Bbradley62 Posted January 24, 2015 Report Share Posted January 24, 2015 From a recent ACBL Robodupe[hv=bbo=y&lin=pn|human,~~M43260,~~M43258,~~M43259|st%7C%7Cmd%7C3SQAH8JD478KAC348J%2CS49H57QKAD56C679T%2CS3678H2349D39TQCK%2C%7Crh%7C%7Cah%7CBoard%201%7Csv%7Co%7Cmb%7Cp%7Cmb%7Cp%7Cmb%7C1N%7Can%7Cnotrump%20opener.%20Could%20have%205M.%20--%202-5%20%21C%7Cmb%7Cp%7Cmb%7Cp%7Cmb%7C2S%21%7Can%7CCappelletti%20-%20spades%20and%20a%20minor%20--%203-%20%21H%3B%204%2B%20%21S%3B%2011-%20HCP%3B%2011-12%20total%20points%20%7Cmb%7C3D%7Can%7C2-5%20%21C%3B%202-5%20%21D%3B%202-5%20%21H%3B%202-5%20%21S%3B%2015-17%20HC%7Cmb%7Cp%7Cmb%7C3H%7Can%7C8-%20HCP%3B%20twice%20rebiddable%20%21H%3B%208-9%20total%20points%20%7Cmb%7Cp%7Cmb%7Cp%7Cmb%7Cp%7Cpc%7C]360|270[/hv]First, the description of 3♦ is absurd; clearly, it shows diamonds. Second, 3♥ cannot mean what the description says, since North passed over 1N. Third, why isn't North happy to sit at 3♦? (Yes, the answer to the third question is probably that he believes the description provided for 3♦.) Quote Link to comment Share on other sites More sharing options...
1eyedjack Posted January 24, 2015 Report Share Posted January 24, 2015 I expect that 3D is a systemically impossible bid on the grounds that GIB cannot imagine a hand consistent with 1N opener that might want to back in with 3D opposite a passing partner in non-protective seat with opponents bidding. I am not commenting on whether GIB is *right* to lack that imagination, but continuing on that assumption, the question then arises what assumptions GIB is supposed to make when it hears a bid from partner that goes off the end of the scale. That question does not just apply to this auction, of course. A 3-level response to Stayman would be another example. It looks like GIB just falls down, gives up, ignores the bid and just goes by the last definition. Quote Link to comment Share on other sites More sharing options...
Bbradley62 Posted January 24, 2015 Author Report Share Posted January 24, 2015 Yes, you are almost certainly correct that GIB basically fails to have any useful assumptions about partner's "impossible" bid, and this is a system-wide problem that could/should be addressed. And it's been noted that GIB also sometimes fails to incorporate prior passes into explanations (such as providing the given explanation of 3♥ despite passing opposite 1N), which should also be addressed. And the part of the "brain" that makes bids like 3♥ should better communicate with the part of the brain that explains such bids. Quote Link to comment Share on other sites More sharing options...
Stephen Tu Posted January 24, 2015 Report Share Posted January 24, 2015 To be fair this auction really isn't supposed to exist. I'd want them to prioritize fixing the definitions for auctions that are plausible and common before working on filling in reasonable meanings for impossible auctions which are practically infinite. Quote Link to comment Share on other sites More sharing options...
1eyedjack Posted January 24, 2015 Report Share Posted January 24, 2015 I'd want them to prioritize fixing the definitions for auctions that are plausible and common before working on filling in reasonable meanings for impossible auctions which are practically infinite.If it were possible to set a 1-size-fits-all algorithm for such occasions it might not require individually programming into each of the individual "practically infinite" auctions. One possibility would be to require GIB to pass. That would be the simplest algorithm, and almost certainly could be improved upon, but even that would probably be an improvement on permitting GIB free reign to make its best guess what to do. As to priorities, I agree that it is low. Certainly not worthwhile if just considering an individual occasion, but the frequency of a default algorithm spanning a vast array of possibilities may change things. Quote Link to comment Share on other sites More sharing options...
Bbradley62 Posted January 25, 2015 Author Report Share Posted January 25, 2015 It shouldn't be too hard to program "if I open NT and then bid a suit on my own, I have 5 of them and not 4 of anything else". Quote Link to comment Share on other sites More sharing options...
1eyedjack Posted January 27, 2015 Report Share Posted January 27, 2015 It shouldn't be too hard to program "if I open NT and then bid a suit on my own, I have 5 of them and not 4 of anything else".It may not be possible to program that in without at the same time telling GIB under what circumstances the robot should be freely introducing a suit in a contested auction having opened 1NT opposite a passing partner when the robot has your hand and is the 1N opener. That may not be so trivial. Quote Link to comment Share on other sites More sharing options...
georgi Posted January 27, 2015 Report Share Posted January 27, 2015 Thanks for reporting. Fixed in v33 about opener's possible continuation over cappelletti ( assuming both others have passed before or after intervention). GIB won't do, but it will recognize the length. It will promise 5 cards and balanced shape. In the example -> 3♦ -> -- 2-3 ♣; 5 ♦; 2-4 ♥; 2-4 ♠; 17 HCP; 18- total points 3H would be -> -- 3- ♦; 4+ ♥; 8- HCP; 8-9 total points But GIB will be more eager to stay based on 5 cards. Quote Link to comment Share on other sites More sharing options...
iandayre Posted January 28, 2015 Report Share Posted January 28, 2015 Thanks for reporting. Fixed in v33 about opener's possible continuation over cappelletti ( assuming both others have passed before or after intervention). GIB won't do, but it will recognize the length. It will promise 5 cards and balanced shape. In the example -> 3♦ -> -- 2-3 ♣; 5 ♦; 2-4 ♥; 2-4 ♠; 17 HCP; 18- total points 3H would be -> -- 3- ♦; 4+ ♥; 8- HCP; 8-9 total points But GIB will be more eager to stay based on 5 cards. I disagree Georgi. If GIB has 3 or even two of the suit opener bids, it should NOT attempt to escape to a 4 card suit. In fact in all auctions GIB should not be voluntarily introducing 4 card suits at the 3 level or higher. Quote Link to comment Share on other sites More sharing options...
Bbradley62 Posted January 28, 2015 Author Report Share Posted January 28, 2015 In the example -> 3♦ -> -- 2-3 ♣; 5 ♦; 2-4 ♥; 2-4 ♠; 17 HCP; 18- total points 3H would be -> -- 3- ♦; 4+ ♥; 8- HCP; 8-9 total pointsI disagree Georgi. If GIB has 3 or even two of the suit opener bids, it should NOT attempt to escape to a 4 card suit. In fact in all auctions GIB should not be voluntarily introducing 4 card suits at the 3 level or higher.In the example, responder cannot have 4+♥ and 8-9 total points, because he didn't bid Stayman over 1N. The only possible meaning I can figure for 3H would be: 4+♣; 1-♦; 4♥; 4♠; 7- total points. (It should be responder's only possible non-pass call over 3♦.) Opener probably doesn't have four of any of ♣/♥/♠, and responder shouldn't be eager to run to a 3-3 fit. Quote Link to comment Share on other sites More sharing options...
georgi Posted February 6, 2015 Report Share Posted February 6, 2015 GIB won't introduce 4+ that easily, but in theory it's possible to have such hand( as per how GIB uses Stayman to have 8TP ). 9TP unlikely. I.e. the hand xx.KQxx.xx.Jxxxx will pass Stayman Later won't bid 3♥, but it fullfils the requirements for 8TP and 4+. Quote Link to comment Share on other sites More sharing options...
iandayre Posted February 7, 2015 Report Share Posted February 7, 2015 Well certainly we should all be able to agree that with 3 of the suit partner bids, GIB should never correct to a 4 card suit. And your description should reflect that. Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.