBO is a single-dummy solver and perhaps more realistic.

But you are right that DDS is a tool that needs some expertise and experience to analyse the rights and wrongs of a contract, bidding and line of play.

[pescetom]

paulg, on 2020-September-06, 14:38, said:

GIB's implementation on BBO is a single-dummy solver and perhaps more realistic.

Is this really so?

[grahamg]

paulg, on 2020-September-05, 10:09, said:

It is easier to download your BBO hands into Double Dummy Solver as you can do this in batch mode (for example, get all my hands from BBO for the last week) - http://bridgecaptain...ownloadDD.html.

This will also show you how you've performed against par (the score that neither pair can improve on with perfect play) on each hand.

BBO has made lots of improvements although, perhaps, not those that Vampyr considers most important. In my opinion, quite reasonably, most seem focused on stabilising the platform with the increased workload and support for virtual clubs rather than more visible new features.

Thank you for that. Bridge Captain is certainly a good option.
I'd still like to have that table of "make-able" contracts on the BBO site on the hand record as it is a very good indicator if used with an experienced eye and if a detailed analysis such as provided by Bridge Captain is not needed..

PS your link does not work (at least in my browser) because it has a full stop included at the end.
Here it is without that.

Bridge Captain

[freds1729]

I use this excellent Chrome Extension from Bridge Solver.

[johnu]

paulg, on 2020-September-06, 14:38, said:

It is an exact science because it is a double dummy solver. GIB's implementation on BBO is a single-dummy solver and perhaps more realistic.

If by single dummy solver you mean that GIB does a sequential series of double dummy calculations after each card played, then I guess you could call it a single dummy solver. I would not call it a single dummy solver.

[paulg]

johnu, on 2020-September-07, 23:04, said:

If by single dummy solver you mean that GIB does a sequential series of double dummy calculations after each card played, then I guess you could call it a single dummy solver. I would not call it a single dummy solver.

I think this is the definition of a single dummy solver

GIB is not a double dummy solver because it cannot see the unplayed cards. I see to recall that the robots on OKBridge could do that, but my memory is poor and it was a long time ago.

As far as I am aware, GIB's algorithm as a single dummy solver is to generate hands that are consistent with the play and bidding to date and choose the line that works on most of these hands (using a double dummy solver on the possible hands). Presumably one difference between the basic and advanced robots is how many hands they generate. The inferences that GIB takes from the auction and carding to date are the key factors in how good a player GIB is. Assuming that everyone plays GIB 2/1 is not ideal and one reason why its play and defence are sometimes subject to ridicule. Anther reason is that it assumes that the opponents are similar machines and will remember all the cards played, which is why some apparent defensive howlers appear.

But the fact that it cannot see all the hands when making its calculations means it is a single dummy solver. No longer close to being the best but far stronger than the majority of BBO players.

[johnu]

paulg, on 2020-September-08, 01:07, said:

I think this is the definition of a single dummy solver

No, still a double dummy engine. Running a double dummy engine multiple times doesn't magically make it a single dummy solver. GIB still assumes double dummy defense and declarer play at every stage.

paulg, on 2020-September-08, 01:07, said:

GIB is not a double dummy solver because it cannot see the unplayed cards. I see to recall that the robots on OKBridge could do that, but my memory is poor and it was a long time ago.

Seeing the unplayed cards is not what a double dummy solver means in the context of bridge playing programs. GIB "deals" a number of different hands based on the bidding (and play) to date, and then applies actual double dummy analysis to determine what to play next. If a 1% line of play succeeds on a particular hand when a 99% line of play would fail, GIB will take the 1% line of play on that sample hand. This is aggregated with other sample hands that have been dealt.

paulg, on 2020-September-08, 01:07, said:

But the fact that it cannot see all the hands when making its calculations means it is a single dummy solver. No longer close to being the best but far stronger than the majority of BBO players.

And yet, it uses a double dummy solver to determine the plays. This means that finesses are never misguessed, you always correctly play for bad splits, or good splits, and always make the best leads and subsequent plays. That is not a single dummy solver IMO. This is also why you will see GIB randomly lead singleton kings, or cash top tricks in other suits (potentially setting up tricks for the opponents) when there is still work to be done on the trump suit for example. GIB assumes that if there is a 2 way finesse or favorable split available then there will be a favorable result. That is not single dummy analysis.

[barmar]

I think there's some confusion here.

When GIB is playing a hand, it's playing it single dummy. It uses a double dummy solver internally, by dealing out numerous hands that fit the known conditions, performing double dummy analysis of the results of each possible play, and choosing a play that has the best expected result. Also, the declarer robot has an more complex algorithm called "GIBson", which is used by advanced robots after several tricks have been played; I don't really understand the details of this algorithm, but it's still single dummy.

But after a hand is over and you're viewing the results, you can click on the "Show double dummy" option. Then it shows a double dummy analysis of the result of each possible play.