[benlessard]

Difference of tricks between Double dummy analysis "Deep finesse" and between real play.


Using Bridge browser is it possible to see the difference of trick between Double dummy and real play.

Double dummy greatly help declarer even if defence can make some killer leads.


My guess is a bit less then half a trick (0,4) per deal.

[inquiry]

It is a statistical thing. Deep Finessee shows you how many tricks can be made with double dummy play and double dummy defense. Many of the declarer plays that take the maximum number of tricks will never be made in the real world, so the maximum number of tricks taken often be less, sometimes much less, than shown. On the other hand, the same applies for the defense. Double dummy defense is particularily hard to find (and often impossible).

So on a typical hand, declarer will not find the magic line to take the right tricks, and on others the defense will not find the magic line to hold declarer to the "right" number of tricks. If you look at enough hands (lets say thousands), the deep finessee estimate of number of tricks will be very close to the number taken, but ony because there is a lot of overtricks and undertricks being made. I thin if you look at enough hands, deep finessee average to what humans make (in the same contract) will be a lot closer than even your guestimate of 0.4 tricks per deal. Actually someone looked at this at one time, and it was indeed close, but alas I don't remember how close.

BTW: every HomeBase tourney will have Deep Finessee calculated contracts, so you will be able to check our travelers against the "makeable" contracts and see for yourself overtime. The travelers will be posted on our webpage.

[joker_gib]

Maybe a silly question (I tried to find another thread on this...) but :

What will be the Homebase Club, Ben ??

Alain

[FrancesHinden]

I'm not sure what the average over all contracts will be, but the higher level, the better I expect DF to be (considering contracts bid to make: for sacrifices the effective 'level' of the contract is much lower).

In most slams, and definitely most grand slams, there is not usually a killing lead and frequently little interest in the defence but much interest in the declarer play. I would expect DF to average over half a trick better on hands with interest in the play. But once you remove the hands that can claim, DF will average less than half a trick better.

For lower level contracts I would expect it to even out, so overall I think 0.4 tricks is too much.

[tysen2k]

inquiry, on Feb 7 2006, 08:53 AM, said:

Actually someone looked at this at one time, and it was indeed close, but alas I don't remember how close.

Peter Cheung did several years ago. On average real play takes 0.1 tricks more than DD. For slams, usually the DD declarers will do better. In partscores, the real declarers will do better. For games, they are very close.

More details here.

Tysen

[inquiry]

joker_gib, on Feb 8 2006, 03:46 AM, said:

Maybe a silly question (I tried to find another thread on this...) but :

What will be the Homebase Club, Ben ??

Alain

Ah.. at last a sucker... er.. educated questioner.

HomeBase Club is yet another private bridge club on the BBO gaming site that, among other things, will run tournaments (fee-based). In addition, homebase will try to offer some club-room activities that will be "mostly" free (those require some work first).

I don't want to fill up the BBF with post on this topic, so check out the link in my signature, and read the news forum discussion section at that site. We also are posting the relevant bridge laws and how we apply them given the nature of what we can do with the existing netbridgevu software. I think if you check it out, you will like what HomeBase has to offer.