zhasbeen

"This may be a bug..." This is the extreme version, but it has many cousins that have 2 elements: 1) rebidding a 5-card major when better bids are available. As I said initially, I have relaxed my stance on rebidding 5 of a major. There are times when it is accceptable, such as some 5-3-3-2 hands with a decent major, etc. However, it makes no sense to me to rebid 2H with xx AJxxx AQJx xx after a 2C response. Rebid 2D, showing hearts and diamonds. What could be simpler and make it easier for partner to know what you have?. 2) Taking the 2-card preference for 5-2 major fit, when 8, or 9 even 9-card fits are available in a minor. This happens fairly often, even when the 2nd suit has been mentioned. That said, it is probably not as easy for the programmers to fix as it might look. They may have to start over from scratch. It's not like a Windows patch that Microsoft can do on the fly.

Time, and time again, GIB rebids a 5-card major before mentioning a 2nd suit, even when the 2nd suit is 5-cards or more. I can cope with it, so far as competing in robot tournaments, since it’s the same for everyone. However, I would like to see it acknowledged that this isn’t just a minor problem, and I don’t mean on just this hand. I’ve mentioned this same GIB logic on several occasions before. Since then I’ve given a little on my opposition to rebidding 5-card major. I’m still not crazy about it with 5-3-3-2, but I have no great objection if it’s in the framework of system. Even with a strong 5-card major and weak 4-card minor I’m ok with it. However, I don’t think I’ll ever accept with 5 of a major + good 4-card 2nd suit, and never ever with 5-5 and up. The definition does say “Opener reverse; 5+ clubs, 4+ hearts, 3-spades, 3-diamonds”, but we shouldn’t even need a definition for this one. GIB chose to rebid spades rather than support hearts with QJ10xx! Bug report submitted. Btw, it does make 5 in either suit without diamond lead and continuation, 4 without; but in the great majority of cases a 9-card fit will work better than 7-card fit, especially when you need to take ruffs in either hand. A picture is worth a thousand words. [hv=pc=n&s=skqha765dj7cak652&w=sj9hk83dkt4cj9874&n=sat873hqjt42d62ct&e=s6542h9daq9853cq3&d=s&v=0&b=11&a=1cp1sp2hp2sp3sp4sppp]399|300[/hv]

I have developed more tolerance for GIB the past week or so. More experience and reading through some stuff, such as fstrick604 thread on this forum has helped. I already knew that creating a software program that can play a decent game of bridge is extremely difficult. This one from smerriman post really caught my attention: "The catch is that the best bidding system for a robot is completely incomprehensible for a human, and as stated near the end of that PDF, trying to make it human-understandable usually gives a much worse result" I downloaded the PDF and read-skimmed through it. I don't know if you've watched basketball much, but most fans have seen times when the coach yells "oh no, oh no, NO!" as one of his players launches an impossible 30-footer. Then, as the ball swishes through the net he screams "yes, yes, YES!" as jumps up off the bench. Well, GIB has given me many of those moments. Sometime there is a method to his madness, and you gotta take the bitter with the sweet. Now, if I can just make it through today without breaking any furniture...haven't played a robot tournament yet.

“You should post hands in separate threads” Will do--I wasn’t sure if what I had in mind was feasible, or if it is appropriate to talk about hands that aren’t necessarily focusing on GIB problems. There is probably another BBO forum for that. One question for you—is there anything you can do except pass when LHO opens 15-17 NT and you also hold one as next bidder? I’m talking about a balanced NT with no 5-card suit. I’ve been passing with those, but usually with some sense of frustration.

I dunno about that one. Two diamonds is a reverse, and I would think a jump to 3 would show more distribution. In hindsight, however, I don't like my 3H cue bid. North already had a chance for NT and didn't take it. Four clubs is another bid I could have made, and 5C happens to make. Bot might accepted with an ace and 3 trump, plus the J in my diamond suit.

7446-5 How do you bid this hand? I finally passed after laboring for a while, playing bot for longer spades but too weak to bid over 1H. Several players opened 2NT with my hand I did manage to make 3S with this 4-2 fit, but got 6% with more players than not in NT. However, my 3H was probably overdoing it, since I already advertised strong hand. Bot could have made neg double over 1H, but I can't be overly critical with 2 of 6 hcp being jacks, and one of them in opps suit. As it turns out that jack was big for no-trumpers. I wonder if opening the hand in question 2NT will work out best in the long run. I will sometimes open NT with this shape, but want stoppers in both doubletons to do so. [hv=pc=n&s=skqh73dak98cakq97&w=st72hakqt4dq53c82&n=sa654hj865dj2ct54&e=sj983h92dt764cj63&d=n&v=0&b=5&a=pp1c1hpp2dp3cp3hp3sppp]399|300[/hv]

This is a hand type that comes up reasonably often and always gives me fits. The 2 main components are 6-4 pattern with too many points for a jump rebid. For this hand the jump shift rebid in clubs was easy, but can cause all kinds of problems after north’s 2nd response. It will be hard to ever recover the heart suit, and what what happens if he raises my clubs? Should I pass 3NT or “correct” to 4H now, figuring he will know I have 6? But, what if he has stiff heart or void? What alternatives do I have to club jump shift rebid if that happens? I need a set way to handle this type of hand when it comes up rather than go into tank for a half hour. It has many cousins too, such as 6-3-3-1 pattern and 19+ hcp. [hv=pc=n&s=sajhq98743dacakq2&w=skt862h5dktcj9753&n=sq7543hkjdq83c864&e=s9hat62dj976542ct&d=e&v=0&b=14&a=p(Bot)1h(Me)p(Bot)1s(Bot)p(Bot)3c(Me)p(Bot)3n(Bot)p(Bot)p(Me)p(Bot)]399|300[/hv]

All the hands I am submitting are from ACBL 12-Board Robot Tournaments. In all cases the human is declarer, sitting south in diagram. The human is never dummy in this type of tournament, although positions could have been rotated if GIB was first to name final contract. They are not necessarily criticism of GIB, or the human player, but there could be errors on one or both sides. None of them had easy answers for me, or I probably wouldn’t be submitting them. The first question will always be “how would you bid this hand?” It can then be answered with the logic behind your answer and any other comments or criticisms you want to ad. work in progress…more hands to come

Man, are you ever right about that. It's the reopening double that I've been burned on a few times. I'm very careful about those, but GIB seems to be o.k. with the michaels. I'm 1 for 2 the last two times I took a chance and reopening--one top and one bottom. I'm thinking that it might be because michaels defense is part of system, while reopening doubles are judgement calls. As you said, "who knows?"

A timely post for me as I'm trying to learn michaels and Unusual NT defenses. I'm trying to memorize them before a game I have next week that I'm a little nervous about. This is one of the easiest, since double of michaels or unusual NT cue bids are always for penalty. You don't have to stop and think if it is upper, lower, 4th suit etc. This one and direct raise of suit partner opened are easiest. Anyway, I think it's a must today, since there is more interference bidding than ever. I wouldn't have known what to do month ago and am still far from confident that I've got it all.

Thanks diana. Can I delete a post or does admin have to do it? I know about the preview. I've created handviewer diagrams manually and pasted tinyurl links into my posts, but was trying to combine the two with no success so far. I've seen some some of the explanations but has yet to sink in. Will come back to this later and play around some more.

TESTIING -trying to insert tinyurl win handviewer Is it possible to delete a post? I want to delete this one and start over

Is this the way you play it? The robot mouseover I'm looking at says "4+ spades". I'm more comfortable with Bergen-Cohen style, but I'm learning that there are lots of good players who play it differently. Speaking of the mouseover help, fixing it to a level where we could always trust what it says makes the top 2 of my most important upgrade list. That ("always") would be ideal although it is probably impossible considering that it has to be in sync with all possible bids that it might encounter. Even though I'm not comfortable with robot bidding style I think I could adjust to it if I could count on those explanations.

Dang, that was so easy. Short and simple play, as well as explanation. Of all the noise going on in this old brain I only needed to ask myself 2 questions. Does east have 2 or less diamonds, and what can he do to hurt me if I play spades first and continue until I have 3 more tricks in the bag. Entry worries and tricks needed solved. Geez...

I think you can make it, but not quite as easy as it looks at first glance. If you duck first D and win next in hand you won’t be able to take C finesse without using ace of D. So, you lead a club to dummy and continue if west ducks. Let’s say he ducks, wins 2nd club and takes out your last diamond. If west wins first club and takes out ace of D, you can get back to your hand by overtaking club without losing an extra trick in that suit because east has 9xx. However, it could make you kinda nervous at the table. I know I would be So it looks like you are ok now, but what happens if east refuses to take king on first spade finesse? You have 3 clubs, 2d, 1h and 2 spades and are still a trick shy with no entries back to dummy after you repeat spade finesse. I keep trying these scenarios and don’t see enough tricks with only 1 heart and possibility of EW ducking first round in black suits. However, I noticed early in the play that you can make 3 heart tricks if your entry situation will allow it by covering anything that east plays when you lead from dummy to south hand. ………………………………………………………………………………………………… Now it’s time to see if entries and timing will allow it before opps can take 5 tricks. It’s looking good so far if I’m seeing this right. You can make it with only 1 club trick: 1C, 2D, 3H, and 3S. Hold of on first D and win continuation in dummy. Next, lead heart to your hand, planning to cover anything that east plays. West wins heart and takes out your last diamond. If east puts up Q first you just win it and lead a heart towards dummy’s 10. No matter what they do you now have enough entries to set up 3rd spade if east holds off on first one, but lead small to dummy rather than running 10. You'll need to finesse east in heart suit again before repeating spade finesse if west won first heart finesse with K. I’m not so sure I’d make this hand at table, but one thing I’ve learned is that robot can be very good at messing with your entries. That’s why the entries are the first place I look for possible robot traps. Although the robots make mistakes, they are very capable of making your life difficult, and are sometimes brilliant. I find most robot tournament hands to be challenging, even when they look easy at first. Double-check me on this. I could have missed something.

Stephen, for what it’s worth, when I see you made a post marked “today” I always look at it. I’m curious about what you have to say. I laughed when I read some of this response. I mean a real laugh rather than a snicker. I’m argumentative by nature and it I looks like I got a rise out of you with the “weird” comment. There, I just laughed again. You never know for sure how a person is coming across when you just read stuff rather than talk. I don’t think either of us is quite getting our point across to the other. For one thing, I’ve rebid 5-card majors a couple times just in the past few days. I suggested the “promise 6” rule because I thought it would be easier to program, and superior to frequently rebidding weakish 5-card majors in those situations. The 5-card suits I rebid were good ones, however, and I treated them as 6 when I could find no better bid. “I think you erred in not bidding 3nt with extra values, that will be hard to show later,” I have a copy of the GIB 2-1 card at my side when I play and I also look at the cursor messages before making some bids. It says that 3NT promises 13-15 with trump support. I would have liked to bid 3NT as you say. This is one change I’d be highly in favor of, since I see little need for the 13-15 bid. This isn’t the first time I’d wanted to make the stronger 3NT bid on hands with 2 trump rather than 3. So, you are right that it would be hard to show later. I labored for a minute or more before finally settling on hearts. Zeroing in on where I have the most (but not only) problem with robot style is the 5-2 major fit. Not only do bots like to rebid 5-baggers, but they love to take a preference for major with only 2-card support. This happens even when I’ve shown a second suit that I’ve bid and bot has support for it. I’m not just talking about the forcing NT situation where opener promises only 3 for minor rebid. Time and again the bot will pass up a fit of 8 cards or more in favor the 2 card preference for the major. This drives me crazy. I know that major suits score more, but minor suit plusses give you more match points than going down in a major. With a normal 4-2 break and only one ruff in your hand you can have your work cut out trying to make the hand. I see 5-2 fits in my sleep. They come in all forms too, not just after opening a 5-card major. This seems so against the grain of law of total tricks model that is supposed to be integrated in GIB program. However, I will give GIB credit for being good in many competitive situations by being right about when to push to 3- level and when to let them have it for 2. It seems like the 5-2 major thing is a more of a blind spot. I’m sure you have counters potential problems that I worry about, but I am assuming they would be more complicated than what most players are able, or willing to try and absorb, including me. I assumed that something similar to what I play is what most people play nowadays. I’m not talking about the pros that have pages of notes to cover all the ins and outs of their systems, but maybe I've made a false assumption all the way around.

This one, although not a serious error for a single board, is a serious problem with bot bidding over time because it comes up so often (my opinion). It makes my life difficult every time “he” does it, although I’m not caught off guard now like I used to be. Once again, the robot makes rebidding a moth-eaten 5-card major as his highest priority. Allowing for that possibility I rebid 2NT for my 2nd response. Next, he does “IT”, showing his spade suit. I’ve seen him do this with 6-4, 5-5, 7-4, etc. You never know what to expect. With 5 of a major and 5 in a minor, he rebids the major first before showing the minor. Stephen, I can’t believe you would bid this way! I just saw your name in tournament results from last week. I know those names, so you have to be a strong player to keep that kind of company. If you do really bid like that you must be one of those card playing wizards that can get away with weird bidding because you declare and defend so well. I won't drop any names regarding tournaments, etc unless you say it's o.k. Btw, I used to go to 24-hr fitness in Milpitas until switching to City Sports that is only a couple blocks from where I live. My choices for bot’s first rebid would be 2 spades or 2NT in that order. Reverses do not promise extras playing 2/1. Considering that his stiff is the 10 in suit that I started 2/1 auction with, 2NT seems reasonable. Rebidding 2 hearts first would be my last choice. Here’s the hand. Although not a total disaster, it came on last board of a 43-table match in which I had 78% score after 11 rounds. The 39% I got knocked me down to 74.5 and 2nd place. I still thought I’d won but knew it wasn’t a sure thing. Sure enough, some other guy had a 75.8%. Don't know how to change opening lead to East, but not much to play after bot lead stiff Kd and I played 3 rounds of trump, making 4. Eighteen other pairs played in NT making 5 and 6 for 78.4% and 98.7%. I guess I could have still played in NT, with QJx of diamonds being a sure stopper. However, I was thinking more in terms of over tricks if bot had stiff, which seemed to be a good possibility off the bidding. I was hoping this was one of those hands where he did have good hearts--you never know. Maybe a case of over thinking, since I figured a lot of pairs would be in NT. :[hv=pc=n&s=saj2hk8dqj5ckqj73&w=s843ht4dt97432c64&n=skqt9ha9653da86ct&e=s765hqj72dkca9852&d=n&v=b&b=12&a=1hp2cp2hp2n3sp4hppp]399|300[/hv]

True; and I actually did get a spade lead. When the dummy came down I figured that there was less than a 10% chance that bot had lead away from the king. The great majority of the time they make a passive lead. However, there was no other possibility for a 12th trick so I let it ride. Down one.

You don’t know half of it. I was leading a 25-table field after 11 rounds and got knocked down 2nd after 26% on this board to lose by 1.4%. That wasn’t the worst, however. On Saturday I went into last board of 43-table game with a 78% score. I’m in process of inserting a handview file of this board in that thread. Anyway, I got a 39% on the board that knocked me down to 74.5 % and 2nd place. I was confident that I’d won anyway but knew that it wasn’t a sure thing. Sure enough, someone had a 75.8%, giving me another tough beat on last hand. The 78% would have been a new record for me since I started playing in the robot games 10 weeks ago. I always play in the 12-board ACBL games. I had 2 other close 2nd’s the last couple days, but will spare you the details. In one I didn’t play well at all and deserved to lose. All that complaining aside, I’ve gotten plenty of breaks from robot, and some good plays too. I believe that the breaks even out over time. And, although we would all like to see improved from robots, I doubt that it would change our scores much, if it all. We all play with the same robots—it’s the same for everyone, as the saying goes. The big plus would be in form of fewer frustrations. It’s always more enjoyable to play with good partners, and the robots never criticize us. I can't begin to imagine how difficult it is to program these guys. What I can imagine is that there is a domino effect. When you make one change it has an effect on a bunch others that you have to deal with.

I hear you, but in both cases the only thing robot knows about the south hand is that it has made a quantitative raise. With that knowledge the robot has deemed the square 12 on top more worthy of going to 6 than the bottom hand with 12 points and a decent 5-card suit. I think my decision to make the 4NT bid is more open to question. It was on conservative side but I see the robot open quite a few 11-12 pointers and so do I. I hate turning a plus into a minus by making a risky bid at match points where games and part scores can be worth more than slams. The quantitative raise is asking "are you in the bottom or top half of minimum range?" That top hand looks like bottom of range to me.

These were 2 hands I played recently: Deal 4 (top): As you can see, robot had a square 12-pointer and accepted quantitative invitation by bidding 6. I went down 1 when queen of spades was off side. There is no other play for 12th trick. Deal 12 (bottom): This one started almost exactly the same as #4, but this time robot passed with what was arguably a better hand because of diamond suit. This time robot passed and I made 6 that required only a 3-2 diamond split. working on handviewer diagram....top is deal 4, bottom deal 12. Both say deal 1 now but Idon't know how to change them now. when I clicked "D" it spit out random numbers rather than 1-12. I it possible to edit hv diagram after you have inserted it? [hv=pc=n&s=sa73hatdkqj2ckq82&w=skjt84hj964d74c93&n=sq65hkq3da63cj765&e=s92h8752dt985cat4&d=n&v=0&b=4&a=1cp1dp1np4np6nppp]399|300[/hv][hv=pc=n&s=sakt2hakqdq84cq76&w=sjh9854d952c85432&n=s83hjt7dak763ckj9&e=sq97654h632djtcat&d=n&v=0&b=12&a=1dp1sp1np4nppp]399|300[/hv]

[hv=pc=n&s=sakt2hakqdq84cq76&w=sjh9854d952c85432&n=s83hjt7dak763ckj9&e=sq97654h632djtcat&d=n&v=0&b=1&a=1dp1sp1np4nppp]399|300|made 6[/hv]

“you only looked at the first post of that thread and concluded that the first post was the only one you needed to read, and there was no change from then” You are mostly correct, although I did read several of the other posts. If there were comments about updates I either missed them or forgot. I saw 2011 last upgrade on the first post, and combined it with general mood of the forum to make my assumption. And now that I think of it, another thing that has had a major impact on my pessimism towards effort by programmers is the bidding help, or whatever you call it when you run your cursor over it. The complaints have been coming for a long time and it is of critical importance (my opinion) to fix. I know it must be very complicated to sync the help with all the possible bids. I would now rate it as #1 over 5-card major overhaul as most important upgrade focus. I want to believe that somebody is listening and that fixes are being worked on. My tolerance for problems goes way up when I know people are trying. So, its good know that there are actually people working on the problems that are brought up. “That's your opinion. I think most good players will disagree and think double is better.” I called someone that I believe even you would classify as a good player. His response was somewhere between your opinion and mine, but more towards yours. He would overcall 2 clubs with the hand in question but had no great objection to a double. His plan would be to make the overcall, and then reopen with a double if bidding died at 2 hearts. He included preemptive value, and hand not being quite good enough to double then rebid clubs, in his reasoning for choice. He went on to say that with the same distribution and less concentration of hcp in the club suit he would favor an initial takeout. So, not as much concern about holding only 3 spades as I expressed, but still a concern Major suits are a much broader topic that was trying to narrow down. One thing I’ve learned as I’ve gained more experience playing robot tournaments is that you can’t give up when every hand feels like a disaster. Everyone has to deal with the same wild distributions, double-dummy defenses, and bidding quirks. If you keep your cool and think it through you can still score well.