Siegmund

Getting out at 4 is going to be hard unless responder has limited his hand and opener decides he is minimum enough to pass. Getting out at 5 - which on the posted cards is a pretty safe spot - ought to be pretty easy for any good cuebidding system, I would think.

From the GIB 21 sticky: "Tweaked the inverted minor sequence. GIB will now provide more accurate information and point limits as both opener and responder." From a robot tourney tonight: ♠AK52 ♥K8 ♦K86 ♣9842 Robot opens 1C (3+C, 11-21 HCP, 12-22 total points), I respond 2C, robot rebids.... 3C (4+C, 11-21 HCP, 12-22 total points.) Well, yeah, it does have 4+ clubs and 11-21 HCP, as it turns out. But I can think of three or four other bids it might have made on this hand that would have conveyed considerably more information to me. I would like for it to have bid 2NT with both majors stopped; I wouldn't have been surprised by a natural 2S or a cheapest-stopper 2D. What IS the bot's rebid style after an inverted raise? And whatever style it is, why did it decide repeating a 98xx suit was its best bid here?

More an interesting bridge hand than a bug report. The bots have a nasty habit of always finding the killing shift in situations where human players make obvious-but-wrong returns of partner's suit, etc. Tonight I knew what I was afraid of, and got fooled when a bot didn't see how to set me: http://www.bridgebase.com/tools/handviewer.html?lin=pn|Siegmund,~~M29663,~~M29661,~~M29662|st||md|1S36QH8JAD56KC5JKA%2CS2JAH479D3TAC678T%2CS578TH5QKD248QC9Q%2C|rh||ah|Board%203|sv|e|mb|1C|an|Minor%20suit%20opening%20--%203%2B%20C%3B%2011-21%20HCP%3B%201|mb|p|mb|1S|an|One%20over%20one%20--%204%2B%20S%3B%206%2B%20total%20points%20|mb|p|mb|2N|an|Jump%20in%20notrump%20--%203-5%20C%3B%202-5%20D%3B%202-4%20H%3B%20|mb|p|mb|3N|an|4%2B%20S%3B%207-11%20total%20points%20|mb|p|mb|p|mb|p|pc|H4|pc|HK|pc|H2|pc|H8|pc|D2|pc|D9|pc|DK|pc|DA|pc|DT|pc|D4|pc|D7|pc|D5|pc|H9|pc|H5|pc|HT|pc|HA|pc|D6|pc|D3|pc|D8|pc|DJ|pc|S4|pc|S3|pc|SJ|pc|S5|pc|SA|pc|S7|pc|S9|pc|S6|pc|S2|pc|S8|pc|SK|pc|SQ|pc|C4|pc|C5|pc|C8|pc|CQ|pc|DQ|pc|H6|pc|HJ|pc|C6|pc|C9|pc|C2|pc|CJ|pc|CT|pc|CK|pc|C7|pc|HQ|pc|C3|pc|CA|pc|H7|pc|ST|pc|H3| I have 7 off the top, can set up 1 diamond, but if I lose a diamond to East, 2 diamonds and 3 spades through my queen might set me before I can get a 2nd diamond or a spade trick. My plan is to keep West on lead. West helpfully takes his ace when I lead a diamond to the king, and pushes another diamond right back. I'm not sure it was right for me to duck this trick - but when I DO duck this and East doesn't overtake and switch to a spade, I figure, hey, obviously East doesn't have the DJ or he would have jumped on it to beat me. Sooo... I get back in, finesse the 8 of diamonds feeling confident of my contract... and it turns out they were 3-3 all along and East had the jack. Pow, down two.

When you admit to 2 keys plus the queen and the ♣K, partner pushes to a less-than-laydown contract, but one with lots of chances: spade hook, spade drop, finding the club queen, or various squeeze possibilities. [hv=pc=n&s=sakj3hq3dakt9ckjt&n=st5hak8764d852ca4&d=s&v=b&b=7&a=2np4dp4hp4np5sp5np6cp7hppp]266|200[/hv] West opens the ♦Q. You win, and (your first good news) the trumps break 2W-3E. Question #1: What's your plan from here? There may be a followup, if a position that came up at my table comes up at yours.

In Alaska we tried building a new capital city from scratch in the 1970s. Discovered it was very expensive, and gave up before we got very far. The owner of one of our bridge clubs lost so much money in the ensuing land speculation fiasco that he was working into his 80s rather than retiring.

That doesn't necessarily mean anything is wrong with the system. There is no such thing as a hand type that is very likely to take exactly 8 tricks. The only time you ever play in 2NT is when you attempted to get to 3NT, failed, and are in serious danger of already being too high. With ~24 combined HCP and two semibalanced hands, you'll get 7 or fewer tricks about 30% of the time, 8 tricks about 40% of the time, and 9 tricks about 30% of the time. With 25 HCP, you would be making 3NT often enough that you should bid it. By the time you are down to 22 combined HCP, the chance of making even 8 tricks drops off fast. Indeed, I would say that if 2NT made more than about 40% of the time, it would be a sign of serious underbidding; the best of the hands you are stopping in 2NT are almost surely good enough to attempt 3NT. On the other hand, if you were making only 20 or 30% of the time, it might be a sign you were getting too high too often.

Speaking as a Westerner... anywhere on the east coast is fine with me. Keep the hot-air pollution downwind of the rest of the continent, thank you very much.

Pro-ams, and all other formats that restrict what kind of pairs may enter the event (mixed pairs, etc), are hit with at least a 20% reduction in masterpoint awards because of those restrictions. I am actually a little surprised that the ACBL didn't apply a similar reduction to the robot tourneys vs. the all-human tourneys. (Presumably that means they buy the argument about putting everyone on an even playing field, combined with the fact that, as in an individual, there is no restriction on who may enter.)

I was thinking "South."

I played negative slam doubles with one regular partner for many years, and had good success with it. I don't recall any penalties collected but do recall several cooperative sacrifices found that would have been nailbiter decisions or quiet 1430s without it. (Also at least once where we chose between 1660 and 1700 but that was the fault of a subminimum preempt not the double.) It is true you have to guess some what to do with Kx/Qxx. But you have to guess whether those are tricks when you sacrifice without information from the double too. I have been a big believer in them ever since I first read the 2-page description of them in Kearse, and was quite surprised to discover they were so uncommonly used. Back in the mid-90s I had the misapprehension that they were in very wide use by advanced players and it was just a choice between negative and positive. But I've encountered widespread fear of them, much as described in this thread, when I have proposed them to several semi-serious partners the past couple years.

Link to hand I don't much care for GIB thinking that a 6-2-3-2 1-count is a penalty double after its partner shows a minor 2-suiter, and I don't much care for it selecting a spade lead against 5Hx either. Looking at the explanations, the root of the problem appears to be its inability to believe that sometimes partner is bidding on distribution, or not bidding to make at all: it took my 2NT-followed-by-5D as "29-30 points"?! The same hand caught half the field in a robot tourney today on similar auctions. I really should have learned from this accident, but no, half an hour later it did the same thing to me after the other side had a strong Michaels auction: Second hand In both these auctions, humans grasp the idea that the side that makes the strong cuebid is bidding on strength / to make, and the side that just bids suits is bidding on shape / to sacrifice, and even have forcing-pass rules accordingly. I don't expect the bot to know that much but I do think it badly misevaluates its defensive prospects when partner makes a shape bid.

I plan to force, starting with 2S, but have a feeling it'll be to 4S when it comes back to me no matter what I do. As much as I love fitjumps, neither the trump support nor the diamond quality nor the club length quite measures up.

Yes, the xstring package was a godsend. I tried to do this with only ifthen some years ago, ran into all sorts of errors I couldn't understand, considered settling for \hand{spades}{hearts}{diamonds}{clubs} but hated the four sets of brackets, beat my head against trying to learn enough plain TeX to handle strings myself... I havent really given much thought one way or the other to a CTAN distribution. It seems like quite a small market, and my LaTeX is somewhat less than professional... I ought to tidy up quite a few things in the code before I published it widely, not to mention finishing a couple things I never did get done (a 2-hand auction, and a nice facility for labelling example hands, so that I could have "Opener", 'Responder A", "Responder B", "Responder C" side by side, for a bidding lesson handout for instance.) Partly I was waiting to see what variations on my initial product I had a need for - so I could, for instance, make my most frequently-used format of auction "\auction" and my second-most-frequently used "\auction*". I didnt really try to build an environment to automatically typeset alerts under or beside auctions mostly because I havent decided what I would want one to look like if I did. I do feel, philosophically, that things like a box around the outside of the auction or hand diagram shouldn't be part of the base command, but either a new command or just use \fbox around the base command. But everyone has their own idea what they want.

"Never" is an excellent robotic rule, re bidding Stayman on 4333, and almost as good re bidding Puppet on 3334.

Nice to know I am not the only one using LaTeX for bridge. I also have developed a package, mostly for my own use. You are welcome to examine it and borrow anything from it you find useful / mix and match between our approaches: http://taigabridge.net/latex/grbbridge.sty http://taigabridge.net/latex/grbbridge_features.pdf

A subtle difference in wording I would like to point out: The ability to play in opener's minor at the 2-level is, yes, a fairly minor concern. The ability to make a natural nonforcing bid] in 2 of a minor has considerable value - most significantly when opener has 3-card support and corrects back to 2 of responder's major, finding a bunch of 5-3s and playable 4-3s even when responder is too weak to use any of the inv+ asking methods under discussion. The more unwilling to raise immediately on 3-card support you are, the more valuable the weak 2m bid becomes.

Leaving aside the contents of Marvin's article... a lot of good things can happen with Checkback. The one downside is losing the ability to sign off at 2C; the upsides of saving an extra step of bidding after 1c-1M-1N can be several. Being able to rebid 2M over opener's 2D instead of 3M over his 2N is probably the biggest of them. I would wholeheartedly agree that 'always 2C' is better than New Minor. All the Polish Club folk have agreed with that idea for years. Playing either convention, it is difficult to handle ALL the forcing bids with one convention. Far superior to the "usual" way, IMO, is to keep responder's second round jumps forcing, and put all of the invitations (as well as a few of his game forces when he is unsure of strain) through 2C. Playing that way I feel no need for two-way checkback, and rather dislike XYZ, although part of that is because of it having a "responder shows" feel instead of a "responder inquires, opener shows, responder sets" feel to it.

Ive never held more than 27 myself. I directed a game in 2003 where someone had a balanced 31 count -- I don't recall how many winners -- and responder had enough that 2210 was a bottom and 2220 was a top.

Expecting, no, but with a human partner, it would certainly be possible, if he didnt hold a good long suit. A lot of things about the hand didn't make sense, and there are other ways either of us could have bid. But my question still remains about whether the cuebidding style has changed / whether GIB now has a tendency to psychic cuebids.

Whether a double last round would be heart support or show values in diamonds, I wish I had done it, to help partner make an informed decision this round. Now, I let it go.

GIB says it does: "8+ 8421 points in clubs," not 4+. This is the first time I have seen it lie about 8421 points. It tends to be very reliable (7- if it bypasses a suit, 3- if it passes it twice, 6+ in trumps after replying positively to a queen-ask.) If GIB switched to Italian cuebids, it just did, and the explanations have not been updated to match. (I would be surprised if it had, to be honest, since they are a lot harder to program well.)

With spades agreed as trump, GIB cuebids 4C (promising the ace) holding 876xxx x Kx KQxx. At my next bid, I used RKC, got a 0-or-3 reply, and knew with absolute certainty it had be three because of the cuebid. Down two in a grand. Had GIB done the asking -- the normal thing to do with second-round control of the side suits but no ace, when you have agreed to do aces-first cuebidding -- we would stop safely in 5. Linky I have never seen GIB lie about the number of keycards it holds, or give the wrong reply to Stayman, as a result of a simulation; it always replies to Blackwood with the correct response. I wish it was equally unwilling to lie with other bids that carry very specific meanings, promising or denying the possession of a particular card. Can it be discouraged from creative cuebidding?

I understand liking to play the hand, and I understand going out of your way to play in the major at matchpoints. But I don't understand GIB rebidding an 87xxx suit rather than showing 4-card support for my minor at his second turn: I am 4-1-3-5, he is 1-5-3-4 and weak. Would anyone NOT rebid 2C? (And to think I rebid 1S rather than being tempted to hog the hand myself with 1NT, partly out of fear of a heart rebid.) Linky

I was under the misapprehension they already weren't supposed to sanction games that aren't run in accordance with the Laws, and didn't think a new motion was needed to remind them of it. (In particular, I believed that a club policy to ban all psyching was grounds for losing their sanction, but apparently that hasn't been enforced for a long time either.) Admittedly I have requested and received permission to award masterpoints for an odd game before -- an auction bridge game run in accordance with the 1920 Laws, as part of a pioneer-days type community festival -- and I would be sorry of the successor to the American Auction Bridge League was no longer able to carry on that organization's work. But I have a feeling that isn't what the legislators were taking aim at. I am not 100% sure I would agree that the robot games are not in accordance with the laws, under the provision for non-random deals that is used for par contests. In a sense that is what robot games are, with all the contestants having the other three hands played for them in the "same" way.

From the OP, I got the impression that, if you alerted and were planning to tell your opponent to look at the card, you expected the agreement to appear there. (Perhaps the agreement was "OK, everything on this cc looks good" and you couldn't remember what was there. I think that constitutes having an agreement, if one appears on the card, even if you can't remember it or forgot to even look at one line of it.) If in fact you had not made an agreement of the everything-on-this-card or everything-on-the-yellow-card or everything-we-play-with-partner-X nature, and had no agreement about the call, a "no agreement, but these are possibly-relevant other facts" type of speech is appropriate, and I can understand your side of it better. I remain sympathetic to your opponent's desire to stop your speech, and think that, when an opponent does so, your obligation is to stop it at once (if you feel you have an obligation to share something more, excuse yourself and say you need to speak to the director, perhaps.) The fine points of the law are debatable - it's the old "giving UI isn't an irregularity, only using it is" argument. And no, I'm not calling you a cheat. But as you know, in "could have known"-type cases, we give people the same score we would give a cheater without calling them cheats. And they hate it. Stepping back from the particulars of your case, if there is an agreement, I do feel that a fuzzy start of an explanation that causes one's opponent to cut one off in fear of a "taking it as" is a clear situation where one could have known that one's manner of explaining might mislead and damage the opponent.