wbartley

This was in IMPs play with human sitting South. [hv=pc=n&s=sk84hj765d753cj42&w=st763hdqjt86cq865&n=shakqt982dak94ct9&e=saqj952h43d2cak73&d=n&v=0&b=1&a=1h1s2h2s4hppp&p=d2d3dtd4d8dah4]399|300[/hv] This has to be some sort of bug, right? Unless there is some way I can't see how ducking the diamond at trick 1 can benefit declarer. It looks like you're losing 2 clubs and a diamond and there's nothing that can be done about it. So, why risk a ruff (or two) by ducking? Funny.

A cue bid is the smallest lie, with pass being at the top of the list of worst lies.

GIB already passes when it doesn't have a definition for a bid. So, if there is a rule that says jumps to the 4 level are splinter bids, change it to say, jumps to the 4 level are splinter bids unless the bidder has previously bid the suit naturally. You don't have to define 4♦. Just don't misdefine it.

Basic robots. Human South. [hv=pc=n&s=sj62h93dajcajt862&w=saq9543haj82dt7c4&n=st87hk64dkq96ckq5&e=skhqt75d85432c973&d=s&v=0&b=11&a=1c1sppp]399|300[/hv] This is not a trapping situation with the North hand so GIB shouldn't leave it to South to reopen here. KQx should be good enough club support for a cue bid.

Just play bridge robots. Human South. [hv=pc=n&s=sakqjh762dak972c9&w=s76432hk98d65ckq2&n=s8hajt53dq843cjt7&e=st95hq4djtca86543&d=s&v=e&b=3&a=1dp1hp1sp3dp3hppp]399|300[/hv] 3♥ non-forcing in this sequence is unplayable. South has shaped out and North has a choice of games (or possibly 4♦ not-forcing).

Yes. Really. The double is far more useful as business. Partner can pull it with a poor hand and a biddable suit. At total points scoring methods, there's no reason to compete in the balancing seat with a hand that couldn't make an initial takeout double of 1♣. Just defend. Look at it this way. If your side can make 2 of a major, what are the chances that 1NT is making on the lead of partner's better major suit?

Just play bridge robots. Human South. [hv=pc=n&s=sa986haj5dkcqjt76&w=sqt4ht4d8754ca542&n=s753h8732daq632c8&e=skj2hkq96djt9ck93&d=n&v=0&b=1&a=p1cp1nppdp2hppp]399|300[/hv] First of all, this shouldn't be a takeout double. This double implies that South has clubs and was trapping over the 1♣ opener. Secondly, the description of the double says, 3+♦; 3+♥; 3+♠. So, why is GIB bidding 2H on xxxx when it has AQxxx of diamonds? Title of post is wrong. Can't believe I did that. It's supposed to be, "Double of 1NT shouldn't be for takeout."

I see it's a text format and I suppose, given sufficient time, examples and energy, I could reverse engineer it. But, a spec would be nice.

Thank you for the web page. I have two questions. 1. Did you work from some sort of specification or did you reverse engineer LIN format from examples? 2. If you don't have a .LIN specification, would you be willing to publish the server side code that does the conversion so that I can streamline the creation of pbn output and conversion to lin format in a different program?

Thank you. Where is the specification for LIN format?

Is there a way to import hands for use at teaching/bidding tables? Any import format would be okay with me.

The only South hand shape where correcting doesn't work out as well as passing is 3316 and even then you end up playing the 43 heart fit which is okay.

Just play bridge robots. Human South. [hv=pc=n&s=sa73haqt72dj5c642&w=skqj92h94dt3caj85&n=s86hkj85d87642ckq&e=st54h63dakq9ct973&d=w&v=b&b=4&a=1sp2spp3dppp]399|300[/hv] I realize this is a balancing auction and I've already been told that GIB balancing needs work, but this is a fairly common situation where North is much better served by doubling in the balancing seat and correcting a 3♣ response to 3♦ rather than unilaterally bidding 3♦.

The principle of restricted choice is sort of confusingly named. If you assume that a player holding equivalent cards (who knows that the cards are equivalent) will play one of the two with equal probability, then it comes down to the fact that there are two possible hands that contain exactly one of the two cards and only one containing exactly both, thus making the case of a singleton twice as likely. I would have thought that GIB would have always gotten this right without being aware of the human interpretation.

Okay, make your 2 of clubs the 2 of hearts and your two of spades the two of diamonds. Now you have a six loser hand that only rates a total tricks raise to 3♥ and you have the same play for game as the 64 hand. You're leaving no wiggle room. A single raise is such a wide range of possible hands and the North hand with the club ace, concentration of spade values and four trumps is a super max after LHO raises to 3♣. Look at it this way. If you had the North hand and you knew LHO was going to raise to 3♣ before you bid, wouldn't you be tempted to bid 3♣?

Your 3♥ bid is awful but that doesn't matter. North has no business bidding 4. The explanation of 3♥ clearly indicates no game interest. Either the explanation is wrong or GIB is off its rocker. On second thought. If this is advanced GIB, it may be running some simulations given the auction. With 3♥ showing six way more often than five and the opponents holding 9 clubs more often than 8, it probably figures 4♥ will have play. So, this one may be your fault. Make your three of clubs the three of hearts and 4♥ has good play.

It rebids diamonds instead of its four card heart suit when you could still have four hearts but bids hearts when you've denied holding four.

Agreed. This is the only sensible treatment with natural bidding methods. Phil Read, a local San Diego expert, used to say, "If you get to the wrong contract, it's because you didn't support your partner's suit." In this case, rebidding spades with 3 hearts makes the subsequent auction impossible.

To be fair, the request wasn't to force the bot to lead his/her suit. It was to force the bot to lead partner's suit unless there is a strong reason to believe another suit is better. The real question is, what constitutes a strong reason to believe another suit is better? GIB generates a number of hands that match the bidding. It then runs double dummy analysis to decide which card is the most effective. Since the generated auction is what GIB would bid and GIB doesn't concern itself much with suit quality when overcalling, that might be skewing the results toward not leading partner's suit. I wonder if GIB leads partner's suit when choosing from what it believes to be equally effective options. If it doesn't, that change alone might make GIB lead partner's suit a lot more often than it does.

Thanks for that very valuable info, Stephen! Basic GIB does indeed bid 3NT in response to a 3♠ bid. However, give it ♠7 ♥AQT3 ♦KQ9862 ♣K2, and the auction goes, 1♠=2♦=2♠=3♥=3♠=3NT when we're ice cold for 6♦. Interstingly, with ♠7 ♥KQT3 ♦KQ9862 ♣K2, basic GIB bids 4♠ instead of 3NT.

Is there some way to set up a specific layout at a practice or bidding table? I haven't been able to figure out how to do it. If I try to configure the hand generator, it times out if I specify too many specific cards. Apparently, rather than dealing the specific cards and then generating the hand with the remaining cards, the deal generator keeps generating random hands looking for one with the specific cards. The syntax is really awkward in any case so it's not a practical way to set up layouts, even if it did work. The reason I ask is that I'm curious whether GIB will bid 4♠ over a 3♠ rebid in this situation.

Priority should be given to rebidding even a five card suit over raising the second suit. There are many hand types where you're manufacturing a reverse with three card support. x Kxx AKx AQJxxx is a good example. I don't want to have to make a non-forcing 3♣ rebid with this hand so I bid 2♦, which is forcing for one round, expecting my partner to rebid hearts with five or more. If I do this with GIB, I end up in 5♦ opposite the given hand. Upon further reflection, GIB doesn't play Walsh style responses, right? With 4-4 in the red suits it will bid 1♦ in response to your 1♣ bid. So, the raise of diamonds here automatically implies five or more hearts.

Just play bridge robots. Human South. [hv=pc=n&s=sakqt53h2dat4c843&w=s986haj85dqj73ca5&n=s7hkqt3dk9862ckq2&e=sj42h9764d5cjt976&d=e&v=n&b=2&a=p1sp2dp2sp3hp4dp5dppp]399|300[/hv] The explanation of the 2♠ bid included 3-♥. GIB seems to be forcing the auction past 3NT. One might argue that South should rebid 3♠ but having already suppressed diamond support once, it seems like masterminding to suppress it again. This is an interesting followup to one of my previous posts about GIB not supporting hearts (1D=1H=1N=2S=3C=3N=4H was the auction with GIB first to bid). In the replies to that post, it was pointed out that NMF is available to ask for three card heart support and also that since GIB had denied holding 4 spades, the 2♠ bid was needlessly giving away information. But, the 2♠ bid at least has the benefit of showing five or more hearts. In this case, the 2♦ bidder has already shown five diamonds so there seems to be no point to bidding 3♥ here unless you're afraid of NT, which you aren't.

How is 3♥ the optimal contract when 2♠ is off 3? You're right. The bots certainly love to bid on this particular layout.

I'm pretty sure the explanation for 2♠ was 4+ spades; 5+ hearts, blah blah blah. If it weren't, I wouldn't have posted this. But, it's possible I made an error. If 2♠ doesn't systemically show a strong hand with at least 45 in the majors and more hearts than spades, what does it show? What is the systemic bid with 5-6-2-0? 2♣?