[virgosrock]

zhasbeen, on 2018-May-10, 14:13, said:

I can’t speak for everyone, but I haven’t noticed any improvement in GIB’s overall play since I started in August of 2017. I have played only tournaments with 3 GIB’s and 1 human at each table, so my sample doesn’t include the other forms of play that BBO offers. Of those,more than 90% have been the ACBL sanctioned 12-board matches. The 1065 matchpoint games (tournaments) I’ve played have averaged 29 tables. Average tables for 335 IMP games has been 22. I also keep a blizzard of other stats.

My objective when I started this thread is pretty much described in the heading. Venting of some frustrations has been a byproduct.

I don’t consider myself to be a BBO basher or a chronic complainer. When I realized that I was moving in that direction I stopped posting. I’d much rather play than complain, and have. However, I believe that BBO is making a mistake by continuing to go with GIB for the indefinite future, unless there is more proof that the problems can be fixed than what most of us have seen so far.

I am aware that there is no easy solution, and replacing all the GIBs would not be cheap. It might not even be possible, but keeping an open mind to change would be a start. Maybe it could be done more gradually than all it once.

If BBO actually believes that GIB can be fixed, don’t keep us in the dark about what the next upgrade is going to correct, and give us a reasonable approximation of when. I’m all for fixing up ol’ GIB if I can see noticeable improvement in some of his more serious flaws.

In addition to fixing the wild jumps to slam, here’s another recurring problem hand type. It would be an understatement to say that GIB does not handle two-suiters well. A robot can not claim to play bridge and bid this way. Anything that calls itself an upgrade has to correct these 2 problems, imo.
More accurate definitions would be right up there.

http://tinyurl.com/ya4akhbv

http://tinyurl.com/y7h7rjes

http://tinyurl.com/yb52y5fw

I’d much rather find something to cheer about than complain

zhasbeen
Hand 1 i think your bid is 3H over 2S bidding out your shape. This has worked very well for me. I rarely reverse without a stiff though.
Hand 2 you have 18 HCP, yet 1 NT. When dealing with GIBidiot I try to be as explicit as possible. After 3D, nobody would fault you for bidding 4D.
Hand 3 my sympathies. Who knows what GIBBO's addled brain is thinking. I would bid 3C. Looks like GIBidiot considers 2S as forward going and showing 5.

I could be wrong on the third one and it is a bug.

vrock

[zhasbeen]

virgosrock, on 2018-May-11, 06:25, said:

zhasbeen
Hand 1 i think your bid is 3H over 2S bidding out your shape. This has worked very well for me. I rarely reverse without a stiff though.
Hand 2 you have 18 HCP, yet 1 NT. When dealing with GIBidiot I try to be as explicit as possible. After 3D, nobody would fault you for bidding 4D.
Hand 3 my sympathies. Who knows what GIBBO's addled brain is thinking. I would bid 3C. Looks like GIBidiot considers 2S as forward going and showing 5.

I could be wrong on the third one and it is a bug.

vrock

Hand 1 - my thinking:

I was about as light as I could be for reverse but I had choice of that, rebidding 2C, or opening 1D and rebidding 2C. Take your pick. I’ve tried the latter several times and it usually turns out bad. The reverse seemed like lesser of evils to me and it shows longer clubs than diamonds, typically played as 1-round force rather than GF. I realize that the definition with 2S bid says “forcing to 3NT” but I’ve learned to take that definition with a large grain of salt.
I like pass of 2S more than the 3H call with only 3. The 2S call is the real stinker, imo. My 2NT rebid says that I have at most 2S and a strong implication that I have at least 2 hearts with a useful honor. I would have bid diamonds again with a 5th one, so probably 1-3 or 2-2 in majors.

Hand 2 - The note I made with this link says “cruel and unusual punishment” in reference to what GIB did to me for opening 1NT with 18 hcp. This hand was a setup made to punish people like me, he-he…

Hand 3 - KQ is normally good, if not great, support for a major that has been rebid—even by GIB. As it was, I did make 4. GIB’s ignoring hearts was horrendous bidding nonetheless


One thing about playing with GIB and these tournaments in general, is that it gives you lots of interesting hands to think about and play. Great practice if you are able to put the competitive aspect out of your mind.

[virgosrock]

Vampyr, on 2018-May-11, 07:02, said:

Yes, the assumption that people who don’t come on to the forums to complain are satisfied with GIB is either foolish or disingenuous. And this argument that many players are weaker than GIB, so it doesn’t matter, is doing less-experienced players a great disservice, because GIB is useless as a learning tool; why wouldn’t a learner think that the things GIB does is correct? And this learner would not even know there was a reason to complain.

I have learn(t, ed) a lot from GIBBO vis-a-vis bidding but only when no competition. With competition GIBBO is a BIG ZERO.
Card Play also is fascinating by GIBBO. Lot of learning here. Goes off the rails about 15% of the time and that is my reading and I am not an expert.
Play of doubled contracts generally superb.

Sorry to disagree with those who don't think thus.

vrock

[manudude03]

virgosrock, on 2018-May-09, 12:44, said:

2c-2d-4h as well?

vrock

Almost, 2♣-2♦-4♥ drops the "opener's suit" part of the desciption. Everything else is the same though.

[steve2005]

jdonn, on 2018-May-09, 10:31, said:

"here we are in 2018 with no noticeable improvements in GIB"

I know it's just frustration speaking, but comments like that are obviously untrue and don't contribute to the discussion. There have been nearly 40 I think upgrades in those years, with several thousand problems worked on or fixed.

The fixes made are mostly to bugs that have been found. There also have been some improvements to bidding database and even some new conventions.
There are however many flaws in Gibs bidding. Without interference the problems are livable. and the real problems people complain about can't be fixed without a huge amount of work. In interference Gibs bidding and the choices offer to the humans is often just not even close to playable. Again these errors are too numerous to list. Almost all of these competitive bidding problems I don't believe can ever be fixed by the current methods. The entire bidding part of the program would have to rewritten to fix competition problems.

jdonn, on 2018-May-09, 10:31, said:

The fact is, there are a few dozen frequent complainers about GIB out of tens of thousands of users. Now, many of those complaints are quite valid (many are not, but that doesn't really matter). I am certainly not advocating for GIB jumping to a grand slam on this hand (although if I was forced to guess the final contract with his hand over 4H for my life it would probably be 7H). And problems do get worked on, albeit not at a pace most would prefer.

For some of this errors it is irrelevant how many people mention a bug it should be fixed it is so fundamental
For example a while back If Gib was in 7NTX with an Ace and on lead. Gib would never cash it and would allow contract to make. Even now there are cash-out situations where Gib can set contract and doesn't. I do understand that Gib may be trying to get -2 by not cashing but in the 7NTX case it clear there was something wrong with the part of the Gib program which chooses a card to play on defense.This fundamental problem was never fixed. Instead Gib was made to lead an Ace if this happened. But whatever was causing this is still there.
Sure, didn't happen often but when it did it was repeatable and everyone in 7NX made but those in 7 suit went down.
Can you imagine if you went to Starbuck and the coffee machine dispensed tea once in a while. They would fix it or replace it. No I do not expect perfection from Gib, but I would tell you some of the things Gib does if a human partner did this I would not be playing with them again unless they were my wife <sarcasm>. or best friend.
People using the commercial Gib program have found cases where BBO Gib makes a play even a beginner wouldn't but commercial Gib makes the correct play. I believe a fundamental bug has been introduced into at least the defensive play of Gib.

jdonn, on 2018-May-09, 10:31, said:

But the fact is, BBO has made the decision long ago to stick with GIB. It has a ton of repeat users who seem satisfied with what they get. I have been receiving these threats to stop using GIB altogether for years, and while I'm sure some people have done so (which is not good of course), the vast majority of users clearly won't. Don't forget, most GIB users are on average worse at bridge than GIB is, often much worse. BBO is very honest and forthright about what you are getting with GIB. Sorry but this is how it is, for the foreseeable future.

That's good. And your right Gib is better than the majority of users. Even good players may not be used to playing with Gib or playing with an unfamiliar partner.
But BBO is trying to increase it's presence with ACBL and some other NBO's which i want too. Those players will be better players on average. All the mistakes Gib makes that a beginner would not make are giving bad press for BBO and won't help increase attendance in this area or increasing number of special tournaments being held on BBO. Even before anyone played in the NABC robot championship there were people who wanted to stop this sort of tournament as it wasn't real bridge.

[virgosrock]

steve2005, on 2018-May-19, 23:41, said:

[/size][/font][/color]
[/left]The fixes made are mostly to bugs that have been found. There also have been some improvements to bidding database and even some new conventions.
There are however many flaws in Gibs bidding. Without interference the problems are livable. and the real problems people complain about can't be fixed without a huge amount of work. In interference Gibs bidding and the choices offer to the humans is often just not even close to playable. Again these errors are too numerous to list. Almost all of these competitive bidding problems I don't believe can ever be fixed by the current methods. The entire bidding part of the program would have to rewritten to fix competition problems.

For some of this errors it is irrelevant how many people mention a bug it should be fixed it is so fundamental
For example a while back If Gib was in 7NTX with an Ace and on lead. Gib would never cash it and would allow contract to make. Even now there are cash-out situations where Gib can set contract and doesn't. I do understand that Gib may be trying to get -2 by not cashing but in the 7NTX case it clear there was something wrong with the part of the Gib program which chooses a card to play on defense.This fundamental problem was never fixed. Instead Gib was made to lead an Ace if this happened. But whatever was causing this is still there.
Sure, didn't happen often but when it did it was repeatable and everyone in 7NX made but those in 7 suit went down.
Can you imagine if you went to Starbuck and the coffee machine dispensed tea once in a while. They would fix it or replace it. No I do not expect perfection from Gib, but I would tell you some of the things Gib does if a human partner did this I would not be playing with them again unless they were my wife <sarcasm>. or best friend.
People using the commercial Gib program have found cases where BBO Gib makes a play even a beginner wouldn't but commercial Gib makes the correct play. I believe a fundamental bug has been introduced into at least the defensive play of Gib.

[color="#000000"][font="arial,verdana,tahoma,sans-serif"][size="3"][left]
That's good. And your right Gib is better than the majority of users. Even good players may not be used to playing with Gib or playing with an unfamiliar partner.
But BBO is trying to increase it's presence with ACBL and some other NBO's which i want too. Those players will be better players on average. All the mistakes Gib makes that a beginner would not make are giving bad press for BBO and won't help increase attendance in this area or increasing number of special tournaments being held on BBO. Even before anyone played in the NABC robot championship there were people who wanted to stop this sort of tournament as it wasn't real bridge.

Steve 2005 its quite simple. Till it hits the pocketbook no major changes are likely to be made. This does not imply Sir Donn not performing his knightly duties. Changes needed are major and need human/compute resources. So, you keep MB players quiet by having a fairly low table fee. Sheeple are so sh*t scared and overawed by the game they won't or may not even notice the finer points or don't have the time. So that leaves the cognoscenti who are few in number. Thus I am not seeing any major motivation to make any major changes.

vrock