Jacoby 2NT

The use of a 2NT response to a 1 major opening has evolved over the years. Typically it would have been somewhat natural 60 or so years ago, either 10-12 and invitational in Acol style, or stronger and game-forcing in old-fashioned American style. If your system includes 5-card majors with 1NT forcing, you don’t need the bid in such a natural mode (you can always respond 1NT and then follow-up with 2NT to show an invitational hand). Jacoby invented a new use for it – to show a game-forcing raise of the major suit, including 4-card support, with the idea of facilitating slam investigation.

The original rebids by the opener showed singletons at the three level and voids at the four level. Using all those bids just to show shortage is obviously very inefficient, and nobody plays that way these days. The modern version taught to beginners uses 3-level bids to show shortages, while rebidding a new suit at the 4-level indicates a good second suit, a source of tricks. That’s clearly better, but still not great. It does have the virtue of being easy to remember, but experts (who are prepared to pay the price in memory strain) usually go for more information. The version described by Larry Cohen (at larryco.com) is fairly typical, I think:
After 1 major – 2NT;

• 4 major = absolutely awful minimum (always 5-3-3-2)
  
• 3C = any other minimum (could be 5-3-3-2, but at least some redeeming feature)
  
• 3D = non-minimum, with a side singleton or void somewhere
  
• 3H = non-minimum, any 5-4-2-2 distribution
  
• 3S = non-minimum, 6+ trumps (no singletons or voids)
  
• 3NT = non-min., 5-3-3-2 (can be defined range eg 18-19)
  
• 4 new suit = decent 5-card side suit (should have ace or king)
  

After opener’s 3-level rebid, responder can bid the next step as a relay, and there are step 2nd rebids by the opener to locate shortages or second suits. Larry says that the memory strain isn’t too bad, but I don’t think I want to take it on.

At Bobby Knows Bridge (www.bobbybridge.com, a site well worth visiting) there is something a bit different. This is called Jacoby-Roman 2NT, and it concentrates on discovering opener’s shape with mostly natural bidding. In addition, the recommendation is to have the original 2NT response promise only 3+ cards in support of opener’s major. While there is therefore a known fit, there is also the possibility of discovering a 4-4 fit in another suit and playing there.
After 1 major – 2NT; opener has five types of rebid:

• 4 major = minimum with 6+cards, no side suit or singleton
  
• 4x = splinter with 6+card suit, no side suit
  
• 3NT = minimum with 5-3-3-2, not forcing
  
• 3 major = 15+ 5-3-3-2 or 6-3-2-2 or 7-2-2-2
  
• 3x = natural side suit
  

Of course, bidding a side suit will be the commonest rebid for the opener. After that, responder in turn has five possible continuations:

• 3NT = good stoppers, only 3-card support for the major. Opener will often pass, and should have a shortage or substantial extras to go on.
  
• 4 major = minimum response with 4-card support.
  
• 4x = 4-card support for opener’s second suit, suggesting slam.
  
• Cheapest bid other than 3NT = relay.
  
• Not the cheapest bid = control cue-bid with 4+card support for the major.
  

The relay bid asks opener to further clarify his shape. Opener continues in a natural style:

• 3NT = 5-4-2-2, not too much extra since 3NT can be passed
  
• repeating the major = 6-4 shape
  
• repeating the side suit = 5-5 shape
  
• new suit = natural, probably 3 cards in a 5-4-3-1 distribution
  

This all looks quite attractive to me. Memory strain is negligible, since all the bidding is essentially natural – there is only one artificial bid to recognize, the relay asking for more shape info. And responding 2NT on 3-card support makes sense to me also, since it increases the frequency of occurrence – the usual Jacoby 2NT really doesn’t come up very often – and allows you the inference that a 2-over-1 response followed by raising opener’s major will show a decent 5-card suit as well as 3-card support. (There is no longer any need to fudge the initial response on a poor suit).

I realize it is far from perfect. There are a number of distributions where the full sequence will take you past 3NT when ideally you might want to stop there. But that isn’t disastrous, because 4major should always be playable. I would be interested to give it a try.