nullve

I know a lot of pairs play 1D-(2C)-?: X = negative 2M-1 = 5+ M 2S = inv+ D raise That's a very interesting idea. It seems like I won't be sacrificing much by playing the double your way, because over 1M-(2x) I tend not to make negative doubles on hands with < GF values unless I have 2M2-x, anyway. I think I still have to do something on those hands, so, using your method, maybe I should often just X or raise to 2M, temporarily pretending to have 3-card support? (Ok, maybe a bit off-topic.)

Could you expand on that? I notice that you play 1C-1S; 1N-2m = to play, so I guess you have design goals that are different from mine, at least.

Sorry, didn't mean to use home-grown notation! I was assuming that the notation '1C-1M-1; 1M' is familiar shorthand for '1C-1D; 1H or 1C-1H; 1S', so I wanted to introduce the suggestive(?) notation'1C-1OM-1; 1OM' for '1C-1D; 1H or 1C-1H; 1S' when talking about a T-Walsh substitute where the 1D and 1H responses to 1C show spades and hearts, respectively, i.e. where 1D and 1H can be thought of as "transfers" to the "other" major, the major Responder hasn't shown. I believe the notation 'OM' for 'other major' is widely used, although I have to admit I introduced in a rather unconventional way. For example, I hadn't already introduced the notation 'M' for the known major. Yes, I know there are different versions of T-Walsh, since I've played quite a few myself. Maybe I chose the wrong title, because I never intended the thread to be about GCC legal substitutes for certain types of T-Walsh only. The OP was intended more as a recipe for turning your favourite version of T-Walsh into something GCC legal without completely destroying its identity. The substitute will no longer be (your favourite version of) T-Walsh, but hopefully something pretty close.

Judging by the responses to David Loeb's (dloeb's) article http://bridgewinners.com/article/view/a-piece-of-transfer-walsh-in-a-gcc-event/ on Bridgewinners, it may not be generally known among ACBL members that the scheme 1C-?: 1D = 4+ S 1H = 4+ H 1S = ART GF 1N = nat., NF, which enables the partnerhip to play something like T-Walsh when Responder has spades, is actually GCC legal. But a comment by Mike Bell (MickyB) points to an even more exciting possibilty: a full-fledged GCC legal T-Walsh substitute. I believe the real idea behind his comment* is that if Responder has some conventional way of dealing with Flannery hands with less than GF values**, then there's no compelling reason why Opener's 1S rebid should promise 4+ S***, so 1S may as well be used as "transfer acceptance" similar to 1M over 1C-1M-1 in T-Walsh. Of course, there's no transfer involved here unless one thinks of 1H as a transfer to the major Responder hasn't promised. (Same with the spade-showing 1D response.) By combining dloeb's scheme with MickyB's idea, one could play 1C-1OM-1****; ? 1OM: corresponds to 1M over 1C-1M-1 in T-Walsh. 1S = 4+ H1S(M=S): corresponds to 1S over 1C-1D in the emulated T-Walsh style, but now with 4+ H instead of 4+ S [corrected stupid error 5th August: '1S(M=S)' was just '1S'] 1N: corresponds to 1N over 1C-1M-1 in T-Walsh [corrected 3rd August ('1C-1M-1' was '1C-1D')] 2C+: obvious stuff, but depending on the emulated T-Walsh style But notice that, unlike in T-Walsh, Opener's 1OM/1S rebids are now not easily passable*****, and this could potentially be a huge problem (for more than one reason). But here are three ways to go about it, corresponding to different restrictions one might want to put on Responder's ability to respond 1D or 1H on "air" (as is common when playing T-Walsh). a) [restrictions on both 1D and 1H] Responder will always have something resembling an inital positive response when he makes his forced rebid, so continuations, apart from the now "impossible" sequence 1C-1OM-1; 1OM/1S-P, may be roughly as in the T-Walsh style one wants to emulate. b) [restriction on 1H only] Responder is allowed to pass 1C-1D; 1H/1S because of the following variation of the above scheme: 1C-1OM-1; ?: 1H = 4+ H, NF 1S: corresponds to 1M over 1C-1M-1 in T-Walsh, but denies 3-S4+H if M=S 1N+: as above Continuations, apart from the now "impossible" sequence 1C-1H; 1S-P, may again be more or less as in the emulated T-Walsh style. c) [no restrictions] Responder cannot easily pass Opener's rebid, so the continuations have to reflect this. E.g.: 1C-1OM-1; 1OM-?: 1S = 4+ H, as above 1N: as in T-Walsh, but could be extremly weak. (Conventional follow-ups may be necessary, depending on the emulated T-Walsh style.) 2C/2H+ = standard XYZ 2D = subminimum, 5+ M****** Any comments? * "Would 1C:1S as 5H4S be allowed? This would free up 1C:1H,1S to be artificial." ** To be GCC legal, a two-level response (not 1S) has to be used as Flannery, though. An interesting alternative, but which effectively takes the Walsh out of T-Walsh, is to respond 1D, canapé-style, with 4S5+H. *** Well, Opener may now have to rebid 2C with 4135/4045, depending on the emulated T-Walsh style. But this is the kind of problem that standard or T-Walsh player have with 1435/0445 over 1C-1S and 1C-1H, respectively, but which disappears when the response with 4+ S is 1D instead of 1H or 1S. **** Yes, the notation '1OM-1' suggests that we may continue to view the 1D and 1H responses as transfers to the other major (=OM), the major Responder doesn't necessarily have. This is convenient when making comparisons with T-Walsh. ***** The same is true of the sequence 1C-1D; 1H-1S, but that seems like a tiny problem in comparison, so I'll ignore it. ****** This is the easiest, assuming Responder would always replay 1S (ART GF) to 1C with 4+ M and GF values. A more advanced version would allow Responder to bid again over 1C-1OM-1; 1OM-2D; 2M(= to play, assuming Responder has subminimal values) on some hands with GF values.

Assuming Rule of 20, (14)15-17 NT and a decent AKQJ-normalised hcp count method (not basic Milton Work, which fails miserably on the hand Frances gave), then with 4351 I would * open 1D and raise 1S to 2S with 11-13 hcp * open 1N with 14-16 hcp* * open 1D and reverse into 2H over 1S with 17+ hcp The main idea is to keep the 2S range so tight that bidding game shouldn't depend on whether Opener is pointwise max or min. So I'd have at most 13 hcp for my raise to 2S with 3451. *I know the ACBL wouldn't let me to this systemically.

Not specifically, but see post #64. Sort of. Unless pass shows something crazy, but we've already discussed this. I do expect to be less lucky next time. Regarding the hand in 62), I was just trying to use the following defence:

I was zooming in on one of the conditions under which the 2C-2D; 2M mechanism threatens to collapse, as in glen's example. (Doesn't happen easily when Responder is 0-7 or 16+, I think) I honestly don't know what an honest long simulation would look like, since people tend to want to be able to overcall but I don't want to force a set of agreements upon them. But I thought this 100 deal simulation could still serve as some sort of (qualitative) illustration of what to expect under the given conditions. Just keep in mind that: The accidents in 30) and 57) wouldn't have happened if East had doubled, Dixon-style. The accident in 31) wouldn't have happened if East had made a normal-looking 2S overcall The accident in 44) wouldn't have happened if West had made a normal-looking, but marginal, 2S overcall The accident in 62) wouldn't have happened if West had made the kind of double you were talking about in an earlier post. The accident in 77) wouldn't have happened if East had doubled, Dixon-style, or overcalled 2N The accident in 90) wouldn't have happened if East had acted with his very strong hand The accident in 95) wouldn't have happened if East had made a normal-looking, but IMO misguided, 2S overcall The accident in 96) wouldn't have happened if East had made a normal-looking S overcall (either 2S, 3S or 4S) The accident in 99) wouldn't have happened if East had made a normal-looking 2S overcall or West had made a "gwnn-style" double So if EW were playing normal bridge (instead of passing throughout), as many as 10 of the 11 accidents could have been avoided. The ones in 90) and 99) even benefitted NS. In the "lab" I currently require 6M (i.e. 6M3-OM) or 5M5+m when vulnerable. But that's (the current) me.

I did a 100 deal simulation on BBO today with the following constraints: N: 0-9 hcp and either 6M3-OM or 5M3-OM4+m* S: 8-15 hcp, 1444 Assumption: N is dealer and EW pass throughout. Result: 11 accidents (i.e. S passed 2S when N had 5H4+m), which was actually a lot less than I expected. This is what they looked like: 30) N: K98-QJT65-2-QT43 E: QJT3-A3-Q943-A76 S: 4-K972-AT65-KJ82 W: A7652-84-KJ87-95 DD result: 2S-3 Par: 4H=(NS)/4SX-2(EW) [i.e. depending on vul] 31) N: A92-T9854-void-T9863 E: KQT863-3-J732-AQ S: 7-KQ62-AT96-KJ52 W: J54-AJ7-KQ854-74 DD result: 2S-5 Par: 5HX-1(NS) 44) N: K3-97643-A5-Q432 E: QJ98-K-86432-KT9 S: 6-AT82-QJT9-AJ65 W: AT7542-QJ5-K7-87 DD result: 2S-2 Par: 4H=(NS)/4SX-3(EW) 57) N: JT8-AK964-8-T972 E: K53-52-AQT94-A85 S: Q-Q873-K765-QJ43 W: A97642-JT-J32-K6 DD result: 2S-4 Par: 3S=(EW) 62) N: K76-T9432-T964-6 E: T43-Q5-J87-AT982 S: 5-KJ76-AQ32-KJ43 W: AQJ982-A8-K5-Q75 DD result: 2S-4 Par: 3S=(EW) 77) N: T62-K7642-J732-Q E: KQ5-A853-K98-KT5 S: 4-QJT9-AQT6-AJ93 W: AJ9873-void-54-87642 DD result: 2S-4 Par: 4SX-1(E) 90) N: T8-AT752-42-QT76 E: AKQ7532-K-AQ9-85 S: 4-Q983-KT73-A943 W: J96-J64-J865-KJ2 DD result: 2S-5 Par: 4S=(EW)/5HX-2(NS) 95) N: T72-97532-AKT5-9 E: AKJ43-void-J732-AQT2 S: 9-AQ84-9864-KJ53 W: Q865-KJT6-Q-8764 DD result: 2S-4 Par: 3S=(EW) 96) N: T-T8742-KJT-QJ74 E: AQJ7543-KJ5-Q75-void S: 8-AQ93-A942-965 W: K962-6-863-AKT82 DD result: 2S-6 Par: 4SX-1(EW) 99) N: 642-AT542-9872-J E: AK9873-97-Q3-A98 S: J-KJ63-JT54-KT54 W: QT5-Q8-AK6-Q7632 DD result: 2S-6 Par: 4S=(EW)/5HX-3(NS) 100) N: T8-Q8643-T9-AQ92 E: K93-K7-KQ532-T65 S: Q-A952-AJ74-8743 W: AJ76542-JT-86-KJ DD result: 2S-4 Par: 3S=(EW)/4HX-1(NS) (Yes, the 29 first deals were accident-free!) Note: The DD and par results were (quickly) calculated by me. Check them if you want. *Although I've described 2C as showing 6 M or 5M4+m, I don't really open 2C with 6M4+OM or (54)(40). Also, I don't use 9 hcp as the upper limit regardless of shape, so the above constraints on North are somewhat simplified.

That's correct.

Instead of chucking the 2M rebid, I've so far played "system on" after 2C-(X), except that 2C-(X)-?: P = intending to play 2CX opposite 4+ C P = 4+ C XX/2D+: never agreed on anything here, but in practice maybe: XX: not sure (SOS with 5M3OM4D?) 2D = 4+ D 2M = 6 MXX = long suit (particularly useful when Responder wants to get out in 2D) any = P/C Two posts ago I argued that 1M3OM is conceivably a problem of the type discussed only when Responder has about 8-15 hcp and either 1M3OM(54) or 1M3OM(63). But after 2C-(X), Responder may pass with 1M3OM(54) or 1M3OM36, intending to play 2CX opposite 4+ C, and redouble with 1M3OM63, intending to play 2D.

I think Multi 2D vs. Multi Squared/MuSHroom is a good idea (and I believe an adaptation of something like Multi vs Multi or Granovetter's Best Defence to Multi may be the way to go), but you seem to tnink that * After 2C-(2D[=Multi]), I need to say bye bye to the kind of continuations I use after 2C-(P). This is not true, because it's certainly possible to play "system on" with X replacing the 2D relay. You also seem to imply that * After 2C-(2D[=Multi]), saying bye bye to the kind of continuations I use after 2C-(P) will somehow hurt. This is also wrong, because the 2D overcall enables Responder to pass many potentially misfitting hands with < inv values, to the point that he no longer needs a mechanism to find out whether Opener has 6 M or 5M4+m on partscore deals. In other words, there's no reason, other than simplicity, to play "system on" here.

gwnn has repeatedly been talking about or alluding to the (very real, but possibly exaggerated) danger of ending up in non-fit after 2C-2D; 2M-P when Responder has 1M3OM (a case of small length disparity between Responder's majors). To keep things straight, let me just point out that 1M3OM is conceivably a serious problem only if Responder has about 8-15 hcp and either 1M3OM(54) or 1M3OM(63). Reason: If Responder has * 0-7 hcp, then 2M on a non-fit ought to be good, since opps will have 24+ hcp between them. * 16+ hcp, he will be able to continue with 2N (inv+ relay) over 2C-2D; 2M * 15- hcp, 1M3OM7+m, he'll be determined to play 3m, anyway, and may therefore not have responded 2D in the first place. Even 8-15 hcp, 1M3OM36 may not be be a serious problem, since Responder always had the option to pass 2C. (Yes, I forgot to list Pass as a possible response to 2C.) A good reason for doing that (apart from avoiding non-fits, if that seems like a good reason) is that he may not be happy to play 3D on a likely 4-3 fit after 2C-2D; 2OM(6 OM or 5M4+m)-3C(P/C); 3D(5M4+D)-P. Similarly, with 8-15 hcp, 1M3OM63, Responder would rather avoid playing 3C on a likely 4-3 fit after 2C-2D; 2OM-3C; P(5S4+C), but what can he do? I can imagine creative attempts/systemic alternatives such as 2C-2H(P/C, usually with [edit:] 2+S2H or 3+S3H); 2S-3D(to play with 1363?) 2C-2S(P/C, usually 2S3+H); 2N(6H3-D3-C?)-3D(to play with 3163?), which I believe is similar to what helene_t suggested early in this thread, but for now, let's just say that with 8-15 hcp, 1M3OM63 may be more of a problem than 1M3OM36.

If simplicity is important, why not think your 2C rebid as a standard Precision 2C opening, except that it shows 11-15, 2-S6+C or 2-S4R5C instead of 11-15, 6+ C or 4M5C? Then you could play 1C-1H; 2C-2D(relay); ?: 2H/2N+: same as 2H/2N+ over 2C-2D in standard Precision, except that Opener has now already denied 3+ S 2S = 2-S4+ D, but with continuations analogous to those in standard Precision after 2C-2D; 2M

Which hand types with 2-S5C do you include in your 2C rebid? Just 1(43)5 or 0445 outside your 1N rebid range?

That's especially true in a weak-only version where of Multi Squared where 2N+ are also available rebids. Yes. So it looks bad compared to e.g. Muiderberg, but ok (or even good!) compared to some versions of Multi 2D. It has worked ok in practice, but in the "lab" I've toyed with a "muiderbergified" scheme: 2C-2D; 2M-2N; ?: 3C = non-max, 6 M or 5OM4+m (i.e. a third Multi!) 3C1 = allowed! 3D = reality check 3M = 6 M 3OM = 5OM4+m 3M = P/C 3OM = P/C3D = non-max, 5OM4+D 3H = max, 5OM4+C 3S = max, 5OM4+D 3N = max, 6 M EDIT: Structure corrected. I was mixing up two versions, but this is the simplest one. 1 [11 Aug 2016:] Meant to write 'P', not '3C'. Added, starting 21 January 2017:

Yes, unless either pass is something really out of the ordinary, like F1. No, not to me. Only when the meaning of Pass by Advancer allows me to pass with some hands as Opener, e.g. when 2C-(P)-2D-(P)=F1. So I think you've got everything right.

The subject of full disclosure is an interesing one, and although I haven't thought very deeply about it, here's what I think: Basically, I view bidding theory as a kind of two-player game, often played by one bidding theorist against himself for lack of opponents, but also by partnerships against other partnerships in real life encounters, and to a lesser degree by forumers against other forumers in threads like this. This game (actually only half of it - see below) respects a move order where Player I starts by assigning a meaning to at least some of the following calls: P, 1C, ..., 7N. Player II then assigns a meaning to some of the sequences (P)-P, (P)-1C, ..., (P)-7N, (1C)-P, (1C)-X, (1C)-1D, ..., (1C)-7N, and so on. Then it's Player I' turn again, and he may assign meanings to some sequences such as P-(P)-P, ..., P-(P)-7N or 2D-(X)-P, 2D-(X)-XX, 2D-(X)-2H, ..., 2D-(X)-7N. And so on and so on. There is also a dual game (the other half of the whole game) where Player II starts first. I think this is the only reasonable way the "game" of bidding theory can be played, so if e.g. an opponent wants to know my 1N defence before he decides upon his 1N range, I think he's got it exactly backwards. Still, I'd be happy to tell my opponent about my structure after 2C-(P)-2D; 2M, for each of his choices of P.

I absolutely agree with you here. But AKJxxx-x-Qxx-xxx is a quite marginal overcall, anyway, and I thought the idea was to pass also with stronger hands.

I know you're half joking, but you're touching a couple of interesting points. To describe the uncontested responses to 2C is the same as describing Responder's options over 2C-(P), where P has the standard/expected meaning. If your meaning of P is non-standard*, in which case I'm entitled to know about it, I might adapt by using a different scheme of responses. For instance, in the extreme case that P is forcing, I might make heavy use of Pass, thereby gaining a lot of extra bidding space and maybe also putting pressure on Advancer. If the only difference between a standard P and yours is that you will occasionally pass on a good hand with 6+ M, I personally wouldn't change anything, because even though the 2C-2D; 2M-P accidents are now more likely to occur, your (tongue-in-cheek?) tactics may backfire. You said that opps are well equipped to avoid disasters (and I assumed you were right, for the sake of argument), but if you choose not to overcall on some good hands with 6+ M, how can you always avoid being stolen from when the bidding goes (2C)-P-(2D)-P; (2M)? After all, it's precisely when opps have good hands that the 2C-2D; 2M-P accidents have any chance of benefitting the bidding side. Ironically, you may have given me one of the advantages of a Multi 2H opening, even one you've been talking warmly about, for free. * or frequently psyched, which effectively changes its meaning

To give you a counterexample, I'd need to know your defensive method, or else you could just choose or make up one ad hoc. But I'll be generous and assume you're right. My previous post was largely about disasters I can't avoid even in principle - unless opps are doing something really unexpected to help me out, that is. So am I partly relying on opps to rescue me? Just barely, as I believe (based on my own imperfect experience) that disasters of this type would be quite rare even if opps were allowed to choose their defence ad hoc. In practice, opps - even you? - will continue overcalling on, say, good hands with a 6-card or longer major, thereby automatically rescuing me from at least some disastrous 3-1 or worse fits.

I thought I should say a bit more about the different scenarios after the "Multi" 2M rebid in the case that Responder has < inv values and a potentially misfitting hand, although it seems that people are beginning to get the basic idea now - hopelessly flawed or not. Recall that if Opener is weak, he will always have one of the following hand types: 1) 6 H 2) 6 S 3) 5H4+m 4) 5S4+m Now assume that Responder has one of the following (common) shapes: a) 1444 or 13(54) b) 4144 or 31(54) c) 1-S6H or 15(43) d) 6S1-H or 51(43) The bidding for each combination: 1a) 2C-2D; 2H-3C; 3H-4H; P 1b) 2C-2D; 2H-P [assuming 6 H] 1c) 2C-2D; 2H-P [hoping for a 7-card or better H fit, hitting p with 6 H] 1d) 2C-2D; 2H-P [assuming 6 H] 2a) 2C-2D; 2S-P [assuming 6 S] 2b) 2C-2D; 2S-3C; 3S-4S; P 2c) 2C-2D; 2S-P [assuming 6 S] 2d) 2C-2D; 2S-P [hoping for a 7-card or better S fit, hitting p with 6 S] 3a) 2C-2D; 2S-P [assuming 6 S. Oops! 3-1 fit or worse] 3b) 2C-2D; 2S-3C; P/3D-P 3c) 2C-2D; 2S-P [assuming 6 S. Oops! 3-1 fit or worse] 3d) 2C-2D; 2S-P [hoping for a 7-card or better S fit] 4a) 2C-2D; 2H-3C; P/3D-P 4b) 2C-2D; 2H-P [assuming 6 H. Oops! 3-1 fit or worse] 4c) 2C-2D; 2H-P [hoping for a 7-card or better H fit] 4d) 2C-2D; 2H-P [assuming 6 H. Oops! 3-1 fit or worse] As the spectacular accidents in 3a), 3c), 4b) and 4d) are bound to happen*, like giant earthquakes in California, Multi Squared will look like rubbish unless one is able to deal with it philosophically when they occur. That may depend on the (perceived) frequency of the accidents and the (perceived) size the damage, just as when playing structures like * EHAA (Opener has no forcing opening bid available on strong unbal. hands) * standard Precision 2C (inevitable accidents when Responder is weak, occasionally the size of a game swing) * the Turbo convention (heightened risk of missing 2+ key cards compared with RKC)* * assumed fit preempts (p may not have the assumed fit) or even * RKC (see http://www.eurobridge.org/competitions/05riccione/Bulletins/21ThuPg4.htm for an amusing example) * preempts in general (p may have less support than hoped for) * just about any bidding structure, really (bad trump breaks, finesses are off...) In my (adimittedly subjective) experience, which you may have good (even mathematical) reason to distrust, accidents as in 3a), 3c), 4b) and 4d)) aren't nearly as frequent as some people might think. And although they tend to benefit the opponents, sometimes by more than game swing, they will occasionally benefit the bidding side, as when opps are cold for game in the same major suit. All in all, I'm not claiming this is a risk free way of playing bridge (there isn't any), but it might still be good bridge. * As gwnn pointed out, the risk is greater the lesser the suit length disparity in Responder's major suits is. ** The comparison with Turbo or RKC might be the most interesting, but I won't say more about it now.

I actually include fewer patterns than many who play Garbage Multi. Where I live, they typically include 5M(332) as well as 6 M and 5M4+m That has never caused a problem for anyone here although we're trying to follow the WBF Systems Policy. As for the WBF definitions of '(standard) weak two's' and 'Muiderberg' you're impliclty referring to, I wasn't aware they existed. Maybe, but then the same is true of an identically defined 2D opening. I wouldn't even if they let me. Agree. I think "strong options according to taste" has been misleading people into thinking that at least one strong option has to be included, although 'zero or more strong options' was what I intended. Sorry. Agree. I've always found it difficult to come up with good names for bits of bidding structure. But if the structure doesn't have a name, or if the name is too long, it becomes very cumbersome to talk about it. Also, if the name I give it isn't descriptive enough, it may be hard to remember what it's referring to. So I've tried to find catchy-but-descriptive names. 'Multi 2C' was unfortunatley taken, so I decided to go with 'Multi Squared' instead, which is sort of descriptive (to me, at least). 'MuSHroom' also serves its purpose, I think.

Yes, I agree with him in the sense that I wouldn't be happy to open, say, 2S (or whatever) with K-KQJ-xxxx-xxxxx. In fact, I would pass, pretending to be balanced. Of course, if that's a tendency of mine on hands like this (singleton king, low ODR), I really shouldn't say that Pass denies a singleton or void. But in a "meme-sized" (i.e. rough, easily remembered) description of P-P P, that seems (or seemed) accurate enough. Maybe it is, but isn't that at least debatable? I've presented three ideas/conventions that are related in the way that Multi Squared = rubbish => MuSHroom = rubbish => P-P P (with 2C+ = MuSHroom) = rubbish Fortunately, the opposite implications don't hold, so there is no reason to buy the whole package. My own view is that Multi Squared and MuSHroom are sound conventions while P-P P may be nothing more than an interesting (i.e. not obviuosly bad) idea.

Another example: weak hands with (41)44 shape. Using the proposed scheme of openings I would have to choose the least lie among Pass (denying a singleton/void), 2D/H (promising 5+ m unless unless 4+H/5+S) and 2S (5+m4+Om). Yuck.

2D is the correct response. As for the guessing, I think it's an exaggeration to say that it will be 'almost blind', but I see what you mean. Probability calculations might throw some light on this issue. EDIT: I think our guesses will be a bit luckier than rough probility calculations would indicate, for the simple reason that opps have chosen not to overcall, either directly over 2C, over 2C-2D or over 2C-2D; 2M.