[awm]

It's becoming increasingly popular to play 3♣ as puppet stayman over 1NT (combined with 2♣ as regular stayman). This seems like a bad idea to me. The typical reasons given (and my responses):

1. It's important to find our 5-3 major suit fits.

When holding two balanced hands, the statistics favor 3NT over 4M on a 5-3 fit. This becomes even more true if opener can use judgement as to whether 1M or 1NT should be opened on 5332 hands in range. The 3NT contract also tends to be more difficult to defend, such that double-dummy may understate the 3NT advantage.

Of course, responder doesn't have to be balanced (for example 3451 or 3361 hands) but in these cases pinpointing the shortness will help more than just checking for 5-3 major fits -- even without an eight-card major fit it can easily be right to avoid 3NT on such a hand (playing a 7-card major fit or 5/6m when opener is weak opposite the shortage).

2. Using puppet instead of regular stayman conceals opener's hand from the defense.

While the statement is true, there are some other factors. In particular, the puppet sequence gives the opponent who is not on lead the opportunity to double the major suit that responder doesn't hold (i.e. 1NT - 3♣ - 3♦ - 3♠ showing four hearts and the opponent in fourth seat can double). This is potentially very costly as it allows the opponents to find an unusual lead that will often set (or reduce overtricks). Sure, sometimes the opponents can double the final 3NT contract to ask for the same lead in a standard auction, but this is quite a bit more risky (not to mention occasionally it might not obtain the right lead).

The puppet auction also gives up more information about responder's hand (for example 1NT-2♣-2♦-3NT and we don't know which major responder holds, whereas the puppet auction reveals this). Probably not as useful to the defense as knowing opener's shape, but it will help on opening lead.

3. The 3♣ response to 1NT is otherwise infrequent and not that useful.

Whether this is true of course depends on the rest of your 1NT structure, but I find that a lot of the systems out there miss some hand types. One that is frequently lost seems to be the three-suited and semi-three-suited patterns (like 1444 or 1453). Sure you can stayman or puppet stayman, but just because you have no 8-card major fit doesn't make 3NT the best contract! It can easily be right to play elsewhere when opener has weakness opposite the shortage.

[Vampyr]

awm, on 2013-November-25, 19:16, said:

It's becoming increasingly popular to play 3♣ as puppet stayman over 1NT (combined with 2♣ as regular stayman). This seems like a bad idea to me. The typical reasons given (and my responses):

1. It's important to find our 5-3 major suit fits.

When holding two balanced hands, the statistics favor 3NT over 4M on a 5-3 fit. This becomes even more true if opener can use judgement as to whether 1M or 1NT should be opened on 5332 hands in range.

This is not true of every hand, though. As you mention, the responder might not be balanced. Using 3♣ as the beginning of a sequence to show shortness might be useful, yes. But the second sentence above is usually not the case, unless the suit is so good the player is happy to rebid it.

Quote

2. Using puppet instead of regular stayman conceals opener's hand from the defense.

While the statement is true, there are some other factors. In particular, the puppet sequence gives the opponent who is not on lead the opportunity to double the major suit that responder doesn't hold (i.e. 1NT - 3♣ - 3♦ - 3♠ showing four hearts and the opponent in fourth seat can double).

This is easier when playing a weak NT, since there are no worries about "rightsiding", and responder can just bid the suit she has. And don't forget that many hands with 4-card majors will have bid regular Stayman anyway, depending on the agreements in place. Remember that it is normal for opener's 3♦ bid here to deny a 5-card major but remain mute about the presence of 4-card majors.

Quote

3. The 3♣ response to 1NT is otherwise infrequent and not that useful.

Whether this is true of course depends on the rest of your 1NT structure, but I find that a lot of the systems out there miss some hand types. One that is frequently lost seems to be the three-suited and semi-three-suited patterns (like 1444 or 1453). Sure you can stayman or puppet stayman, but just because you have no 8-card major fit doesn't make 3NT the best contract! It can easily be right to play elsewhere when opener has weakness opposite the shortage.

It seems that your main point is not that 5-card Stayman is a bad use of 3♣, but that a shortness relay is much better. This may be, but playing, say 4m instead of 3NT may not be such a successful strategy; don't forget that they haven't led the short suit yet. The prospective gain in IMPs is obviously much larger in 3NT, and in matchpoints you will be with any pairs in the field who don't have this shortness ask, so if 3NT makes you will have an average if you are in it.

[aguahombre]

The improvement, that 3D doesn't guarantee a 4-card major prevents some leakage.

Using it with a 4-card major only when holding exactly 3 of the other reduces the off-lead double situations you are concerned with.

Using it as a semi-automatic bid enroute to 3NT allows us to remove a direct 3NT with a wide open Major, since Responder will be wide open in the Majors (alertable). We enjoy the expression on the opponents' faces when it goes:

1N-3N
4N-5m
P

[WGF_Flame]

awm, on 2013-November-25, 19:16, said:

1. It's important to find our 5-3 major suit fits.

When holding two balanced hands, the statistics favor 3NT over 4M on a 5-3 fit. This becomes even more true if opener can use judgement as to whether 1M or 1NT should be opened on 5332 hands in range. The 3NT contract also tends to be more difficult to defend, such that double-dummy may understate the 3NT advantage.

Of course, responder doesn't have to be balanced (for example 3451 or 3361 hands) but in these cases pinpointing the shortness will help more than just checking for 5-3 major fits -- even without an eight-card major fit it can easily be right to avoid 3NT on such a hand (playing a 7-card major fit or 5/6m when opener is weak opposite the shortage).

I would like to see numbers on this, anyway obviously there are hands who wish to play in 4M opposite 5 card major and not opposite 4 card major.
and the responder can use puppet for those.

In general i think the reasson for the popularity is not that its an amazing system. its because its very simple and definetly not bad.

[MickyB]

I prefer 1N:2N as puppet, as Meckwell play, partly because then oppo don't get to double 3C as well as 3M. When I don't want to find a 5-3 fit, I'll normally use Stayman, because 1N:2N, 3M:3N gives away a lot about opener's hand. I definitely think puppet comes out ahead when responder is 4-3 majors, although obviously we will sometimes lose when the other table bids 1N:2C, 2D:3N and they are scared off leading the other major.

I pretty much only bid 1N:2N without a four-card major if I've got a low doubleton, I'd be surprised if 3N rated to outplay 4M on such hands. Another argument in favour of Puppet not mentioned in the OP is that it is also good for balanced slam-tries.

[broze]

MickyB, on 2013-November-26, 05:43, said:

When I don't want to find a 5-3 fit, I'll normally use Stayman, because 1N:2N, 3M:3N gives away a lot about opener's hand.

Aren't the losses pretty negligible in that auction though because you'll be using Stayman anyway? As I see it the relevant case are (using 2N as puppet):

A) Loss: 1N:2N, 3M:3N (vs when the 2N bidder would have used Stayman - i.e. the knowledge leaked is the extra Major card in opener's hand) Rare. Substantial leak.
B) Loss: 1N:2N, 3M:4M (vs 1N:2C, 2M:4M) Again only extra Major cark leaked. Rare.
C) Loss: 1N:2N, 3D:3M, 3NT (vs 1N:2C, 2D:3N) Opps will not lead the Major responder has shown. Common. The biggest leak, but only relevant to the lead.
D) Gain: 1N:2N, 3D:3M, 3NT (vs 1N:2C, 2M,3N) Common. Huge gain, both on the lead and in the play.

A and B are both pretty rare cases I imagine (even more so given that opener will sometime choose to avoid 1N with the 5cM for one reason or another). Whether using Stayman on hands where "you don't want to find a 5-3 fit" is a net gain seems to depend on the size of D above. I can't say for sure but I imagine it comes out on top. Unless I am missing something.

This post has been edited by broze: 2013-November-26, 07:03

[MickyB]

broze, on 2013-November-26, 06:35, said:

Aren't the losses pretty negligible in that auction though because you'll be using Stayman anyway? As I see it the relevant case are (using 2N as puppet):

A) Loss: 1N:2N, 3M:3N (vs when the 2N bidder would have used Stayman - i.e. the knowledge leaked is the extra Major card in opener's hand) Rare. Small leak.

It's not the extra card that's the biggest issue.

If a 1NT opener shows 4-5 spades, he has 2-5 cards in each minor. If a 1NT opener shows 5 spades, he has 2-3 cards in each minor.

[broze]

MickyB, on 2013-November-26, 06:47, said:

It's not the extra card that's the biggest issue.

If a 1NT opener shows 4-5 spades, he has 2-5 cards in each minor. If a 1NT opener shows 5 spades, he has 2-3 cards in each minor.

Yes - good point. So the leak is fairly substantial. Still it wouldn't surprise me to find that it was fairly negligible due to low frequency. I guess this depends on your style - seems difficult to quantify.

[rhm]

I sympathise with your mood.

I play puppet Stayman over 2NT in most partnerships, but not over 1NT. I play it, but I am not convinced.
Puppet Stayman was originally designed not to reveal opener's holding. It was not designed finding 5-3 major suit fit. This was more or less an unintended side effect.
Nowadays puppet Stayman is used to reveal opener's major suit distribution almost indiscriminately.
Of course sometimes the 5-3 major plays better giving the puppet players a field day. If 3NT would have been better it is often not even realized or shoulders are shrugged.
What I would like to see is statistical evidence over a large number of deals a la Pavlicek from top level play, where in one room puppet was used and in the other not

I have still to see a comprehensive, yet simple set of responses where responder can show a singleton or void to the balanced hand below 3NT.
These hands are rare, but depending on opener's holding in the short suit, 3NT can be the only game or slam in a minor laydown while 3NT is down off the top.
These hands are infrequent, but annoying when they come up because there is little you can do as responder.

Rainer Herrmann

[Vampyr]

rhm, on 2013-November-26, 07:42, said:

I have still to see a comprehensive, yet simple set of responses where responder can show a singleton or void to the balanced hand below 3NT.
These hands are rare, but depending on opener's holding in the short suit, 3NT can be the only game or slam in a minor laydown while 3NT is down off the top.
These hands are infrequent, but annoying when they come up because there is little you can do as responder.

I believe that showing such a method was the point of the OP. Hopefully he will enlighten us soon.

[Cyberyeti]

rhm, on 2013-November-26, 07:42, said:

I have still to see a comprehensive, yet simple set of responses where responder can show a singleton or void to the balanced hand below 3NT.
These hands are rare, but depending on opener's holding in the short suit, 3NT can be the only game or slam in a minor laydown while 3NT is down off the top.
These hands are infrequent, but annoying when they come up because there is little you can do as responder.

Rainer Herrmann

I devised one which does this but it's not simple. Broadly speaking, if you play the dual stayman approach (and we use a 5M inquiry which is not puppet), direct 3♦/♥/♠ and even 3N can be available for this.

[antonylee]

MickyB, on 2013-November-26, 05:43, said:

I pretty much only bid 1N:2N without a four-card major if I've got a low doubleton, I'd be surprised if 3N rated to outplay 4M on such hands. Another argument in favour of Puppet not mentioned in the OP is that it is also good for balanced slam-tries.

A nice writeup of the usefulness of Puppet for balanced slam tries (and for other stuff) by Justin Lall here: http://justinlall.co...er-a-1n-opener/

Also, while 1N-3♣-3♦-3M-3N reveals responder's 4cM, if you were not playing Puppet you'd likely have bid something like 1N-2♣-2♦ or 2oM-3N, so the opponents still had a lead-director available for clubs (although it's marginally riskier for them to double 2♣ rather than 3♣), and instead of having the opportunity of doubling 3M for the lead, they have a bit more information about opener's hand (whether he has 4 cards in the other major or not - this assumes you play that 3♦ does not say anything about the possession of a 4cM). It's easy to be unhappy when a lead director against 3M costs us the contract or some overtricks but sometimes you don't even notice that the info leakage via regular Stayman actually cost you something.

[JLOGIC]

I think point #1 is the most important, if you believe it then you should use 3C as something else obviously.

Adam brings up using judgement in when to open 1N with 5332, a counterpoint is you can use judgement on when to puppet. I am on board with going 1N 3N on a lot of hands that might have a major suit fit if looks right, but I think I can also use my judgement to find a 5-3 fit on hands where that will usually be correct if my system permits it. I also think a lot of people like me open 1N whenever it seems in range for 1N and cannot use judgement in whether or not to do so.

[glen]

JLOGIC, on 2013-November-26, 15:48, said:

I think point #1 is the most important ...

It is the most important, but is actually:

1. It's important to find some of our 5-3 major suit fits.

The some includes some where both hands are balanced/semi-balanced. The actual set of hands (where the 5-3 fit is better) is quite complex, and discussion is likely best for another thread.

[mikestar13]

JLOGIC, on 2013-November-26, 15:48, said:

I think point #1 is the most important, if you believe it then you should use 3C as something else obviously.

Adam brings up using judgement in when to open 1N with 5332, a counterpoint is you can use judgement on when to puppet. I am on board with going 1N 3N on a lot of hands that might have a major suit fit if looks right, but I think I can also use my judgement to find a 5-3 fit on hands where that will usually be correct if my system permits it. I also think a lot of people like me open 1N whenever it seems in range for 1N and cannot use judgement in whether or not to do so.

I've been following this idea more and more as I get older: use judgement, but when in doubt, reach for Hamman's Hammer.

[awm]

Mostly I was just curious why a method which I think is not particularly good was rapidly increasing in popularity. The feeling I'm getting from reading these posts is:

1. People don't agree with me on the relative merits of 3NT vs. 4M on two balanced hands (a sim could help, but I'm not going to run one). Also, a lot of people play systems where they have to open 1NT on 5332 hands in range and cannot use judgement in that aspect of the auction.

2. The idea of having responder bid his major (instead of the other major) when playing weak notrump is interesting and might reduce the disadvantage of the lead-directional double, but this is not the way I'm seeing puppet played at most tables. Otherwise people seem to feel that the knowledge of whether opener has a four-card major is more significant than I believe it to be and/or that the danger of the lead directional double is less.

3. A lot of people play methods where they can't properly show shortage on a lot of hands, and in some cases seem to have never really considered the problem... if you can't show shortage anyway at least you can find your 5-3 major fits when responder is (say) 3451. Also, people seem to come to this method from a base where 3♣ was "both minors weak" which is basically a terrible usage, so almost anything is an improvement.

Interesting...

Anyway, one person asked what I play, and while our current methods are rather complicated the basic idea here is simple and fits well with four-way transfers:

3m = 4+ in the bid minor, short in the other minor, usually 3-4 in each major (we don't have another bid for 3271)
3M = 4 in the bid major, short in the other major

At this point we can scramble for a 4-4 major-suit fit over 3m, with most continuations being natural. 3NT is always an offer to play (but opener should remove responder's 3NT with a weak holding in the short suit, and responder can pull opener's 3NT with serious interest in slam). Opener bidding the short suit shows a "perfecto" with nothing wasted; responder bidding the short suit asks if opener has such a "perfecto" or else shows a void if opener has already denied it. This lets us find some thin slams when the hands are fitting just right, while bidding to the right game contract virtually all the time. These bids are relatively frequent and have been substantial winners when they come up (in particular avoiding bad 3NT contracts to play an excellent 5m or 4-3 major fit, and finding some slams that are otherwise tough to bid).

For those who were using 3♣ as "weak both minors" that hand can be bid via 2NT (transfer to diamonds) and pass opener's 3♣ (I don't like diamonds). Transfer to clubs followed by 3♦ can take over as "GF both minors" (since clubs with short diamonds responds an initial 3♣). With major-suit shortness but lacking four cards in the other major, transfer to the long minor and then show the shortage (note that this could be only five cards in the minor transferred to if 3145 or so... we find that opener's holding opposite the short suit is typically more important than distinguishing 3145 vs. 3136 immediately on these hands).

[Zelandakh]

MickyB, on 2013-November-26, 05:43, said:

I prefer 1N:2N as puppet

Another option that can be used is to play 2♣ as Puppet Stayman with follow-ups that handle the 4-4 major fits. This tends to reveal a lot about Responder's shape but is not as bad as the common form of Puppet in revealing Opener's (mainly since 2♦ is simply no 5M rather than showing a 4 card major). Notice that you can also use the "denial bid" form of Puppet follow-ups to your advantage in handling the hands that Adam is highlighting. So, for example, 1NT - 2♣; 2♦ - 2♥; 2NT - 3♥ can show a hand with heart shortage and 1NT - 2♣; 2♦ - 2♠; 2NT - 3♠ spade shortage. Minor suit shortages can be shown with, for example, 1NT - 3M, or for stronger hands 1NT - 2♣; 2♦ - 3♣(GF both majors); 3♦ - 4m.

In any case, my sentiments for using Puppet lie along the lines of Justin's. You can have Opener show judgement in whether to open 1M or 1NT or have Opener open 1NT with practically every 5M332 in range and then have Responder use judgement in when to look for a 5-3 fit. The former has a negative impact on the rest of the system whereas the latter has only a small impact on the 1NT structure as a whole and everything else is easier. So the question is why would you not want to use such a method, assuming you have the other important hand types already accounted for?

[Jinksy]

MickyB, on 2013-November-26, 05:43, said:

I prefer 1N:2N as puppet, as Meckwell play, partly because then oppo don't get to double 3C as well as 3M. When I don't want to find a 5-3 fit, I'll normally use Stayman, because 1N:2N, 3M:3N gives away a lot about opener's hand. I definitely think puppet comes out ahead when responder is 4-3 majors, although obviously we will sometimes lose when the other table bids 1N:2C, 2D:3N and they are scared off leading the other major.

Bird and Anthis' book looks in depth at this auction (1N:2C, 2D:3N), and concludes that you should treat it basically as 1N:3N - ie major leads dominate (DD) almost as much as they do in the latter.

[Siegmund]

Glad to see awm posted this. It's an idea I wish got more airing.

Re the 4M vs 3NT with game values issue, the results of the sim depend a bit on one's assumptions, but if you include most hands with a 3-card major for responder, it is very close to a dead tie at MPs, and a very small win for the major at IMPs. Did this sim two different ways a year or so ago and never got round to writing an article about it.

It's certainly true that the times where the 5-3 fit gains are going to be times when responder is (for instance) 3415 -- but a lot of those are hands where you would gain from finding the much-more-frequent minor fit. Puppet is aimed at such a rare target I am astonished it ever became popular. I will take most any other meaning for 1NT-3C over Puppet.

Over 2NT, the worse problem is the trouble Puppet causes for 4-5 / 5-4 major hands for responder, unless you make some big changes to the structure.

[Vampyr]

Siegmund, on 2013-November-30, 00:37, said:

Over 2NT, the worse problem is the trouble Puppet causes for 4-5 / 5-4 major hands for responder, unless you make some big changes to the structure.

With 4-5 you can transfer to hearts and then bid spades; with 5-4 there is no problem so long as you interchange the 3NT and 3♥ rebids -- so 3♥ is no 4- or 5-card major and 3NT is a hand with 5 hearts. It's not a big change.