smerriman

You're right - it would actually be very clear-cut for GIB if it simulated, because a response guarantees a very old-fashioned 6 HCP (with the only exception being if holding heart support). And a quick sim suggests game is making double dummy about 65% of the time opposite a known 6 count. How much that would be dragged down when including weaker hands like the one shown, I'm not sure though; would need to sort out some specifics about exactly what constitutes a response (and also what rejects the invite later).

Yeah, 3♥ wouldn't have been my choice - it won't bid 4♥ without either 20+ total points (nope, it counts 18) or 18+ total points (yes) with 'strong rebiddable hearts' - which is AKQ or 4 of the top 5 honors (nope). And is told not to simulate to see if it's worth digressing from the rules. In saying that, Kx was the perfect trump holding; switch the position of the major honors or replace it with a couple of minor suit Qs and game is looking a lot shakier. And even with Kx, you still needed friendly diamonds. I guess if it's going to have a fixed set of rules, they're probably not the worst definition.

There aren't any rules, at least not in this situation (but perhaps others like leading from a sequence). It's not bidding related, it just sees 2 and 5 both score identically and chooses one at random.

You mentioned being surprised - to clarify, was your expectation 1♠, 1N, or double? I would pass myself (and probably double as South), but I've been known to make some silly balancing choices, so I have no idea what would be optimal. Probably a good ATB poll.

You passed, showing potentially nothing, and then doubled, which is 100% undefined. I don't know what the exact algorithm is, but assuming the opponents play GIBberish is always one given, so when everything is completely haywire you're not going to have much luck expecting GIB to figure anything out.

Whenever someone makes general statements about what "no one","we", or "everyone" thinks or wants, I replace those words with "I". It makes the post a lot more accurate. "We" get what you think; no need to imply everyone else agrees.

No, it doesn't care what heart you used at all. East showed 19-21 HCP. They clearly :rolleyes: hold the ♥J, because 10 HCP is exponentially closer to 19-21 than 9 HCP B-) . So you can get at most 1 trick; a diamond ensures this trick, locking declarer in dummy (required if declarer holds the ♥T too).

I feel like you haven't been reading my posts very closely. I said at the very beginning that without Soloway, I would continue with 3♠, and end up having no choice but to leap to 6♠. All of the comments after that suggested that by going slowly, you'd be able to bid more constructively, so I assumed this meant people were advising not leaping to 6♠, and was looking for what the alternative was. I meant, of course, if continuing to play some form of strong shift, that said shift moved on from the old fashioned strong to Soloway, being considered an upgrade exactly as you just said yourself. To the extent that it changed what the acronym SJS usually stands for. But it seems from Cyberyeti / jillybean that I was wrong.

Sorry, I more meant in the sense that there are some very old fashioned bids that you would be very surprised to hear played today - like strong 2s - because more modern methods - weak 2s - usurped them. I had mentally placed strong [non-Soloway] jump shifts into the same category, and wasn't aware anyone still played them.

@mike: OK, that sounds better. So if partner has a singleton spade, they raise to 4♠, and we can keycard or cuebid from there. If partner bids 4 of a minor, the only bid is 4♠? And 3nt from partner, having already denied a stopper, shows.. a partial stopper? @jillybean - ah, your first S meant strong, not Soloway. I have no idea how they work then, I thought they died out years ago when everyone moved onto Soloway.

"SJS asking partner to cue" is not how SJS work, so the auction won't go like that at all. I see we've had more comments that 3♠ in the original auction shows a self sufficient suit - can someone respond to my earlier comment about how auctions are meant to proceed after you were looking for a heart stopper if 3♠ is no longer an option? Maybe this is what nullve was getting at but it wasn't clear what 3nt meant - something artificial?

The hard part to me is figuring out which. Specifically, 2♥ here might be looking for a heart stopper for 3NT - say, 6322. So after hearing we don't have one, my 3♠ feels more like I still don't know what trumps should be. Partner might have a void spade, so if they bid a minor over this, it's probably natural, still looking for the best spot. Now the only spade bid I have left is surely a signoff. Which is why on this hand I feel like I'm going to have to end up arbitrarily gambling on a jump to slam. (And why I like Soloway, though the strong one suiter is very rare; the splinter comes up much more often).

When logged into BBO, and looking at a hand in the bottom right corner, click the button to show double dummy. Previously, it would take about half a second for the numbers to appear, both initially and every time you click next trick. That half a second no longer exists.

If I were forced to play without Soloway, I would bid as you did, then try 3♠, then in a couple of bids time realise there's no way to keycard or cuebid or investigate slam, gamble by jumping to 6♠, and hope I'm right.

.. the bot doesn't give or interpret any signals / previously played cards in defense other than some lead rules like 4th best, and switches because it thinks you are just are likely to hold Kx or Kxx as K. In a *very* early version of GIB (possibly even prior to BBO picking it up), it used to weight hands based on the conditional probabilities that each played card was a mistake double dummy. This was disabled, either because it was too slow, or because it simply led to erratic behavior (consider the numerous issues that GIB already has putting too much weight to the bidding, and imagine that applied to every single played card..)

[hv=handviewer.html?nn=Argine&n=SQHA95DAT6532CQ96&d=w&a=p1d(4+!d; HCP 11-23; 4-card !d opener)P1S(4+!S; HCP 4+; natural - Forcing)P2D(5+!D; HCP 11-16; unbalanced)P3H(4+!D; 4+!S; HCP 13+; Splinter, shortness in this suit and big fit in last suit bid - Game forcing)P4H(5+!D; 3-4!H; HCP 11-15; to play)]450|200[/hv] Robots just can't get splinters right, can they? Admittedly, this may be yet another case of the logic not matching up with the description, given 4♥ seems a reasonable bid otherwise.

A worthwhile maxim in general, but you already know GIB can't possibly have a trump stack from the bidding. I'm not sure how you'd define penalty and takeout double in your human partnership here (does the distinction even exist?), but it's definitely not doubling under a specific expectation that you'd take it out; it will double with lots of other shapes (e.g the same shape with singleton in a side suit instead). In the old database, GIB believes double here would show 24+ HCP and 14+ total points. Yep, 24+ HCP, because the rules tell it that it should double if the partnership holds a minimum 24 HCP and 26 total points between them, and South has shown 0 HCP in the description to date. >= 24 and >= 14 combined with its original pass of < 12 and < 13 combine contradictorily to give the final description. In fact, in the old database, there are 0 rules which match in this situation - not even pass, but the book bid defaults to pass if nothing matches. But when simulating, it's allowed to choose between pass, double, and 5♣, and 5♣ wins by a country mile. Something has changed since then; in particular, I believe that GIB has been programmed that if nothing matches, it's not allowed to pass out a contract if it knows the partnership has a minimum 24 total points combined fullstop, ignoring HCP. It thus doubles with this hand, and with KQx-void-Jxx-Kxxxxxx, but not with a total point less; but since this isn't a real rule, it still uses the old description which includes 24+ HCP when outputting the result, which is where the 11 HCP comes from. For unknown reasons, this must also be preventing it from simulating and coming up with the 5♣ bid. All a bit of a shambles really, but not the classic penalty vs takeout double issue at all.

Click Edit -> Click Use Full Editor -> modify the title.

The last time GIB was updated was over 4 years ago. I'll leave the rest to you..

It's easy to complain about many things on BBO, but I have to say.. the double dummy update is *amazing*. I'm not sure how it works (yet? ;)) but double dummy scores now load instantly (inside BBO - not the standalone handviewer yet), including as you step through trick by trick. No network requests are going out anymore, suggesting it's possibly a local DD solver (or maybe it's all precomputed somehow, though I don't see that in the request).. but the difference is huge. Thanks.

Possibly nonobvious, but I feel as though this is one they'll all pull off without actually intending to.

I don't think the pitch is relevant to the hand - the 10th trick comes from the marked finesse / endplay in spades whatever West throws. Same outcome.

Also, 1 in 3 reporting some form of disconnect is a lot higher than I thought (though of course there will be some bias that people experiencing disconnections may be more likely to click the thread - but not excessively so). And more than BBO have probably made out to be the case in the past. Definitely no 'alibi' here :)

The results seem pretty clear and line up with previous threads on the topic. - 2/3 of people experience no problems at all, with another 15% rare disconnections - 10% experience occasional disconnections - 9% experience regular disconnections with the usual reports that: a) temporarily switching to a different app results in a disconnection b) any slight interruption to the connection / loss of a data packet results in a disconnection a) is a result of b) since e.g. iOS in particular places heavy restrictions on background apps. And b) seems to come down to the fact that BBO constantly pings the client and a single lack of acknowledgement results in a disconnection, which is a very poor system. So the conclusion is that people like Kc 11 above most likely *do* have some form of issue with their connection, which may be somewhat resolvable locally, but BBO should have a better system in place regardless, since such connection issues don't affect any other areas of internet browsing.

I mentioned abstaining if you expected disconnections, so you don't qualify to abstain :) Agree it could be more robust to packet loss but at this point I'm more just narrowing it down to the fact people experiencing issues must be experiencing packet loss first (and that this is rare), rather than having a perfect connection as others claim. Of course, they probably think they have a perfect connection because the issues don't affect other websites. But the poll should include them as non-abstainers.