[bluecalm]

Quote

If they knew Bluecalm's opinion that it sucks, they probably would stop again.

Maybe they have another reason, like reducing variance and playing along with the other table.
No idea, I spent enough time on this one to be confident that at least at imps puppet is not profitable and it's better to just bid 3nt on all those 4-4-3-2, 4-3-3-3 and 5m-3-3-2 hands.
If someone very good had told me "I spent a lot of time analyzing this and my conclusion is that puppet is +EV instead of bashing 3NT" then I would probably reconsider but intuition alone even from someone very good isn't going to convince me as I saw tons of layouts and I don't have selective memory bias on this one (as I saw them consecutively with the goal of determining which is better).

One day when my program is ready which can make reasonable human-like lead on most NT biddings I will just simulate 10k hands and see. For now it have to wait

[JLOGIC]

Who bids puppet on 4333? Obviously it's good to be playing puppet when you are like (43)(15) or 33(16) or something, and on some 4432s I would definitely want to find my 5-3 fit if I could.

Again, what are you going to simulate? There is no doubt that finding your 5-3 fit is good on a certain subset of hands, the bigger question is how frequently do you lose when you look for a 5-3 fit and don't find one vs how often do you gain when you do find one, and how often do you lose by not being able to bid smolen/whatever you play 3C 3D 3M is normally. Good luck simulating that!

And again, there is no chance that meckwell are playing puppet to try and reduce variance, I cannot believe you would actually think that.

[aguahombre]

Jerry Helms, in the Feb Bulletin, attempted to clarify his previous statement about not liking puppet/2NT. Basically, the problem is the 4-5 and 5-4 hands, but an experienced pair, and/or a great pair, can make adjustments to other auctions and include puppet if they have the inclination to do so. It boils down to context. The convention sucks for those who have holes in their structure because of puppet. And is simply a matter of preference for the rest.

[bluecalm]

Quote

Again, what are you going to simulate?

I took some hands which people routinely bid puppet with and look at the possible layouts (100's of them because I am an addict).
My conclusion was that puppet causes more loses (because of better 1st lead) than it ever gains on 4-4-3-2 hands and 5-3-3-2 hands. Also system with puppet is weakier because there is less space for other hand types but obviously this one is difficult to quantify.
Obviously all the muppets and romexex which doesn't allow you to bid 2NT - 3NT aren't even worth considering.

[JLOGIC]

Then perhaps the results of your simulation show that people are bidding puppet on the wrong hand types, not that puppet is a bad convention (btw I do not play puppet over 2N).

[bluecalm]

Btw as to puppet discussion. I have just found out (by reading their cc) that MR play the version of puppet which hides opener shape:

2NT - 3♣
3♦ any without 5M
and following:
3♥ = 4spades or none
3♠ = 4 hearts
3NT = 4-4 in majors

2NT - 3♦
3♥ - 3♠ = any 5-4/5-5/4-5 in majors

This version could be even better for camouflage purposes than normal stayman as defenders often don't know anything about declarer 4card majors (if resp have at least on 4M) if it ends up in 3NT.
I must say I like this idea a lot.

[jallerton]

bluecalm, on 2011-April-10, 05:39, said:

Btw as to puppet discussion. I have just found out (by reading their cc) that MR play the version of puppet which hides opener shape:

2NT - 3♣
3♦ any without 5M
and following:
3♥ = 4spades or none
3♠ = 4 hearts
3NT = 4-4 in majors

2NT - 3♦
3♥ - 3♠ = any 5-4/5-5/4-5 in majors

This version could be even better for camouflage purposes than normal stayman as defenders often don't know anything about declarer 4card majors (if resp have at least on 4M) if it ends up in 3NT.
I must say I like this idea a lot.

Compared with 3♣ 4-card Stayman, this method of bidding over 3♣ will sometimes give away less information, but sometimes more. If Responder is merely seeking a 5-card major opposite and Opener holds 4 spades, that information is given away in the sequence 2NT-3♣-3♦-3♥-3♠-3NT. 4th hand gains the ability to double an extra conventional call too.

However, I am more concerned that you "like a lot" the idea of 2NT-3♦-3♥-3♠ as "any 5-4/5-5/4-5 in majors". Unless their 2NT opening guarantees 3 cards in each major, or they prefer to play seven card fits in 4major to 3NT, they are going to struggle always getting to the normal game contract after this start.

[bluecalm]

I still like normal stayman more but this version seems better than stuff when you give away opener shape on every hand.

[dake50]

With these 2NT structures, would splinters be needed?
If more than game - here's my short to help evaluate tricks/duplication.
Wouldn't the splinter hand just best-guess game/not.
Surely splinters on top of this 2NT structure can be very narrowly defined.

[gnasher]

bluecalm, on 2011-April-10, 15:17, said:

I still like normal stayman more but this version seems better than stuff when you give away opener shape on every hand.

If you don't care about finding a 5-card major in opener's hand, these methods give away nothing about opener's hand, always rightside any game, and don't risk reaching any silly contracts:

3C = both majors (opener responds to Stayman, then over 3D responder bids Smolen)
3D = 4+ hearts
3D-3H-3S = only 4 hearts
3D-3H-other = 5+ hearts
3H = 5 spades
3S = raise to 3NT
3NT = 4 spades, choice of games

The main disadvantage is that you give them a few extra chances to double artificial bids.

[jallerton]

Another (perhaps more serious) disadvantage of that method, Andy, is that you only have the 3♠ puppet and 4-level responses to cover both all minor suit slam tries and all slam tries with 4 spades, not 4 hearts.

I suspect you'd be better off putting the slam tries with 4 spades through your otherwise underused 3♣ bid.

[gnasher]

jallerton, on 2011-April-11, 14:49, said:

I suspect you'd be better off putting the slam tries with 4 spades through your otherwise underused 3♣ bid.

Yes, we do exactly that:
2NT-3♣;3♦-4m = 4 spades and four of the minor
2NT-3♣;3♥-4m = 4 spades and four of the minor
2NT-3♣;3♥-3♠ = slam try with hearts agreed

I didn't mention that part of it (and one or two other twiddly bits) because I was trying to convey the essence of the method, without going into lots of detail.

By the way, the reason that we do this is that you made exactly the same suggestion to me three years ago.

[MrAce]

gnasher, on 2011-April-11, 15:45, said:

Yes, we do exactly that:
2NT-3♣;3♦-4m = 4 spades and four of the minor
2NT-3♣;3♥-4m = 4 spades and four of the minor
2NT-3♣;3♥-3♠ = slam try with hearts agreed

I didn't mention that part of it (and one or two other twiddly bits) because I was trying to convey the essence of the method, without going into lots of detail.

By the way, the reason that we do this is that you made exactly the same suggestion to me three years ago.

Andy, would you mind sending me all the details of this via email ? If you will turkofla@hotmail.com. Thanks in adv regardless u send or not I didn't wanna hijack the thread.

[gnasher]

MrAce, on 2011-April-11, 17:14, said:

I didn't wanna hijack the thread.

It's a bit late to worry about that.

I've emailed you the details.

[MrAce]

gnasher, on 2011-April-12, 01:57, said:

It's a bit late to worry about that.

I've emailed you the details.

Heh, i hear ya. Thanks m8

[MrAce]

Double post

[Poky]

wank, on 2011-March-27, 18:17, said:

is there a better way to do it specifically for inv+ jacoby?

Yes, there is.

Your 'Swedish' structure has many leaks:
1. The cheapest rebid is showing the worst hands,
2. It's asymmetric with a profitable structure when we overcall,
3. Doesn't polarize between weak balanced/unbalanced hands without revealing the shortness.

What do I mean?
1. The 3C rebid spares most space and should be used to show the best hands,
2. Do you really want to bid 3C to show a minimal hand after, say:
(1C) 1S (pass) 2NT
(Dbl) ?
3. Sometimes (and pretty often, I would say) responder has a hand that wants to play game if partner is minimal but unbalanced, without knowing which shortness opener has (to make opponents harder to find the best lead/defence)

All that said, a much better structure should be:
3C = any 17+ (...3D asking, new suit showing shortness, 3M showing 10-11 balanced)
3D = minimal unablanced (...3M/4M to play, step1 asking shortness)
3M = minimal balanced (...3NT to play, cuebid trying for slam)
3oM/4C/4D = shortness, 14-16
3NT = balanced 14-16 with 5M
4M = 6M(322), 14-16

After 3C 17+, 3D relays into the same structure, with one addition:
3M = 6M(322)
3NT = balanced with 5M

And besides all, it is very intuitive, easy to remember and use.