Over a period of time, I came to question many of the conventions used as defenses to opponents' 1NT openings. At some point the conclusion dawned on me that if the defense is to maximise the likelihood of competing with all the possible single suit and two suiter hands that potentially arise, different instruments are needed for the two seats, direct and balancing. Players are, however, naturally resistant to learning different conventions for different positions at the table. Eventually I noticed that by inverting some of the bids of the popular DONT defense, this one convention, suitably adapted, serves adequately for both positions and thereby minimise what needs to be memorised.

Inverted DONT, called iDONT, allows three of the natural single-suited overcalls (2, 2 and 2) to be shown immediately following RHO's strong 1NT opening, and in contrast to regular DONT, defers most of the overcaller's two-suiters until the next round of bidding. The purpose for this inversion is to take advantage of the fact that single-suited overcalls will occur about three times more often than two-suiters¹. For that reason, it is desirable to be able to have their bids available in the direct seat, so as to be able to produce the most frequent disturbance of opponents' Stayman and transfer sequences. Here's the math:

Begin by making the simplifying assumptions that, firstly, the only hands suitable for disturbing are either single suiters with a 6+ card suit, or two-suited (5-4+) hands. And secondly, that the natural single-suit overcalls occur exactly 3x more frequently than the 5-4+ two-suiters. Then we can assign the number 3 to be the relative frequency of single-suited overcalls, and the number 1 to be the relative frequency of the two-suiters. We then find that there are exactly (4 naturals x 3) + (6 two-suiters x 1) =18 theoretical possibilities for disturbing the opponents 1NT opening. However, since a natural 2 disturbs neither Stayman nor Jacoby transfers,² we must eliminate one of the natural overcalls (2) from the analysis, which means we have only (3 naturals x 3) + (6 two-suiters x 1) = 15 actual disturbing possibilities.

Ordinary DONT allows us to get in immediately with a disturbing bid holding any of the 6 possible two suiters, and with one of the single suit naturals (spades). So ordinary DONT will get in (6 x 1) + (3 x 1) = 9 of the possible 15 opportunities. However, since DONT’s 2 does not disturb anything (it is partly natural and yet includes 3 of the possible 6 two-suiters), the equation actually reduces to (3 two-suiters x 1) + (1 natural x 3) = 6 of the possible 15 opportunities, i.e., 40% of the possible opportunities.

Landy, a simple convention in which double is for penalty, 2 shows the majors, 2NT the minors, and 2, 2, and 2 are natural, will get in immediately (3 naturals x 3) + (2 two-suiters x 1) = 11 of the possible 15 times, or about 73% of the time, way better (33%) than DONT in the direct seat. Notice that the Landy 2, unlike the DONT 2 does disturb the opponents, since it is not a natural 2. It makes no sense to use Stayman or Jacoby transfers after a Landy 2 that promises both majors.

Now let's create a partially inverted DONT, by reversing the level of bidding where we show most of our two-suiters and single-suiters. In ordinary DONT we call our two-suiters immediately, and postpone showing the single suiters till after a puppet double. In inverted iDONT we pretty well reverse this, showing most of our single-suiters immediately, and defering most of our two-suiters until after the puppet double. Take a look at this inverted DONT for the direct seat, shown in detail in the table below. iDONT has three natural disturbing overcalls, 2, 2 and 2, and after the puppet double to 2 can show half of the six possible two-suiters. (The other three are shown directly through the partially natural 2 just as in regular DONT.) However, neither iDONT’s double nor its 2 bid showing clubs and a higher suit have any serious disturbing effect. Therefore iDONT actually disturbs the opponents (3 naturals x 3 = 9) + (1 two-suiter x 1) = 10 of the 15 potential opportunities, or about 67% of the time. In other words, iDONT is 27% better than ordinary DONT in the direct seat, and only slightly below the 73% disturbing effect of Landy.

If disturbing the opponents were all that competing against a 1NT opener were about in the direct seat, no frills Landy would win the day, especially as it adds the penalty double which though rare, does occasionally come up and is also a restraint on opponents who might be tempted to shade their “strong” 1NT down to 14+.  (Of course if the opponents are using a weak NT, Landy in the direct seat wins hands down which is why it is the most popular defense in Europe where a 12-14 NT is the most common range.)

That aside, iDONT is a true competitive innovation. Not only does it allow for all six of the two suiters (three of them delayed till after the puppet to 2), it also permits two ways to show a minor two-suiter: 2, showing clubs + a higher suit, and also 2NT to show a more distributional minor two-suiter. Moreover, iDONT has a deferred 2 relay bid (after the double-->2 sequence) available to overcaller to request a second puppet 2NT from advancer, which then permits overcaller with a three-suiter to rebid the suit below her short suit, thereby getting in another occasional possibility for competing. In addition to all this, there is yet a second, overcaller-deferred 2NT which is free for any use the partnership might find for it.

There is a somewhat surprising conclusion from this analysis: ordinary DONT, despite its name, is actually a relatively inefficient disturbing instrument for the direct seat. On the other hand, in the balancing seat where there is no longer any need for disturbing the opponents, DONT has its attractions, mainly in allowing the overcalling side to compete at the two level in all 18 possibilities. In the direct seat, however, I recommend that DONT aficionado’s convert to iDONT, and for other’s to consider Landy whose simplicity belies its power to disturb.

Jock Millenson, 24 April 2006
revised 12 Nov 2006

---