smerriman

I guess theoretically West could have an extra spade in place of a small club. If GIB could only simulate one matching hand and that was it, that could lead to a diamond return. But that's clutching at straws. GIB is broken.

GIB deducts a point for honours in short suits. So it only counts this as 8 points and thus doesn't continue. 3♦ then 4♥ shows "Mild slam try. No shortness - 2+ ♣, 2+ ♦, 6+ ♥, 2+ ♠, 7-8 total points.". Which seems to describe the hand perfectly; not sure why it wouldn't bid this.

Well, I'd have to disagree completely - that's true if you're dealt a long suit in normal bridge, but the way Goulash is dealt involves everyone having freak hands. This is especially true on computer-dealt Goulash hands, which are normally restricted based on a 'freakiness' factor. Using a non-goulash bidding system is never going to work, but always opening a long suit at a high level has its downsides too - you're not only freezing out the opposition; you're freezing out your partner. Your partner needs to know when to compete, when to raise, when to penalise (since the opposition are likely to compete regardless). Obviously there's a lot of luck involved, but regardless of how you play, you should at least have some definition of when you open low (eg obviously it'll never be a standard 5 card suit in a balanced hand). I like to do so with a four card side suit, to at least give us a chance of finding a secondary fit.

All 13 cards lead to the same result of 2C=?

No, no definition at all (this part was also replicated at a bidding table). Have seen humans get into trouble making undefined bids, but not GIB..

Nothing to do with resources, just the usual over-trusting of opponents. South has "7 spades" and leads the queen - no harm in therefore ducking to the singleton king left in north.

6NT shouldn't be too hard to see. Duck a club, then play your club (throwing a spade) and four spades. West has to throw a red card, which gives you your 12th trick. Having seen that, play 7♠ in exactly the same way, except reverse the first two tricks. Win the club, play a club ("throwing" a spade). (In fact, you can just ruff two clubs, without even needing the squeeze, hah.)

http://www.bridgebase.com/forums/topic/78455-capp-leb-confusion/

That's the problem. I proved that it sometimes doesn't make signals like that, on occasions where it cannot possibly cost. That is very easy to program (if double dummy shows 0 difference between the options, opt for the signal), and is how the developers said it was meant to be programmed, yet it still doesn't :(

Yeah, I know, I did the analysis :P But the conclusion was that the fact that it signals wrong a large amount of the time, plus the fact that it's impossible to read anyway, makes it equivalent to assuming no signals at all. In this specific case, your partner might have: - a singleton - any doubleton, since it plays low from doubletons (even small doubletons) 50% of the time - Kxx or better - it normally won't play the 2 from this, but occasionally forgets for no reason - basically any other hand. I wish I could get a response to that thread as to why GIB is so buggy in this situation :(

[hv=http://www.bridgebase.com/tools/handviewer.html?sn=smerriman&s=SA732HK85DQCKQT53&wn=Robot&w=SJ9HQT6DAT763C862&nn=Robot&n=S864HA93DKJ854CJ7&en=Robot&e=SKQT5HJ742D92CA94&d=n&v=n&b=5&a=PP1C(Minor%20suit%20opening%20--%203+%20%21C%3B%2011-21%20HCP%3B%2012-22%20total%20points)P1D(One%20over%20one%20--%204+%20%21D%3B%2011-%20HCP%3B%206-12%20total%20points)1S(5+%20%21S%3B%207-11%20HCP%3B%208-12%20total%20points)1N(3-5%20%21C%3B%202-5%20%21D%3B%202-4%20%21H%3B%202-4%20%21S%3B%2012-14%20HCP%3B%20stop%20in%20%21S)PPD(Balancing%20double%20--%205-%20%21C%3B%205-%20%21D%3B%204+%20%21H%3B%205+%20%21S%3B%2011-%20HCP%3B%2010-12%20total%20points)PPP&p=C2CJCAC3SKSAS9S4DQD3D4D2CKC8C7C4CQC6S6C9CTD7S8D9C5D6H3H7H5HTHAH4DKH2S2DASJD5SQS3STS7H6D8S5H8DTH9HJHKHQDJ]400|300[/hv] GIB east makes a 4 card overcall, then a balancing double of 1NT, which gets passed out. Not the greatest result there :)

Wow. So much conflicting advice. I'm with those who say that you should take the safety play only if it gains > 50%. This doesn't make sense. Nobody in 4♥ is going to take a safety play to make 9 tricks when "you are likely to make 10 tricks" as stated. Therefore you'll never tie anyone in game - whether or not you take the safety play will never affect your scores against those in 4♥. This also doesn't make any sense. You should make an "abnormal" choice when it turns a bad result into a good result (compared with the normal contract). In this case, the safety play keeps a bad result a bad result, and a good result a good result. You should take the safety play if it is the "normal" choice - otherwise all it does is make your score worse compared with the others in 3♥.

GIB doesn't play defensive carding. Take a wild guess.

After 1♥ - 1♠, what would your next bid be? I open 1NT because I don't know how to answer that question (and accept the invite, easily).

GIB will always play identically if everything up to that point is the same. The same random seed is used at both tables.

Not quite sure what your question is. NMF applies over 1♥ - 1♠? In fact, standard is to play both 2♣ and 2♦ as NMF, though GIB only uses 2♣: https://en.wikipedia.org/wiki/New_minor_forcing

They don't beat you. You're compared against a selection of players who played the original tournament - not against others playing the instant tournament.

You don't believe me, because of all of the other threads referring to what people think is the worst bug ever. But I actually mean it. [hv=http://www.bridgebase.com/tools/handviewer.html?sn=smerriman&s=SAT862HAD984CA965&wn=Robot&w=SKHJ8764DKCKJT842&nn=Robot&n=S5HT952DAJT732C73&en=Robot&e=SQJ9743HKQ3DQ65CQ&d=s&v=e&b=3&a=1S(Major%20suit%20opening%20--%205+%20%21S%3B%2011-21%20HCP%3B%2012-22%20total%20points)2S!(Michaels%20--%205+%20%21H%3B%2012+%20total%20points%3B%20forcing)3D(Natural%2C%20not%20forcing%20--%202-%20%21S%3B%20twice%20rebiddable%20%21D%3B%207-9%20total%20points)4SP5C(5+%20%21H%3B%20strong%20rebiddable%20%21C%3B%2029-30%20total%20points)P7H(3+%20%21H%3B%207+%20total%20points)D(5+%20%21S%3B%2019-21%20HCP%3B%2022-%20total%20points)PPP&p=SASKS5S9CAC4C3CQD9DKDAD6D7DQD4C2H3HAH6H9C5C8C7S4H4H2HKS2HQD8H7H5D5S6H8DJHJHTS7C6CKD2S3C9CJDTSJS8CTD3SQST]400|300[/hv] Basic robots, 100% replicable at a bidding table.

According to the descriptions, bidding 2♥ shows you're a maximum of the 0-4 range. 4 points, 5 clubs, 4 hearts actually seems to make game (double dummy) a surprising proportion of the time despite the Moysian fit; I guess due to the fact that every finesse is guaranteed to work. Looks crazy and I don't think it's right but I can actually see why GIB bid it.

I wouldn't say it's impossible to figure out that if 4 bananas represents the number 4, 3 bananas represents the number 3. Sure, you could define a picture of 3 bananas to mean 28 if you really wanted to, but on the silliness scale, I'd rate that sillier than some other uses of the word silly here :)

Puzzle #2 - how were you meant to get 19 or 26?

Maybe GIB assumed South would never lead 7 from 87 doubleton. Who knows :/

.. but whether it's GIB or not makes no difference whatsoever if this comes up in a human context.

Double dummy analysis is never completely relevant, but since this was GIB (came from free weekly instant): - assuming double dummy lead and play - assuming GIB has >= 3 spades, 11 or less HCP, 10-12 total points (GIB definition) - 500 sims Case 1: exactly 3 spades, no singleton 3NT scores 25% in MP over 4S Case 2: exactly 3 spades, no singleton, at least half a stopper (hcp + length >= 4, GIBs definition) in each side suit 3NT scores 28% in MP over 4S Not finding the killing lead may skew this one way or the other, and maybe there are better definitions of when to sit, but surely not enough to make 3NT worth it.