[tx10s]

Can anyone explain why bidding 4 of a suit over a 3NT bid to show shortness is a good idea? This bid comes up every time a robot bids 2S over partner's 1NT opener to show 5-4 in the minors and 3S over a 2NT opener. Every time this bid has come up to me, I have never had a fit for the minor suits and the robot has always had a 3-1 distribution in the majors. I figure by bidding NT, the player is showing no fit for the minors and having the major suits covered. When the program insists on bidding its short suit, the player is then stuck trying to bid a bad fit at the 5 level. The only option is to bid your 3 card suit at the 5 level, but that has never yielded a good result. I know you can get around this by bidding 2NT after a 1NT-p-2S-p sequence, but there is not way to avoid this problem in a 2NT-p-3S-p sequence. Every time this has come up, the robot bid of 4 of a major has taken me out of a makeable 3NT contract. I once even tried to bid 4NT to play, but the robot then bid 5C, promising 6 but only holding 5 to the K nothing.

[Bbradley62]

I presume you're referring (at least in part) to this hand:

---

To help the conversation, I've posted the auction from another table in your tournament because GIB failed to provide explanations for your bids, which were the same as this South's.

[Bbradley62]

Here's another table with the same hand, with North misrepresenting his distribution differently:

---

Of course, it didn't help that South misunderstood 3♠, but I would suggest that South's 4♠ should show 5 spades, not 4-5, since North's 3♠ denies 4 spades. And, either North shouldn't bid 5♦ or the explanation should be changed.

[Bbradley62]

This South did what he was supposed to do, and that didn't work out so well...

---

[tx10s]

Yes, that is the hand I was referring to, but anytime the bidding has been 1NT-p-2S or 2NT-p-3S it has always had a similar result, so this is a more general question as to why the robot should bid showing a short suit after a 3NT bid by the player when the player already knows that the robot has at most 4 cards in the majors and still wants to bid 3NT. As a side note, I find it interesting that when the robot is 5-4 in the minors, it has always been 3-1 in the majors.

[steve2005]

Good example. 3N is no fit with major suits well stopped. Gib's minors are nothing to write home about and 3N should be ending the auction, bidding slam is just ridiculous.

[1eyedjack]

I don't reckon that this is among the more egregious of GIB's shortcomings.

First off, I don't know what 1N-2S-3N is supposed to show. I guess it should show something, but no way am I bidding anything other than 2N over 2S opposite GIB if I don't have a minor to support. Then when it bids its shortage I am still able to bid 3N. That does leak info to oppo, so I would like the option to go direct to 3N.

As for the above example hand where it pulled 3N with 5-5 in the minors after 2N-3S-3N, we would need to do some sims, but I can visualise myself bidding 3N on a lot of hands without a 4 card minor (and perhaps some with a 4 card minor) with the majors considerably less well stuffed than AKQ +AK, which spare values would then be available in the minors.

On a related matter, I only just got my $1.50 back on a 55% rebate tourney, partly as a result of this hand:

---

Personally I would very much have preferred it if partner HAD shown the heart shortage rather than the rebiddable Clubs. Maybe I should have gone for 3N anyway, but I would have been more inclined (maybe in error) to do so after a 3H shortage bid.

But I also wonder whether 2S is an appropriate response. Barely worth a GF, the likelihood of 5m being better than 3N I think is so low that use of 2S should be reserved for hands with rather more serious minor suit interest.

[Stephen Tu]

Agree with 1eyedjack. Over 2s 3nt should probably never be bid directly by GIB, I guess since some humans may bid this anyway it should be defined as minimum with lots of points in the majors, no minor fit, and try to discourage GIB from going on.

I think the main things to be tweaked here are:
- decision to bid MSS on hands that are fairly marginal as slam tries, it should probably be slightly stronger IMO to investigate.
- tighter constraints on opener's 4nt and 5m bids, to discourage robot from bidding on without significant extras, as these bids should be discouraging. GIB bids horribly over splinter bids in general. Also tighter constraints on auctions like 1nt-2s-2nt-3S-3nt, also trying to respect opener's attempted signoff without significant extra values. I think right now 3nt is merely described as "stopper", but it shouldn't be bid with ace empty holdings, or qx doubleton type holdings which I think I've seen GIB do. It should probably show something like 5+ hcp in the suit or something like QJTx.

[Bbradley62]

At least GIB cooperated on this one:

---