smerriman

What's 3♣? Isn't this usually the limit raise when you agree to make 2♦ GF?

steve2005's original hand: [hv=https://www.bridgebase.com/tools/handviewer.html?lin=st||pn|StevE2005,~~v3fakebot,~~v3fakebot,~~v3fakebot|md|3SQ6HAKQ92DA2CK984,SKT94HDQ83CAQJ652,S8732H8654DJ76CT3,SAJ5HJT73DKT954C7|sv|n|rh||ah|Board%205|mb|P|mb|P|mb|1H|an|Major%20suit%20opening%20--%205+%20!H;%2011-21%20HCP;%2012-22%20total%20points|mb|D|an|Takeout%20double%20--%203-5%20!C;%203-5%20!D;%202-%20!H;%203-4%20!S;%2012+%20total%20points|mb|P|mb|1N|an|One%20NT%20to%20double%20--%203+%20!H;%203-%20!S;%205-10%20HCP;%20stop%20in%20!H|mb|2C|an|4+%20!C;%205+%20!H;%2021-%20HCP;%2015-22%20total%20points|mb|D|an|5-%20!C;%203-5%20!D;%201-%20!H;%203-4%20!S;%2016+%20HCP;%20biddable%20!C|mb|P|mb|P|mb|P|pc|D3|pc|D6|pc|D9|pc|DA|pc|S6|pc|S4|pc|S2|pc|SJ|pc|C7|pc|C4|pc|CJ|pc|C3|pc|CA|pc|CT|pc|H7|pc|C8|pc|C6|pc|H4|pc|D5|pc|C9|pc|CK|pc|CQ|pc|H5|pc|H3|pc|HA|pc|C5|pc|H6|pc|HJ|pc|ST|pc|S3|pc|SA|pc|SQ|pc|D4|pc|D2|pc|DQ|pc|D7|pc|D8|pc|DJ|pc|DK|pc|H2|pc|S5|pc|H9|pc|SK|pc|S7|pc|C2|pc|H8|pc|DT|pc|HQ|pc|S9|pc|S8|pc|HT|pc|HK|]400|300[/hv] In the 10-year old version of GIB's database, there is a rule which describes what it should do after the opponents make a penalty double. When testing whether to bid 2♥, there are a few conditions it checks, but the most important is it checks whether [my hearts + min hearts partner showed] is strictly greater than [my clubs + min clubs partner showed]. So in this case, 4 + 5 > 2 + 4, so it bids 2♥. Seems reasonably logical. Well, except it doesn't do that, because the programmers used the wrong code for counting the number of hearts it holds. Instead of (m=.), they've put m. instead. The latter works in some areas of the database, but not in the specific context used in that rule. So the number of hearts it holds is set to 0. In the above case, it calculates 0 + 5, which is not bigger than 2 + 4, so it passes. While it's always possible they updated this rule since then, it seems unlikely, because all other examples still match: - give the robot 7 hearts and 1 club, and it still passes, because 0 + 5 is not greater than 1 + 4. - give the robots 3 hearts and 0 clubs, and now it bids 2♥, because 0 + 5 is greater than 0 + 4.

Wow. So the explanation for this one is.. too wow to be a simple reply to this thread. I'm going to make a new one with a more eye-catching subject.

Not 100% sure what you mean, if you're asking about the value of Jx, its formula for total points is HCP + 1 per doubleton, 2 per singleton, 3 per void, minus 1 for each short suit containing an honor. So you can say it adds an extra point for Jx and Kx, but subtracts them both off again.

GIB doesn't know how to cue bid, fullstop.

Its rules for this situation when it's weak and picking a suit, in order: 1: bid 2H if your longest suit is a 5 card minor, and your second longest suit is 4 hearts 2: bid 2x if x is your longest suit, and x isn't spades, and it's at least 5 cards long 3: bid 2x if x is your longest suit, and x isn't spades, and it's exactly 4 cards long 4: bid 2x if your longest suit is spades, and x is your second longest suit, and it's exactly 4 cards long 5: bid 2x for any x other than spades which is exactly 3 cards long Pretty clear logical error in the database, missing the case when you have 5 spades and 5 of something else. And yes, this means if you make North 5125, it will pass the double, not because it wants to penalise, but because now none of these rules match.

None of those things would change my answer.

No.

It's a known gap in GIB's database. It used to open this 3♠, but BBO changed the logic for preempts in 2017 to be "much more in line with what most bridge players would expect" [sic ;)] In doing so, they introduced a gap where it won't bid 3♠ here due to it having 10 HCP, but also independently fails its original rules for opening at the 1 level which require more total points (it thinks it has 11, since it adds one for the diamond doubleton only; it requires 13 to open a 10 HCP hand 1♠).

[hv=handviewer.html?d=w&nn=Argine&n=SAK8H9872DT4CK862&a=pp2D(1+!C; 6!D; HCP 5-10; weak hand)2N(HCP 15-18; balanced, no 5-card major)P3H(5+!H; HCP 11-; natural)&v=b]250|250[/hv] Argine has a game force opposite a 2NT overcall. It decides to bid 3♥, natural, showing 5 hearts, and 0-11 points. Not even sure where to start here.

Pretty please.. it bids it almost once per tournament, with a different unknown meaning every time. This one similar to thepossum's above, but it doesn't show a single thing it hadn't previously other than a spade fit. [hv=handviewer.html?nn=Argine&n=s9632hkq84daqj93&v=e&d=s&a=1C(4+!C; HCP 11-23; natural)P1D(1+!D; HCP 4+; natural or fit in !c - Forcing)P1N(4!C; 2-4!D; 2-4!H; 2-4!S; HCP 15-17; balanced)P2C(1+!D; HCP 7+; relay - Forcing)P2S(4!C; 2-3!D; 2-3!H; 4!S; HCP 15-17; exactly 4!S and denies 4!H - Non forcing)P5S(1+!D; 4-5!S; HCP 7+; fit)]440|250[/hv]

Right, but that's not how GIB plays it. Agree it's closer than the current description, but descriptions are all generated according to strict rules. Don't think it's really possible to generate such a description based on the database format.

If you wanted (10 year old) GIB to describe its actual logic of when it would bid 1NT over 1C, it would be along the lines of: - doesn't match 13 different rules with various conditions about having values for slam - doesn't match a rule with various conditions that tells it to pass - doesn't match another rule that tells it to bid an unusual 2NT - doesn't match 5 different rules that are various situations when it can double and then, either of these two rules match: - it has 15-18 HCP, its longest suit is 4 cards or a 5 card club suit, its shortest suit is 2 or 3 cards, and it has "stopper quality" = 4 (length + HCP in suit at least 8) OR - it has 16-18 HCP, its longest suit is 4 cards or a 5 card club suit, its shortest suit is 2 or 3 cards, and it has "stopper quality" at least 2 (2 = Kx or Qxx, 3 = A, QJx, or 5 HCP in the suit) But that doesn't really fit in the alert box :) Of course, some of the rules about slam would never overlap with the 1NT definition, but some of the doubles might, so you can't narrow it down to any hand which matches those last two rules. And it's not the same definition for overcalling a major. And that's actually a very simple scenario, if you wanted to know what a double meant here, it's not just the 5 possible rules alluded to above, but 3 others that come later after it rules out 1NT, Michaels, and 4 level bids, then another rule that it only looks at if it fails to match 25 other rules for suit overcalls..

Are you the table host? Table hosts need the extra click as the first one allows them to choose additional options like booting the player. It's a single click for me when not hosting.

I suspect there are very few humans who know the exact conditions that result in a robot making a bid, and while alerting it as a 1NT overcall is silly, there probably isn't a single example of a bid where the displayed definition outlines it precisely. There are thousands of conditions that need to be checked to exclude alternative bids / prioritise others. What would you prefer instead? Simplest solution here is if the human explains their understanding of the robot bid, since that's all that really matters, even if it's inaccurate.

We have put it to the test. Many people claim this, and each time a proper analysis is done - rather than what you 'feel' is true, which is always heavily biased - everything works out at perfectly 50%. The fact people keep claiming it is unsurprising, since studies have shown humans are very poor at recalling this type of information correctly, remembering the failing cases far more than the successful cases. It is, of course, possible that you are taking poor finesses. For example, from the opening lead, if the robot leads a suit and you have AQ in dummy, the finesse will almost surely lose - because the robot almost never leads away from a king. This doesn't mean finesses fail more than 50% of the time.

I don't think declare only IMPs is any more realistic though. In fact, given people play different hands, I would say far less so, as it just increases the luck factor of needing to be dealt swingy hands to score well. (Already an issue with all formats, but this would be the one affected the most.)

So did I - remember reading that one numerous times - but there's no connection to Roald Dahl, is there? I did enjoy several other Roald Dahl books, but reading taste is very individual. Most adult books that get highly rated, I avoid at all costs. I'll admit I usually prefer the simplicity of teen fiction :)

Others have reported the same bug where people getting subbed out of a tournament also get booted out of BBO, so it seems this isn't strictly true, albeit unintentionally: https://www.bridgebase.com/forums/topic/87937-why-do-you-boot-players-when-they-get-subbed/ . You might have misread it though since the only response was about chat logs.

I'm staggered than 5NT wasn't Blackwood, Ace-asking. Well, not according to West, anyway. Might have been 1430 king asking, excluding diamonds, for East? But yes, I enjoyed the Q♥ play.

Of course ruffing hearts can't cost. GIB just doesn't know finessing in hearts can cost. 21 HCP is obviously impossible but it still takes it into account when simulating, and all hands where East had the king were too far away from what it was trying to simulate to be considered.. Age-old bug with some easy alternatives, but nothing was ever implemented.

BBO haven't said anything officially about GIB for many years, so you're unlikely to get an official response, but I can quote from a reliable source that "it's been a long time since someone actively monitored GIB reports".

:( West's double shows 21+ HCP in GIB's system, so it's another classic example of ignoring the fact there's trivial 100% lines in favor of reading too much into the bidding.

Please.. [hv=handviewer.html?nn=Argine&n=st965ha5dkqjckqj5&v=n&d=n&a=1S(4+!s; HCP 11-20; natural)P2N(4+!s; HCP 10+; fit and forcing to game - Forcing)P3N(2-4!c; 2-4!d; 2+!h; 4!s; HCP 15-17; balanced)P4C(4+!s; HCP 15+; control in !c - Forcing)P5S(2-4!c; 2-4!d; 2+!h; 4!s; HCP 15-17; hope of slam)]400|300[/hv]

Try going to the homepage ( https://www.bridgebase.com/ ) and use the Just Declare option there. It's the same one, just doesn't involve being logged into BBO. (Basically, those solitaire games are run on a completely external server, so don't involve any communication to the logged in one, making it think you're not doing anything. BBO have said in the past this isn't easy to fix, presumably due to the fact it's just an inner iframe).