[uva72uva72]

My link

IMPs, ACBL robot individual

I limited my hand by failing to superaccept North's transfer. Then, after looking at the system notes, I selected the bid that showed the weakest hand with ♠ support. Yet the robot still invoked the ineluctable Blackwood on its way to a failing 5♠ and the loss of 11 IMPs to those who were permitted to play 4♠. Paradoxically, those players who made the stronger bid of 3♠ vice my 4♠ were let off the hook when they failed to respond to North's cue bid. I'm guessing that traces to North's reported amnesia about previous rounds of bidding.

I'm curious to know what hand the robot could have constructed, consistent with my bidding, that would have yielded a good play for slam opposite North's collection. But mostly I'm curious to hear about what can be done to prevent North from invoking Blackwood where there is so little likelihood of having a good play for slam.

[Stephen Tu]

The explanation was cut off, did the jump to 4s show a 4th spade card? Did it limit the hand to 15 HCP? It's probably hard to make rules that constrain opener's hands to "bad hands for slam" in GIB's language (hands with mostly slow values in the red suits).

In my opinion opening NT bidder should never really jump to 4 opposite an unlimited responder on this auction. It deprives responder of valuable cue bidding room, and then they have choice of blkwood, 5 level cue bid, or never showing slam interest, all of which can work out poorly. 3S shouldn't be "stronger" than 4s, it should just show support and be unspecific as to strength. 4S can be defined as minimum to discourage exploration if someone chooses it for GIB's rule base, but I personally just never take that call on this auction.

Can you have Qxxx KQx Axxx Ax from its point of view?

[1eyedjack]

Stephen's not going to like this. I certainly don't.

I am not completely certain whether there have been any changes in this area between versions 30 and 31. I rather suspect not, and I can tell you how version 30 behaved:

The 4S here shows any 1N opener with fewer than 4 Clubs, more than 2 Spades and 15-16 HCP.

A 3S bid would have shown 3 Spades precisely and 17 HCP
A 3N bid would have shown any hand with 2 Spades and fewer than 4 Clubs
With 4+ Spade support and 17 HCP you would have super-accepted on the previous round
All suit bids other than Spades promise a Club fit. There MAY be some reserved to show a double fit (there were in previous versions but there have been changes between versions here and I am no longer certain what is current).

So there you have it. "Fast arrival" has been programmed into GIB despite that responder is unlimited. Hope you are not all full of Xmas pud. Waste of prime vomit.

In answer to OP, I can construct a few hands where slam is good. Stephen has posted one. You are also quite unlucky to go down in 5 on the posted hand. That said, I agree that you don't want to be there.

[ArtK78]

How can you go down in 5♠? The obvious way to play the club suit is lead the 10 off of dummy on the first round of the suit, planning to play West for the ♣9 if the 10 is covered.



Having played with the robots for a long time, I would have bid it the same way. 4♠ is the correct system bid on the South hand.

[Bbradley62]

I believe bids show the following combinations of spades and HCP

.......High Card Points
Spades 15 16 17
3......4S 4S 3S
4......4S 4S SA
5......4S 4S open 1S

Maybe that chart should be changed to something like:

.......High Card Points
Spades 15 16 17
3......4S 4S 3S
4......4S 3S SA
5......3S SA open 1S

Of course, the descriptions provided would still be misleading, since the 3♠ rebid cannot be described as "20 Pearson points"...

[1eyedjack]

That may well be an improvement on the existing method, but I think that the whole idea of fast arrival when responder is unlimited is flawed. May work fine if responder is a passed hand.

If they were minded to change the programming in this area I would rather they rethought the whole FA principle, rather than tweak the continuations within those constraints

[Bbradley62]

Fine by me... I am completely indifferent to the difference between the proposed chart I posted above and one in which the three "4S" bids are exchanged with the three "3S" bids. I would just like to see some differentiation. Maybe the two bids could be described as "min with support" and "max with support" rather than trying to give specific details as to the distinction. (Of course, it would be nice if the actual distinctions are explained in the System Notes.)

[ArtK78]

If you are playing in an ACBL robot game on BBO, responder is NOT unlimited. You are dealt the best hand, so responder cannot have a hand better than yours.

[1eyedjack]

ArtK78, on 2014-December-26, 09:58, said:

If you are playing in an ACBL robot game on BBO, responder is NOT unlimited. You are dealt the best hand, so responder cannot have a hand better than yours.

And you think that this factor should be programmed into GIB's system? Actually he is effectively unlimited for all purposes pertaining to this thread if, despite having fewer than 17 HCP by the conditions of contest, he still may be strong enough to try for slam.

[1eyedjack]

Bbradley62, on 2014-December-26, 09:06, said:

Fine by me... I am completely indifferent to the difference between the proposed chart I posted above and one in which the three "4S" bids are exchanged with the three "3S" bids. I would just like to see some differentiation. Maybe the two bids could be described as "min with support" and "max with support" rather than trying to give specific details as to the distinction. (Of course, it would be nice if the actual distinctions are explained in the System Notes.)

I certainly would NOT support a change which simply swapped the 4S and 3S bids.

IMO it would be an improvement over either proposal, albeit still less than optimal, if the 4S rebid were simply removed altogether from the repertoire and you were forced to rebid 3S on all hands with a fit. Responder can show a slam try, and a descriptive one at that, without committing beyond 4S, which slam try can be rejected by opener with a minimum or unsuitable hand, again without committing beyond 4S, and responder STILL has the option to progress beyond 4S in light of that information if so inclined.

That would clearly be sub-optimal as logically there must be some use for the 4S rebid which, if sufficiently narrowly defined, would take some pressure off other rebids while preserving safety in continuations. The key is in defining the bid with a narrowness that befits the consumption of bidding space.

[iandayre]

I don't see a critical need to change methods here. What is needed is improved judgement on GIB's part, it being able to determine that no slam try should be made opposite a minimum opener with support.


Here is somewhat similar hand:

http://tinyurl.com/ncj8csx

Despite me denying 3 Hearts or good C support, GIB bid on to the NT slam that is poor if the Hearts run, and pretty much hopeless if they do not. And I had a prime, full-value hand. Weren't we told that GIB would be less aggressive in bidding NT slams? Can't see that here.

And to revisit another theme - who got the great scores on the hand? Two players who opened 1NT with 20 real, and one who opened 1D, bid all of 1S over 1H, and mustered all their courage to bid 3NT over 1NT.

[Bbradley62]

Same sequence start... I wanted to raise clubs... 4♣ was explained as support w/17HCP and 5♣ was explained as showing a double ♣♥ fit... I didn't poke around to see other explanations... Is there a way to raise clubs without showing a max? (On this hand I wanted to show one, but there might be other hands where I don't.)

---

Edit: Oops... My bad... 2♥ accepted the xfer and showed heart support over the double. I didn't see the double during the auction. But then 4♣ appears to say "I lied about the support".

This post has been edited by Bbradley62: 2014-December-26, 15:53

[ArtK78]

iandayre, on 2014-December-26, 13:34, said:

I don't see a critical need to change methods here. What is needed is improved judgement on GIB's part, it being able to determine that no slam try should be made opposite a minimum opener with support.


Here is somewhat similar hand:

http://tinyurl.com/ncj8csx

Despite me denying 3 Hearts or good C support, GIB bid on to the NT slam that is poor if the Hearts run, and pretty much hopeless if they do not. And I had a prime, full-value hand. Weren't we told that GIB would be less aggressive in bidding NT slams? Can't see that here.

And to revisit another theme - who got the great scores on the hand? Two players who opened 1NT with 20 real, and one who opened 1D, bid all of 1S over 1H, and mustered all their courage to bid 3NT over 1NT.

One of the things I have learned the hard way is to NEVER (OK, almost never) accept an invitation offered by GIB.

GIB is insanely aggressive in slam auctions, and very aggressive in game auctions.

However, sometimes GIB's aggressiveness is due to a bid that you made assumking it meant something rational, but, upon reviewing the meaning of your own call, you find that you have shown 256 HCP and a minimum of a solid 12 card suit. You have to be careful, especially when you make jump bids. I often find myself making what seem to be nonforcing bids when I want to bid a game, knowing that if I bid game GIB will bid slam.

[ArtK78]

Bbradley62, on 2014-December-26, 15:10, said:

Same sequence start... I wanted to raise clubs... 4♣ was explained as support w/17HCP and 5♣ was explained as showing a double ♣♥ fit... I didn't poke around to see other explanations... Is there a way to raise clubs without showing a max? (On this hand I wanted to show one, but there might be other hands where I don't.)

---

Notice that the North hand - GIB's hand - is the type of hand that most players would, at most, transfer to hearts and bid 2NT, believing that partial or game is the limit on the hand and the issues are (1) whether to play in hearts or notrump and (2) whether to play part score or game. Some might transfer to hearts and pass, trying to go plus when partner doesn't preaccept hearts. But GIB makes what most players would think was a game forcing 3♣ bid, showing that powerhouse QT9x second suit. Of course, we find out later that the 3♣ bid is only forcing to 3NT - GIB can (and did) pass 4♣.

[iandayre]

ArtK78, on 2014-December-26, 15:37, said:

One of the things I have learned the hard way is to NEVER (OK, almost never) accept an invitation offered by GIB.

GIB is insanely aggressive in slam auctions, and very aggressive in game auctions.

However, sometimes GIB's aggressiveness is due to a bid that you made assumking it meant something rational, but, upon reviewing the meaning of your own call, you find that you have shown 256 HCP and a minimum of a solid 12 card suit. You have to be careful, especially when you make jump bids. I often find myself making what seem to be nonforcing bids when I want to bid a game, knowing that if I bid game GIB will bid slam.

Quite so. In this case 4NT was my signoff and GIB bid on anyway.

[iandayre]

Bbradley62, on 2014-December-26, 15:10, said:

Same sequence start... I wanted to raise clubs... 4♣ was explained as support w/17HCP and 5♣ was explained as showing a double ♣♥ fit... I didn't poke around to see other explanations... Is there a way to raise clubs without showing a max? (On this hand I wanted to show one, but there might be other hands where I don't.)

---

Edit: Oops... My bad... 2♥ accepted the xfer and showed heart support over the double. I didn't see the double during the auction. But then 4♣ appears to say "I lied about the support".

It's true you erred by bidding 2H. But then, if you have shown 3 or more H, why does GIB pass 4C? Makes no sense.

As far as what to do when GIB transfers, then bids a minor, any new suit at the 3 level shows support for the minor. I don't believe that a max is required but I haven't seen this auction in the latest release.

[steve2005]

4N is listed as 17+ pts. Gib has 14 hcp and singleton with wasted Q♦ has no business bidding Blackwood opposite a known min. If opener has Aces Gib is looking for they don't bid 4♠.