Transfers over 2NT

[mikeh]

I am late to this thread, so much of what I say has already been expressed. However, I dislike Roland's idea for several reasons.

 

1. There will indeed be hands on which playing in responder's major at the 3-level will be best and these will most often be those on which opener has no fit. Responder's hand will often take NO tricks if he has a 5332 2 count or so, if played in 3N.

 

While these hands may be viewed as low-frequency, they are not zero frequency and I fail to see any offsetting benefit from the 'do not accept the transfer without a fit' approach.

 

2. I like to use the transfer to 3♥ as showing either ♥ or some other hand type (currently some unspecified 4441 with slam interest, but I have used others) and this approach requires that opener's reply to the 3♦ transfer be constrained to 3♥ or 3♠, the latter being a super-accept. This approach of using multi-purpose bids conserves bidding space: which is at a premium after a 2N opening bid. Using 3N over 3♦ to deny a fit prevents this useful approach.

 

3. I use the transfer then raise to show a mild slam try. If responder holds that type of hand (say, 6+ suit and 10 hcp points or so), he must, in Roland's scheme, bid over 3N. Now, he can do this in one of two ways. He can retransfer or he can bid his suit directly.

 

However, bidding his suit directly will sometimes wrong-side the contract. While on many hands, the choice of declarer will be irrelevant, the reality is (I think) that, when it does matter, it is more often correct to have the 2N bidder be declarer than responder. Given that we are discussing slam hands, the cost of wrong-siding even a minority of contracts will be high.

 

We can avoid that by retransferring but that carries real costs as well: responder will sometimes be two suited, rather than one suited. While transfers work when the second suit is ♦'s or ♥... with both majors, they stumble when the second suit is ♣. Now responder must bid 4♠ to show the ♣ suit. This may not cost often, but consider this: opener has NO room below 5♣ in which to indicate his liking for ♣. 4N must be conserved as 'to play', while 5♣ merely accepts the transfer. Must opener FORCE to slam whenever he likes ♣? It seems to me that this is a complication that we simply do not need. Compare the auctions 2N 3♦ 3♥ 4♣ to the auction 2N 3♦ 3N 4♠, where responder is showing the same hand: which partnership is more likely to have a comfortable auction to the right strain and level?

[mike777]

let's assume 2nt=20-21....I fail to see the argument for superaccept. I mean by this that non super accept is a winner more often, significantly more often.

[1eyedjack]

In my opinion the main reason for superaccepting is to make slam bidding more accurate, rather than, as suggested here, to find thin games opposite a v weak responder. That (finding the thin game) is an added bonus, but really the icing on the cake.

 

I cannot accept that forcing accepting the transfer actually improves the strength of the slam bidding as suggested elsewhere in this thread. Increasing the hand types that can accept should in theory reduce the accuracy of continuations following acceptance.

[awm]

Super-accepting wins in two ways: (1) Partner would pass if you simply accept the transfer, but 4M is making. (2) Partner has a marginal slam try, and the super-accept and immediate cuebid helps to find a slam that might otherwise be missed. Super-accepting loses in one way: (1) Partner would pass if you simply accept the transfer, and 4M is not making.

 

On a related point, suppose we have a hand which isn't worth a game bid (something like 0-4 hcp), and the hand includes a five-card major. Should we normally transfer? I think the answer to this question is a definite yes. Playing in 3M creates more winners (at least one more, sometimes two more) and also creates entries to the weak hand to take various finesses if necessary. Note that 2NT opening hands usually don't have a lot of tricks, and can't make 2NT independently by themselves. This is generally because there is no long suit to run. However 2NT openings do have a lot of cover cards... the issue is often "not enough winners" rather than "too many quick losers." The major suit contract will normally play at least one trick better and often more, even if opener has only doubleton in the suit (note that a big part of this is because of the existence of entries to the weak hand, and so it doesn't necessarily apply when you give responder some scattered values that can serve as entries).

 

So back to super-accepting. How many points do we really need to make 4M with a nine-card fit? We frequently make such hands on 22-23 high card points, especially if the short hand has a possible ruffing value too. Compare the playing strength of the following two hands when partner transfers to spades:

 

AKxx

Ax

KJxx

KQJ

 

Ax

AKxx

KJxx

KQJ

 

I'd expect the first hand to play at least two tricks better, wouldn't you? Give partner a junky hand like Qxxxx xxx xx xxx and game is excellent opposite the first hand and lousy opposite the second. My expectation would be, in spades, ten tricks opposite the first hand and eight opposite the second. By the way, in notrump I expect maybe eight tricks opposite the first hand would be mildly lucky to take six tricks opposite the second.

 

And this is ignoring the slam-bidding benefits of the super-accept.