1NT rebid?

[gartinmale]

[handviewer='http://www.bridgebase.com/tools/handviewer.html?lin=pn|bubels,~~M15562,~~M15560,~~M15561|st%7C%7Cmd%7C2S267TJKHJQKAD5QC6%2CS3H578D2468C2378A%2CS9H469D7JKAC49JQK%2C%7Crh%7C%7Cah%7CBoard%208%7Csv%7Co%7Cmb%7Cp%7Cmb%7C1C%7Can%7CMinor%20suit%20opening%20--%203%2B%20C%3B%2011-21%20HCP%3B%2012-22%20total%20points%20%7Cmb%7Cp%7Cmb%7C1S%7Can%7COne%20over%20one%20--%204%2B%20S%3B%206%2B%20total%20points%7Cmb%7Cp%7Cmb%7C1N%7Can%7C3-5%20C%3B%202-5%20D%3B%202-4%20H%3B%202-3%20S%3B%2012-14%20HCP%20%7Cmb%7Cp%7Cmb%7C3N%7Can%7C4%2B%20S%3B%2017-%20HCP%3B%2013%2B%20total%20points%7Cmb%7Cp%7Cmb%7Cp%7Cmb%7Cp%7Cpc%7CH2%7Cpc%7CHA%7Cpc%7CH5%7Cpc%7CH6%7Cpc%7CHK%7Cpc%7CH7%7Cpc%7CH4%7Cpc%7CHT%7Cpc%7CC6%7Cpc%7CCA%7Cpc%7CC4%7Cpc%7CCT%7Cpc%7CC3%7Cpc%7CCK%7Cpc%7CC5%7Cpc%7CS2%7Cpc%7CD7%7Cpc%7CD9%7Cpc%7CDQ%7Cpc%7CD4%7Cpc%7CD5%7Cpc%7CD8%7Cpc%7CDA%7Cpc%7CDT%7Cpc%7CCQ%7Cpc%7CS4%7Cpc%7CS6%7Cpc%7CC2%7Cpc%7CCJ%7Cpc%7CS8%7Cpc%7CS7%7Cpc%7CC7%7Cpc%7CC9%7Cpc%7CS5%7Cpc%7CST%7Cpc%7CC8%7Cpc%7CDK%7Cpc%7CD3%7Cpc%7CSJ%7Cpc%7CD6%7Cpc%7CDJ%7Cpc%7CH3%7Cpc%7CSK%7Cpc%7CD2%7Cpc%7CH9%7Cpc%7CSA%7Cpc%7CHQ%7Cpc%7CH8%7Cpc%7CHJ%7Cpc%7CS3%7Cpc%7CS9%7Cpc%7CSQ%7C]400|300[/handviewer]

 

1NT, eh?

[calm01]

Aversion to re-bidding minors is another fairly common GIB behaviour that reinforces the view that original authors of GIB did not fully comprehend the first principle of bridge bidding - namely showing length first and strength as a by-product. In a 5-card major system when an opening minor can be 3 cards, it is even more important to show minor length if you have it.

 

This bidding error is probably a symptom of a generic design flaw that may be readily fixed once fully recognised.

[georgi]

Thanks for reporting. It's incorrect rebid indeed. It's fixed for the new update.

[barmar]

This bidding error is probably a symptom of a generic design flaw that may be readily fixed once fully recognised.

You seem to be assuming that GIB has general bidding principles. It just has an enormous database of explicit rules: "If the auction has shown X, and you have Y, bid Z". There's no "generic design" to fix, we just have to find the rules and modify them.

 

This hand causes problems for human bidders as well. Every bidding sequence for this hand has a flaw. 1♣-1♠-2♣ is supposed to show 6+ ♣. Some humans solve it by opening 1♦ and rebidding 2♣, but that should show ♦ at least as long as ♣ (GIB doesn't bid like this, but allowed for the possibility if the human bids it). And some people do what GIB does, and bid 1NT despite having a singleton.

[cloa513]

You seem to be assuming that GIB has general bidding principles. It just has an enormous database of explicit rules: "If the auction has shown X, and you have Y, bid Z". There's no "generic design" to fix, we just have to find the rules and modify them.

 

This hand causes problems for human bidders as well. Every bidding sequence for this hand has a flaw. 1♣-1♠-2♣ is supposed to show 6+ ♣. Some humans solve it by opening 1♦ and rebidding 2♣, but that should show ♦ at least as long as ♣ (GIB doesn't bid like this, but allowed for the possibility if the human bids it). And some people do what GIB does, and bid 1NT despite having a singleton.

The generic design fault is that it ignores those explicit rules due to what simulations say and those explicit rules weren't well thought out and documented.

[barmar]

The generic design fault is that it ignores those explicit rules due to what simulations say and those explicit rules weren't well thought out and documented.

There are over 4,000 rules in the database. Documenting them all would be an impractical task. It's too late to do anything about this now.

 

And the whole point of simulations is to handle the cases that are hard to describe using explicit rules. Overriding rules with simulations is a design FEATURE, not a fault. It would be virtually impossible for us to make explicit rules to handle all the possibilities, such as figuring out when your honors are well placed, which short suits are good or bad, etc.

 

The simulations do have problems, because sometimes it has a hard time dealing hands that match the bidding so far (especially if simulations were used in earlier bids). But trying to replace simulations with explicit rules is not feasible. We're working on other ideas, I can't go into details.

[fred]

Serious and open-minded students of the game will recognize this situation (ie how should opener bid in a natural system with a minimum hand and 4-5 in the minors) as one of which there is neither any good solution nor any real expert consensus. There are parts of the world in which each of the following basic strategies is considered to be "normal":

 

1) Always open 1C and rebid 2C

2) Always open 1D and rebid 2C

3) Always open 1C and rebid 1NT

4) Choose between 1, 2, and 3 above depending on factors like club suit quality, diamond suit quality, exact point count, and whether or not your singleton is an honor

 

For whatever its worth, if I had to choose between 1, 2, and 3, I would choose 3 (and hence make the same rebid that GIB did). That being said...

 

- I prefer to play a style in which judgment trumps rules (a style that is hard to teach to rule-based computer programs). So I will always really play some version of 4 (and include the possibilities of opening 1NT and Passing) even if our system notes and convention card say 3 (or 1 or 2).

- Since I like to think that I am both a serious and open-minded student of the game, I am not under any illusions that my preferred approach in this area is necessarily "best" from a pure bridge point of view and I know very well that there are many top experts who would disagree with my preferred approach (and with any other approach for that matter).

 

Regardless of what GIB chooses to bid with this particular hand (or other similar hands) there will be many players who will complain because they would consider one of the other alternatives to be "normal". This is a situation where our programmers basically can't win as long as there are GIB users out there who are sufficiently dogmatic in their thinking to believe that this is a non-problem for which their preferred solution is obviously correct.

 

The real lesson here is that even apparently simple situations in bridge can be very difficult (and in some cases have no good solution) even for the very best players. GIB deserves plenty of criticism, but please cut it (and our programmers) some slack when the action you disagree with falls into this category. If your bridge knowledge is insufficient to know when such a action falls into this category, then both your bridge game and your manners might benefit from a little humility.

 

Fred Gitelman

Bridge Base Inc.

www.bridgebase.com

[gartinmale]

Hmmph. As usual Fred's points are excellent. I don't claim to be anywhere near an expert, so I don't really have a leg to stand on here. I've seen GIB rebid 1NT in this auction (or in the auction starting 1C-1H) with singletons many times before and haven't had any issue with it; I never posted such a hand until this one because this one seemed particularly egregious to me (no honors in the majors, concentration in the club suit). But I could certainly see why someone following (3) might bid 1NT here, and why one might program GIB that way. I'll spend some more time thinking about why I reacted so poorly to this 1NT. I apologize if my opening salvo seemed presumptuous.

[fred]

Hmmph. As usual Fred's points are excellent. I don't claim to be anywhere near an expert, so I don't really have a leg to stand on here. I've seen GIB rebid 1NT in this auction (or in the auction starting 1C-1H) with singletons many times before and haven't had any issue with it; I never posted such a hand until this one because this one seemed particularly egregious to me (no honors in the majors, concentration in the club suit). But I could certainly see why someone following (3) might bid 1NT here, and why one might program GIB that way. I'll spend some more time thinking about why I reacted so poorly to this 1NT. I apologize if my opening salvo seemed presumptuous.

My diatribe wasn't directed at you - sorry that I didn't make this more clear.

 

It sounds like my thinking in this area is similar to yours. In general I am also fine with a 1NT rebid with this strength and pattern but like you I am also not thrilled about GIB's bidding with the actual hand in question.

 

Fred Gitelman

Bridge Base Inc.

www.bridgebase.com

[mike777]

Eric Rodwell, The Bridge World January 2012 from interview in 2009.

 

Main system principle is logic in all phases, also:

 

1) must be able to remember

2) system must have firm rules

3) system be impervious to preemption

4) long auctions are bad, dont exchange any more information than necessay.

 

"Many system-makers violate one or more of those principles."