[fred]

Scarabin, on 2012-March-30, 23:02, said:

As long as we do not know exactly how GIB works you must expect us to speculate and often get it wrong. All suggestions can seem destructive but I think xxhong and cloa513 have made contributions which could be used to improve GIB. The same goes for antrax. One of cloa513's questions sticks in my mind and I think merits further investigation: "why do GIB's simulations not cause it to cash out when it has established enough winning tricks?"

Now I am biassed, I want to see a pragmatic reasoning AI become the world champion and purely selfishly I wish you had developed Base lll into Base 17!

Having defended the indefensible for most of my life, I do empathise with Barmar even though he does it much better than I used to.

GIB sometimes doesn't cash out because it *thinks* that the IMP or MP odds are such that cashing out is wrong. Strong human players frequently do the same thing (especially at MPs).

I am not saying that I always (or even usually) agree with GIB's bridge conclusions in such cases, but there is considerable logic behind such actions. We have considered introducing some kind of artificial bias with respect to the value of the making/setting trick (mostly because we think it will improve customer satisfaction even though it will arguably make GIB a worse bridge player).

I disagree with your statement that "all suggestions can seem destructive". Tone matters. There is a big difference between users like antrax and those like cloa513 (and to a lesser extent xxhong). Antrax always expresses himself in a polite, respectful, and constructive manner even when he is being critical.

Fred Gitelman
Bridge Base Inc.
www.bridgebase.com

[pigpenz]

I know I myself have learned to live with the results from GIB....comparing it to playing in an old fashioned open event back in 70's or 80's when pairs events were large.
there would always be an obtuse result here and there.

I think the main problem listening to people talk about GIB's is that there are too many versions of them and the programming keeps trying to make them into different things by the way the upgrades are implemented.

So when people post things they are from all versions, mainly BBO needs to have bugs posted along the lines of cheap gib, middle gib, advanced gib.

From my original line...I notice one thing when playing with GIB's in open room, your results always become skewed when there are humans both North and South....because the Human north doesnt defend in the same ways most of the time like the GIB does.

maybe it would be nice if there were an option to just have open room where human as south and GIB at all other seats....and just those results compared.

[Antrax]

Thanks barmar.

[Scarabin]

fred, on 2012-March-31, 08:48, said:

GIB sometimes doesn't cash out because it *thinks* that the IMP or MP odds are such that cashing out is wrong. Strong human players frequently do the same thing (especially at MPs).

I am not saying that I always (or even usually) agree with GIB's bridge conclusions in such cases, but there is considerable logic behind such actions. We have considered introducing some kind of artificial bias with respect to the value of the making/setting trick (mostly because we think it will improve customer satisfaction even though it will arguably make GIB a worse bridge player).

I disagree with your statement that "all suggestions can seem destructive". Tone matters. There is a big difference between users like antrax and those like cloa513 (and to a lesser extent xxhong). Antrax always expresses himself in a polite, respectful, and constructive manner even when he is being critical.

Fred Gitelman
Bridge Base Inc.
www.bridgebase.com

You are right, of course. Please disregard my last post. I should not interfere and I now realise I am only encouraging rude behaviour.

[barmar]

Scarabin, on 2012-March-29, 22:23, said:

Fair enough, but in "Upgrated GIB" you spoke of adding overriding pragmatic rules to take care of situations like cashing out and where GIB seems to lose its way. Has this also been ruled out?

No, but I'm not going to predict any time frame when we might figure out how to do this.

[jon321]

Why not create some REST API endpoints that allow us to access GIB, ie hit it with bidding problems or cardplay problems, in a way other than playing the game on BBO? This way we could expose weird bidding flaws, etc, and improve the system together. This would not require showing the GIB source to anyone. If someone is smart enough to reverse engineer it then they deserve the benefits accrued to said activity anyway