smerriman

Which tournament are you talking about? BBO lets each tournament director specify their own rules about the number of boards + time limit per board; some have 2 per round, some have 3 per round, etc. So you may be directing your question to the wrong place.

It's not about finding situations where playing 6 was right, it's about how many situations playing 6 wasn't wrong. By my calculations: - a 10 deal simulation will tell GIB to play low 9.2% of the time - a 20 deal simulation will tell GIB to play low 2.5% of the time - a 30 deal simulation will tell GIB to play low 0.7% of the time - a 40 deal simulation will tell GIB to play low 0.2% of the time We don't know how exactly how many deals BBO has configured GIB to simulate, but according to barmar it is "a few dozen" for advanced robots, so playing low will be rare but still within the realms of possibility, even if there weren't any bugs in the intended algorithm. (And of course, with basic bots it would be considerably less rare).

It seems clubs from the top works too :) Were the auctions and lead identical at other tables?

Oops, no idea why I wrote Fred, meant to write Matt. But you're right; it could have originated even earlier. Still seems very strange to be deliberate, regardless of who came up with the idea, compared to most other issues which are unintentional bugs. Was it a common usage back then?

It's not like this is even unintentional. The old version of the database specifically says: 11+ total points, biddable clubs or diamonds, rebiddable major -> bid 2M 11+ total points, rebiddable clubs or diamonds, biddable major -> bid 2M That's it, no other conditions, but both cases specifically listed. Why the programmer Fred Matt, or whoever (the exact same rule existed in his pre-BBO version) thought this was a good idea is beyond me..

If we take a finesse and it loses, they'll clear hearts and then we'll be down every time the ♣K is offside. I think I'd just play clubs from the top. If the opponents clear hearts we then have a guaranteed 9 tricks and free shot at 10 by taking the diamond then spade finesses into the safe hand.

Nah, it's not that at all. It's just the usual situation of GIB not making any deductions / assumptions based on past played cards. So even though there is a 0% chance you would have played the ♦J on the first round with ♦KJ, it simply doesn't know that. In fact, quite the opposite - when deciding what to play at trick 3 it thinks there's about a 65% chance you have the ♦K to make up the points that S/W have shown. (I've said a few times in the past on the forum it used to have this type of logic built in when Ginsberg originally programmed it, but it got removed for being too slow / not working properly).

GIB's forcing pass rules / when to double, pass or bid 5 are 100% broken, and basically force South to double here (and North to pass). The other options that North considers at its first bid also break in similar ways when following the book to the end of the auction, leading to its wacky limit raise. That latter part may also be caused by the fact the old version of GIB only considers 2♠ and 3♥, never a 4♥ preempt for who knows why.. so it might just pick 2♠ as the only route to game, even if it results in defending doubled once in a while..

No logic here. Old version of GIB plays high 100% of the time at both IMPs and MPs. BBO's version seems to play random cards at times for no reason.

If you swap the positions of the J and 9 so that West has JT643 left and East played 9 on a previous trick, it will never lead the T, despite being a 100% identical state from a play position. But here with T9643 and J played on a previous trick, if it decides to lead a high card, it will lead T and 9 equally as often. I guess its rules about when to lead high from equals during the middle of the hand only apply to honors.

I'm not sure why there was a long discussion about which call was the 'withdrawn' one then, since 16A1b about which information is authorized points to 16C, which says *no* information is AI to the offending side regardless. Are you therefore saying in the other situation I mentioned - where if you make an insufficient bid, and when not replacing it with a simple higher alternative, the director warns you that if you make a different bid your partner will be barred - you're not allowed to know that partner will be barred? Everything I've read says that not only are you allowed, but you're encouraged to make a 'sign-off' bid that you would never make otherwise. [Edit] This came before your edit. Where are the other laws?

As far as I'm aware Walsh style only refers to 1♣ - 1♦ auctions, and is quite a simple idea since you'll never miss a major fit. Bypassing 1♠ with a balanced hand after 1♦ - 1♥ is a different kettle of fish, since you'll miss a 4-4 fit if responder is weak - Kit Woolsey has argued if playing MPs you're better off bidding 1♠, but if IMPs then 1NT. So with AL78 playing MP I don't think it's old fashioned at all.

You're probably thinking of v3, and not v4. It's still not very good, but it's able to play complete games of chess.

You have these backwards. A chess game has a finite number of possible states, and each position is a theoretical win, draw, or loss. (And endgames with up to 7 pieces have been 'solved' via a tablebase, so you instantly know what the right move is). It's bridge which has an infinite number of states, since you have to take into account not just the current state of the cards, but also the potential reasoning that was behind every past bid and play from everyone else at the table.

Very nice hand! Goes to show the dangers of using two suited bids..

Long story short, while GIB knows that x was a support double when you made it, later in the sequence it's using rules that are only meant to apply after you made a penalty double, and its rules about what to do after a penalty double / forcing passes / etc are very messed up.

I'm aware. My point was there exist other situations where it's perfectly legal to compensate for the knowledge that partner will be constrained by UI. Why does that not apply here?

You haven't made a specific suggestion. What is your suggestion? If a 20 board match has a side receiving 3 Yarboroughs, redeal the hands? What about 2 Yarboroughs, or 3 hands with under 3 points, or maybe someone getting two 23+ point hands, or two 9 card suits, or ones where almost all of the finesses for one side lose (a much more common complaint) - where would you like to draw the line? (In any set of boards, you'll be able to find once-in-millions patterns that you'd likely never see again; it's a well-known fallacy that rare events like this shouldn't happen.) If you want hand-crafted tournaments that don't use random hands, BBO already provide this functionality; TD's can upload pre-dealt hands (like they do for Goulash). If a specific TD knew what your criteria were, and agreed it was worthwhile, then they can generate hands that meet those criteria and upload them for you. But if they decide a tournament should use random hands, then random hands is what you'll get.

I had the same hand as you a couple of weeks ago, and have just had another where it put me in slam after determining we were missing exactly 2 keycards. I think you're right that it's still playing regular Blackwood despite the alerts. [hv=n=saqj63hq64d42ckt2&w=s2ht8753d965c8643&e=s984haj9d83caqj95&v=e&d=s&sn=smerriman&nn=Argine&a=1d(4+!d; HCP 11-23; 4-card !d opener)P1S(4+!s; HCP 4+; natural - Forcing)P4C(5+!d; 4!s; HCP 15-20; Splinter, shortness in this suit and big fit in last suit bid - Game forcing)P4N(4+!S; HCP 12+; RKCB in !s {!s King is a keycard} - Forcing)P5H(5+!d, 4!s; HCP 15-20; 2 or 5 key cards and no Queen of trumps - Forcing)P6S(4+!s; HCP 12+; to play)X(HCP 17-; penalty)PPP]400|300[/hv] Despite everything in this thread so far, I'm still staggered by how bad Argine is.

If you make an insufficient bid, and replace it with a completely different bid, partner is barred from the auction. The fact that partner will be barred is AI; you're allowed to 'gamble' with a high level bid that you would never have made otherwise. From East's perspective, is the fact that partner will be constrained here due to UI, also AI? It seems like you could get into very meta territory.. is 5♦ suggested because it has a hope of game over a strong 2NT, or is 3♦ suggested because it gives partner another chance to correct their mistake with 3NT.. or can East make a bid that would suggest the opposite of what he actually wants West to bid.., thus making sure West chooses it.. etc :/

Firstly, your numbers are a *long* way off. Over 20 hands, the chance that you'll be dealt three or more Yarboroughs is actually about 1 in 5.4 million, due to the fact you have 20 opportunities to be dealt such a hand and there are many different sets of 3 amongst that. (In the same way, there's a 1 in 92 chance you'll see at least one Yarborough over a 20 board stretch - considerably more likely than 1 in 1828). Secondly, checking to see if someone has been dealt a small number of points in the past, and then intentionally biasing the deals to ensure they receive more points later.. just makes no sense whatsoever. So your partner picks up a couple of huge hands early on.. now you know they'll probably have bad hands the rest of the day to give the opponents a fair shot, and can adjust your decision making accordingly? Random dealing - which is what BBO does - so you're just as likely to receive any hand at all, no matter what happened in the past, is the only sensible option.

Yup, standard GIB; play different cards, get a different random number. No correlation between the play and heart/diamond discard, though I can see a spade affecting the probabilities.

[hv=https://www.bridgebase.com/tools/handviewer.html?lin=st||pn|smerriman,~Mwest,~Mnorth,~Meast|md|1SA3HAT8DAQJ843CKT,ST97HJ2D965CJ9875,SKQJHK965DK7CAQ42,S86542HQ743DT2C63|sv|b|rh||ah|Board%207|mb|1D|an|4+!D;%20HCP%2011-23;%204-card%20!D%20opener|mb|P|mb|1H|an|4+!H;%20HCP%204+;%20natural%20-%20Forcing|mb|P|mb|2N|an|2+!C;%204-5!D;%202-3!H;%202-4!S;%20HCP%2018-19;%20balanced|mb|P|mb|6N|an|4+!H;%20HCP%2012-16;%20to%20play|mb|P|mb|P|mb|P|pc|ST|pc|SQ|pc|S4|pc|SA|pc|D3|pc|D5|pc|DK|pc|D2|mc|13|]400|300[/hv] 18 opposite 18-19 clearly isn't worth investigating grand on. An unfortunate duplication in spades meant we only had 14 top tricks, rather than 15.

It probably should be described as help suit, rather than long suit. Though GIB is pretty bad at both making and accepting them at the right times.

The other tables would have thrown different cards, but there is zero correlation between this and the fact that they got a better score. Almost all of the time, the robot will play a third heart no matter what cards you throw. Since the free robot uses a very small number of simulations, once in a rare while it will deal hands where you just happened to have a doubleton heart in all of them (or all but one, and it includes a psych or has buggy results on another one, which counterbalances it). Its conclusion from the simulation is that a heart isn't any better than anything else, so it may lead something else. Say that happens 1 every 50 times you give GIB the hand.. if someone throws different cards, they get a different random number than you did. In this case, throwing high cards resulted in them getting the 1 in 50 shot.. next time if you got exactly the same hand, throwing normal cards might get you to the 1 in 50 shot and you get the 70% score, or you get a flat 50% because it played a heart for everyone. So there is really nothing to be learned here, other than sometimes people get lucky for no reason. Nothing other tables did made it more likely they'd get a better score.