[rhm]

awm, on 2013-August-27, 18:05, said:

Out of question, do you feel that you get good results from bidding 2♣ with weak hands which are only 4-4 majors and which are not willing to pass a 2♦ rebid by opener? My general experience is that these find a major suit fit about half the time, and that the wins from finding a 4-4 major fit are roughly offset by the losses when we play 2M on a 4-3 (instead of 1NT).

If you read

http://taigabridge.n.../dd/garbage.htm

carefully, you will realize that

a) Garbage Stayman is useful on many more hand patterns and strengths than people realize.
b) It is important to get out of 1NT with a weak hand even if nobody doubles.
c) With a weak hand even a 4-3 fit in a major shows on average a small profit overall over 1NT. (except when responder is 4333)

What matters is not whether you always reach a better contract, but whether your chances reaching a better contract outweighs the occasional loss.
For example with 0-1 HCP it is correct to bid Stayman with 4♠-2♥-4♦-3♣, even though you will have to pass a 2♥ reply irrespective of vulnerability or whether IMPs or matchpoints.
For me 1NT--2♣-2♦-2♥ always shows a weak hand, which does not want to play 2♦. For this reasons I do not like restrictions like one of the majors must be at least as long as the other etc. I am convinced they are inferior.
Opener is supposed to preference to 2♠ only with three spades and two hearts and will pass otherwise. If I bid 3♣ over 2♠ I show a weak hand close to 3♠-4♥-1♦-5♣.
Other continuations after 1NT--2♣--2♦ I do not play as a sign-off, because playing them as garbage would give you a diminishing return and there are better use for these continuations.

Rainer Herrmann

[awm]

There are several problems with this analysis.

1. 1NT has a significant declarer advantage over double dummy (because much depends on the lead). 4-3 major fits tend to be the opposite (because you want to draw trump if and only if they are breaking typically).

2. Opponents are more likely to double 2-suit. 1nt will rarely be doubled (some do not even play penalty) and you will have a better chance to scramble later if they do double.

3. Assumptions about opener shape are never explicitly given. My partners often open 1nt with 2245 and sometimes with 22(36) both of which can be disasters for some of these patterns.

[Siegmund]

Yes, it is a purely double-dummy result, and using Stayman WILL result in better defense.

Whether it is more likely to result in a penalty double is debatable - I can remember many more times I've used Stayman even after a double, and not been doubled in 2M, than times I have been nailed in 2MX after Stayman. One plus is only one opponent knows you are Garbaging in time to stomp on you.

Quote

3. Assumptions about opener shape are never explicitly given. My partners often open 1nt with 2245 and sometimes with 22(36) both of which can be disasters for some of these patterns.

I did several articles at once, and I see I didn't spell it out as clearly as I could have. The simulations allowed all 5422 patterns (I think-- I don't remember excluding 5S4H, and I know I DO open 1NT with 4S5H), and 6322s with a 6-card minor.
The "If you play very old-fashioned conservative 1NT openings" paragraph was based on 5332, 4432, or 4333, no 5-card major. I suppose there might be somebody who allows a 6-card minor but forbids a 5-card major; such a pair should be conservative in their use of garbage stayman too. But most everyone who ever opens on 2236 also opens on 5M332, and they are gaining a lot more from having some of their '4-3' fits really be 5-3, than they are losing on their minor-oriented hands.

(And whether everyone agrees with the findings or not - it is good to know someone is reading. Even if it makes me feel guilty that I have hardly written anything new in the past 2 years.)

[rhm]

awm, on 2013-August-30, 06:24, said:

There are several problems with this analysis.

1. 1NT has a significant declarer advantage over double dummy (because much depends on the lead). 4-3 major fits tend to be the opposite (because you want to draw trump if and only if they are breaking typically).

Not really true here when dummy is broke or entry less. The defense against 1NT then tends to be straight forward, while declarer has rarely good options.
Low level 4-3 fits present some unique challenges to both declarer and the defense, but, as is usually the case, the defense tends to be even tougher.

Quote

2. Opponents are more likely to double 2-suit. 1nt will rarely be doubled (some do not even play penalty) and you will have a better chance to scramble later if they do double.

Penalty doubles are much more common over notrump than suit bids. Nobody plays initial doubles of suit bids as penalty nowadays. The suggested philosophy contradicts common wisdom.

Quote

3. Assumptions about opener shape are never explicitly given. My partners often open 1nt with 2245 and sometimes with 22(36) both of which can be disasters for some of these patterns.

True, it depends on many factors like whether you are going to open 1NT with a 5 card major. However, even if 2-2 in the majors is a possible shape for opener, it will be infrequent enough not to matter much (much less frequent than a 5 card major if you open 1NT with 5M332).
I repeat what matters is not the occasional loss, but the frequency of gain.

Rainer Herrmann