The 6-2 Fit In A Major

[gurgistan]

During the course of bidding, I as responder discover that partner has six-cards in his 1M opener eg a typical 2/1 Game Force sequence such as 1M 2m 2M.

 

I have enough hcp for game. I also have 2 cards in opener's major.

 

The choice seems to be either 3N or 4M.

 

Is it simply a case of we have an 8-card fit and therefore it must be 4M or do I take into account other factors?

 

Other factors, I take to be:

 

1. Our overall hcp count

2. Our overall distributive count (hcp + responder shortage)

3. Whether I have stoppers in the unbid suits?

 

All relevant comments much appreciated.

[kenrexford]

In a 2/1 GF sequence, you have the option of bidding 2NT to get more info before committing.

[Finch]

...Or of raising partner's major to the 3-level, and seeing if he wants to bid 3NT.

It's worth discussing whether 2NT should promise, or deny, doubleton support.

[gurgistan]

...Or of raising partner's major to the 3-level, and seeing if he wants to bid 3NT.

It's worth discussing whether 2NT should promise, or deny, doubleton support.

 

This is a good point. Implicit in this suggestion seems to be the thought of Opener's trump quality.

 

If 2N denies 2-card support then with 2-cards I must go to 3M.

 

If 2N promises 2-card support then without 2-cards I must bid 2N.

 

My gut instinct is that 2N should deny 2-card support.

[awm]

This seems like a good question for some simulations. The basic point is:

 

(1) If we have a 6-2 fit in a major, which will usually play better, 4M or 3NT?

 

(2) Which additional factors have the most impact on this choice?

 

My impression is that the 6-2 fit is usually (but not always) better. Factors I would use in a decision include that any ruffing value (singleton or small doubleton) in the 2-card hand points towards 4M, slow cards in the side suits (i.e. queens, jacks rather than aces) point towards 3NT, a solid major suit points towards 3NT.

[Stephen Tu]

During the course of bidding, I as responder discover that partner has six-cards in his 1M opener eg a typical 2/1 Game Force sequence such as 1M 2m 2M.

 

You should learn that for large portion of better 2/1 GF players, 1M 2m 2M does NOT promise six cards in the major; it is a catchall bid for hands not suitable for something more descriptive.

 

There are exceptions like people who play "Bergen" style 2/1 where 2M does promise 6 and use something else as a catchall (e.g. 2nt).

[han]

I did a quick simulation with these conditions:

 

Both north and south have 12-14 HCP. North has exactly 6 spades and a 6322 distribution, south has exactly 2 spades and a 4432 distribution. South has at least two of the top 5 cards in each suit, and jack-ten does not count.

 

Out of 164 hands, 3NT in the south made 153 times (93%), 4S in the north made 135 times (82%). Don't ask me why I chose 164 hands, it was a misclick.

[han]

I redid the analysis with a larger number of hands (400). I now obtained that 3NT makes 91% of the time and 4S 81% of the time.

 

These results are double dummy, and I don't know how those compare to the real life results for these hands.

 

Another simulation: Suppose we have xx KQxx KJxx AJx. Our partner opens 1S, we relay, and we find partner holds 12-14 HCP and some 6322 distribution. Should we play 3NT from our side or 4S from his?

 

Double dummy 3NT makes 95% of the time and 4S 85% of the time, a similar difference. If we add in that our hand is the unknown, I think that 3NT is the better call.

[aguahombre]

Don't ask me why I chose 164 hands, it was a misclick.

 

That is how I make most of my choices.

[awm]

I think Han's simulation is sort of expected though; he assumed we have two honors in each side suit (rather than one), that opener has no singleton (opposite 6331 the spade contract might look better), and his fixed example hand has some slow cards and no spade honor. This seems like pretty much the "notrumpiest" hand with a 6-2 fit that we could construct. I'm more curious what the results would be if we specified something like 2(443) opposite 6(322) or 6(331) with 12-14 hcp in each hand, no restrictions on honor location.

[bluecalm]

This seems like a good question for some simulations. The basic point is:

 

(1) If we have a 6-2 fit in a major, which will usually play better, 4M or 3NT?

 

(2) Which additional factors have the most impact on this choice?

 

Richard Pavlicek made this analysis on real life vugraph hands some time ago.

 

http://www.rpbridge.net/rpme.htm

http://www.rpbridge.net/9x12.htm

 

I did a quick simulation with these conditions:

 

Very interesting. I will try to replicate once I am home. Those results surprise me.

[Finch]

I think Han's simulation is sort of expected though; he assumed we have two honors in each side suit (rather than one), that opener has no singleton (opposite 6331 the spade contract might look better), and his fixed example hand has some slow cards and no spade honor. This seems like pretty much the "notrumpiest" hand with a 6-2 fit that we could construct. I'm more curious what the results would be if we specified something like 2(443) opposite 6(322) or 6(331) with 12-14 hcp in each hand, no restrictions on honor location.

 

Yeah, I think you've defined your conditions to make 3NT better. What you want (and often have) is a way of offering the choice, rather than having to guess which contract to play.

[fromageGB]

You should learn that for large portion of better 2/1 GF players, 1M 2m 2M does NOT promise six cards in the major; it is a catchall bid for hands not suitable for something more descriptive.

 

But if we are talking of a balanced responder 2(443) then he could easily be playing that 2♣ is a 2-way GF, either a long club suit or balanced, with a 2♦ inquiry. When opener skips the inquiry, I would expect 2M shows a weak hand with 6. So a perfectly valid deduction even if you are playing 2M catchall over any other response.

 

Is it normal to play 4 card 2 over 1 without an artificial sequence?

 

But I would add that when responder has a normal 2 over 1 with a 5 card suit, then the odds must very heavily swing to be better in 4S than 3NT.

[aguahombre]

But I would add that when responder has a normal 2 over 1 with a 5 card suit, then the odds must very heavily swing to be better in 4S than 3NT.

 

From the threads I have read here on 2♣ over 1M, I don't think we can use "normal" when referring to 2♣ with five.

 

Let's just call it our old-fashioned expectation ---allowed because J2N is not in our bag. Nevertheless, it would seem that IF responder does have 5, then the 6-2 fit in the major is odds on to be better than NT.

[han]

I think Han's simulation is sort of expected though; he assumed we have two honors in each side suit (rather than one), that opener has no singleton (opposite 6331 the spade contract might look better), and his fixed example hand has some slow cards and no spade honor. This seems like pretty much the "notrumpiest" hand with a 6-2 fit that we could construct. I'm more curious what the results would be if we specified something like 2(443) opposite 6(322) or 6(331) with 12-14 hcp in each hand, no restrictions on honor location.

 

Why is that interesting at all? Presumably at the table we can look at our hand, and base our bid on the hand that we hold? At the table we will also often know something about the shape of partner's hand, depending on our agreements. In the example I gave, I will know that partner is 6322 and not 6331.

 

I'm not saying you should refrain from your simulation, but I don't see what can be learned from it.

 

Yeah, I think you've defined your conditions to make 3NT better. What you want (and often have) is a way of offering the choice, rather than having to guess which contract to play.

 

Yes, obviously I have. I have given an example of (fairly specific, but not highly unusual) conditions where I thought that it would be right to pick 3NT over 4M.

 

Often it is very hard to determine which game is better, and if we offer the choice to partner then even partner will often at best be making an educated guess. Before we decide to offer, we should determine whether our hand is suitable to:

 

- Bid 3NT and not offer a choice to partner.

- Bid 4S and not offer a choice to partner.

or

- Offer a choice to partner.

 

If we offer the choice to partner on all balanced hands with a 6-2 fit, partner will have no way to make a good guess. Holding the hand I gave in my simulation, I think it is better not to offer a choice to partner, with this hand it is very likely that 3NT from our side is the best game.

[Finch]

Yes, I agree with this. Perhaps the way you phrased it was confusing.

[awm]

I think there are a couple interesting issues to resolve here.

 

Sometimes you have an auction where you don't know much about partner's shape and have to make a decision. The typical example in standard bidding is 1♠-1NT-3♠. Suppose that you have doubleton spade and enough for game after this auction. Should you bid 4♠ or 3NT? How do you decide?

 

Of course, sometimes you have more space to find things out. In these cases you will often have to prioritize information. For example, you can find out if partner has shortness (or no shortness). You can find out more about his suit quality. You can find out if he has a concentration of values (if any) outside his suit. But you can't usually find out all three of these. So we could ask which of the three is most likely to help you figure out the right contract.

 

Pavlicek's page seems to indicate that top experts choose 3NT a little more often than they should on these hands. The decision is apparently non-trivial even for really good players.

[gwnn]

Sometimes you have an auction where you don't know much about partner's shape and have to make a decision. The typical example in standard bidding is 1♠-1NT-3♠. Suppose that you have doubleton spade and enough for game after this auction. Should you bid 4♠ or 3NT? How do you decide?

4♠. :)

[aguahombre]

Yep. Can relax while pard plays it --save energy for next one. Also, if I am wrong in 3NT, we get a bad board AND I am a hand-hog. Nothing for me at home, later. Edited March 13, 2011 by aguahombre

[han]

4♠. :)

 

I agree.

[Free]

Han, why are you limiting responder's hand to 4432? Isn't a 5332 applicable as well?