smerriman

Argine is horrible. [hv=handviewer.html?nn=Argine&v=b&n=SAK86HJT5D6CKJ985&d=w&a=2D(1+!c; 6!d; HCP 5-10; weak hand)X(HCP 12+; take-out - Non forcing)P3D(HCP 7+; cue-bid; strong, or not natural bid available - Forcing)P3S(4!S; HCP 12-18; natural)P4D(HCP 7+; cue-bid, strong, or no natural bid available - Forcing)P4N(4!s; HCP 12-18; Blackwood, asks how many Aces - Forcing)P5H(HCP 7+; 2 Aces - Non forcing)P5N(4!S; HCP 12-18; Blackwood, asks how many Kings - Forcing)P6C(HCP 7+; 1 or 4 Kings and not counting King of trumps - Non forcing)P6N(4!S; HCP 18; to play)PPP]400|250[/hv] I was happy when 4N was described as Blackwood in spades, because I had a strong hand but no way to raise spades without repeated cuebidding. Until Argine decided to play in no trumps, losing the first 6 diamond tricks off the top.

Argine forgets it's playing ACOL and plays SEF here instead. Don't how SEF works, but presumably agreeing one of the majors as trumps somehow.

Saw a hand talked about just today on BridgeWinners where someone held 11 tricks in their hand and people were still suggesting opening at the one level. I guess you can pass, but you'd better be right.. and shouldn't be limiting your reverses out of fear this will happen regularly.

Well, you can use your judgement and pass an artificial 2♣ opener when holding a weak hand with clubs too. But your partner is unlikely to be very happy :) A reverse is just as unlimited as a 2♣ opener.

No, a reverse is forcing.

The usual advice about 6-5 hands is that it's OK to reverse with less HCP than a normal reverse, because you can rebid your 5 card suit a second time, showing 6-5. If you had 6-4, you only bid it once, then go back to your first suit.

Fixed your link. I had the same hand, and elected to bid 3♠ as South. West tried a 4♥ cuebid, intended as Michaels, but described as strong (despite being a passed hand). East was either asking for the minor, or bidding a natural 4NT for some reason, but it ended up being interpreted as Blackwood. [hv=https://www.bridgebase.com/tools/handviewer.html?bbo=y&lin=st%7C%7Cmd%7C2SKJ975H6DAK7CAQ53%2CSAQT62H93DQT9853C%2CS4HQJ87542DJ4CKJ6%2CS83HAKTD62CT98742%7Csv%7CB%7Cah%7CBoard%204%7Cmb%7CP%7Cmb%7C3H%7Can%7C7-8%21H%3B%20HCP%205-10%3B%20preemptive%7Cmb%7CP%7Cmb%7C3S%7Can%7C5%2B%21S%3B%20HCP%2014%2B%3B%20forcing%20-%20Forcing%7Cmb%7C4H%21%7Can%7CHCP%204-11%3B%20cue-bid%2C%20strong%2C%20or%20no%20natural%20bid%20available%20-%20Forcing%7Cmb%7CP%7Cmb%7C4N%21%7Can%7CHCP%2017-%3B%20Blackwood%2C%20asks%20how%20many%20Aces%20-%20Forcing%7Cmb%7CP%7Cmb%7C5C%21%7Can%7CHCP%204-11%3B%201%20or%204%20Ace%28s%29%20-%20Non%20forcing%7Cmb%7CP%7Cmb%7C5N%21%7Can%7CHCP%205-17%3B%20Blackwood%2C%20asks%20how%20many%20Kings%20-%20Forcing%7Cmb%7CP%7Cmb%7C6D%21%7Can%7CHCP%204-11%3B%200%20or%203%20King%28s%29%20and%20not%20counting%20King%20of%20trumps%20-%20Non%20forcing%7Cmb%7CP%7Cmb%7C6N%7Can%7CHCP%2017%3B%20to%20play%7Cmb%7CD%7Can%7C5%2B%21S%3B%20HCP%2014%2B%3B%20penalty%7Cmb%7CP%7Cmb%7CP%7Cmb%7CP%7Cpc%7CH6%7Cpc%7CH3%7Cpc%7CHJ%7Cpc%7CHA%7Cpc%7CD2%7Cpc%7CD7%7Cpc%7CDQ%7Cpc%7CD4%7Cpc%7CDT%7Cpc%7CDJ%7Cpc%7CD6%7Cpc%7CDK%7Cpc%7CS7%7Cpc%7CSQ%7Cpc%7CS4%7Cpc%7CS8%7Cpc%7CD9%7Cpc%7CC6%7Cpc%7CHT%7Cpc%7CDA%7Cpc%7CC3%7Cpc%7CS2%7Cpc%7CCK%7Cpc%7CC2%7Cpc%7CCJ%7Cpc%7CC4%7Cpc%7CC5%7Cpc%7CH9%7Cpc%7CHQ%7Cpc%7CHK%7Cpc%7CS5%7Cpc%7CS6%7Cpc%7CS3%7Cpc%7CSJ%7Cpc%7CSA%7Cpc%7CH2%7Cpc%7CD3%7Cpc%7CH4%7Cpc%7CC7%7Cpc%7CS9%7Cpc%7CD5%7Cpc%7CH5%7Cpc%7CC8%7Cpc%7CCQ%7Cpc%7CD8%7Cpc%7CH7%7Cpc%7CC9%7Cpc%7CCA%7Cpc%7CST%7Cpc%7CH8%7Cpc%7CCT%7Cpc%7CSK%7C]400|300[/hv] Really bizarre how the robots make a bid expecting it to mean one thing, but interpret it as another. Seems to be a fundamental issue with how they're programmed. GIBs approach seems better and better the more I play with Argine! Sadly, misplayed for only down 4 and a 58% score :(

Can't answer the poll since I don't know what others would do. But seems like an easy double to me. Edit - yay, I made another insane bid :) Oh well.

I would start by eliminating browser add-ons as a possible cause: in Firefox, Help -> Troubleshoot Mode -> Restart and see if you can still replicate the issue there.

A link to it, close enough: https://www.dropbox.com/s/tuaonwuwess18gn/TableOptions.png?dl=0

Can the leap-to-5M bid be outright deleted from Argine's database? It bids it so regularly, and (amongst other reasons) makes it impossible to have any form of meaningful slam auction, especially when it seems to have no idea what it's meant to mean. On the most recent occasion, according to the definition and hand itself, it decided to use it on a keycard-less minimum. [hv=?nn=Argine&d=s&n=SK852HQT7643D3CQJ&v=b&a=2C(HCP 18+; forcing to game - Game forcing)P2H(5+!H; HCP 8+; natural - Forcing)P3H(3+!H; HCP 20+; fit - Forcing)P5H(5+!H; HCP 8; fit)]350|300[/hv]

Upload it to a site like imgur.com. You'll then get a link to use here.

If you hadn't yet realised, all of johnu's posts in this forum are sarcastic. The lead is of course ridiculous for the reason you mentioned, but GIB's algorithm is based around double dummy analysis, so assumes the opponents will know it is singleton. It's a fundamental issue with GIB but isn't something that will ever be fixable.

Well, the description is accurate :) Never heard of 1430 king responses though, especially when it's rather tricky to hold 4 out of the 3 total.

Almost certainly this bug again. Couldn't replicate GIB playing the ten in the old version with 100 sims but one of the runs had 84 occasions of playing the ten losing 1 IMP, and 16 occasions of playing the ten "gaining 5 IMPs": DT: 100? D4: -90 which made it almost a toss-up decision as to whether to inexplicably throw away the trick or not. If you're still playing with free robots, even more likely this will happen with less sims.

Well, it's not documented anywhere, including the ACOL system notes, but Argine now plays RCKB. Not specific kings though.

GIB is programmed that in advancing 1♠, it needs 5-7 total points to bid a LOTT 3♠, and not to simulate to overrule its decision (wise in this early stage in the auction IMO). So it passes. GIB is unrelatedly programmed that later in the auction, if it determines it has a 10 card spade fit and the opposition have shown values, simulate a number of deals matching the auction to date, and determine whether 4♠ would score better in the long run than passing. If so, bid 4♠. It runs this simulation, sees that pass is best by a long shot, therefore passes. The free version of GIB is crippled by changing the logic to this: GIB is unrelatedly programmed that later in the auction, if it determines it has a 10 card spade fit and the opposition have shown values, simulate a number of hands for partner, and determine whether 4♠ would score better in the long run than passing. If so, bid 4♠.

If they play 2♥ is nonforcing, then they need another bid to make when they have a good hand with hearts. Which is where NFB comes - you're forced to double or jump shift every time you have a good hand: https://kwbridge.com/bb/nfb-1.htm But it's a very different system from standard bidding, and not something you could ever play without agreements. More likely, they're not playing this at all, and would bid 2♥ with anything, which of course isn't playable.

Ha, this one is really good. So if North isn't a passed hand, 3♥ is as I described before. It's only broken if North is a passed hand. In the ~10 year old version of GIB that I can debug with, it has specific rules that say that after a 1NT response to a takeout double: - a new suit at the 2 level is 17+ if partner is an unpassed hand, but *11 or less* if not - a new suit at the 3 level is 20+ if partner is an unpassed hand, but *11 or less* if not BBO must have fixed 2 level bids, since it now shows a good hand either way, but never adjusted the latter. So I guess the original double showing 12+ and the 3♥ showing 11- wipes out the rest of the description. Why the original code distinguished them based on partner's position, I have no idea. It probably intended to check whether *you* were a passed hand, rather than partner.

Never make an off-shape takeout double with GIB; it can't handle any followups. But your description doesn't match; I assume you meant East opened 1♠, but the description of 3♥ is 20+ HCP, strong rebiddable hearts, and GIB raises to 4♥. Can you post the original hand with alerts?

If you have nothing other than basic agreements, then 2♥ was insane, and you already got lucky. Negative free bids isn't a basic agreement. Double the first time and hope you can bid 2♥ on the second round to say that you were too weak for an initial 2♥. If not, you've said enough.

EDGAR did some testing with robot data (not for determining cheating in robot games), and found the north robot looked suspicious :) (Because its leads always seemed to hit partner with the best hand.)

Thanks all. I did bid 4♥ and was allowed to play there. I seem to have lost the original hand record which is annoying as I wanted to recall exactly what stopped the opponents competing further, but partner turned out to have the absolute nuts - something like KJxx in hearts, K 5th in diamonds, and the ace of clubs. After a spade lead 12 tricks were laydown, but I picked up the 13th in diamonds as well. Really try to avoid resulting, but nonetheless it's not easy to avoid an 'oops' feeling when you go for a somewhat preemptive 4♥ and see 12 immediate tricks. So I'm glad this bid was shared by others.

[hv=handviewer.html?s=HAQT98543DAJTCT4&v=b&d=w&a=1C1D1S]200|300[/hv] How would you bid this one?

The old version of GIB says that 4♣, if bid, should be described as 20+ total points, and 5+ clubs.. .. and the logic part of the rule says that GIB is only allowed to bid it if it has 20+ total points and 5 clubs. Exactly 5. The mistake seems reasonably clear here :)