Research

The strong jump shift

A study of 3,075 deals that kept the strong jump shift at 19 or more points and, playing 2/1 Game Force, limited it to suits whose ordinary response would have been at the one level.

The question

A jump shift is a response in a new suit one level higher than necessary, for example 1♦ – 2♠. Our sources agree it is a strong bid in an uncontested auction, but not on the details. One asks for 19 or more points. Another accepts about 16 and a good suit, and treats a jump to three of a lower suit as invitational. A third says only that the jump shift is strong. A second question was whether a jump shift helps at all when the partnership plays 2/1 Game Force, where an ordinary two-level response already forces to game.

The deals

The study used every stored deal with an uncontested one-of-a-suit opening in which an unpassed responder held 16 or more points counting length and a side suit of five cards or more: 3,075 deals. BidMaestro bid each deal from the opening onwards under each candidate rule, with all four players using the same system.

The candidates

Nine rules were tested. Eight were a jump shift on 16, 17, 18 or 19 points, each in two forms (any strong hand, or only a hand with one long suit), with the jump made over every opening. The ninth had no jump shift at all. They were compared with the rule the app played at the time: 19 points, over minor-suit openings only.

The nine rules were measured on 2 September 2026. The comparisons of 1♥ with 1♠ and of the minor-suit jump below were measured on 4 September, after other improvements to the app’s bidding, so the two sets of figures cannot be compared directly.

How the auctions were scored

Each final contract was scored on the cards as dealt, played with all four hands visible (Double Dummy), from the point of view of the side that opened. Because one layout of the cards can favour or penalise a rule by chance, 2,305 of the 3,075 deals were also re-dealt 200 times. The opening side’s cards were kept and the opponents’ reshuffled, and the contract was scored against the average of those layouts. The other 770 deals were scored on the dealt cards only. To be adopted, a rule had to score better on both measures by more than its margin of error.

What we found

Lower point requirements, with the jump made over every opening, lost points.
All nine candidates jumped over every opening, majors included, so the study could not tell whether a loss came from the lower requirement or from jumping over a major. Together they scored about 30 to 65 points per changed deal worse than the rule the app played. Even the 19-point requirement lost when it was also used over major-suit openings.
No jump shift at all lost points too.
About 70 points per changed deal. Without the jump shift the app missed 55 slams it had reached before and reached 23 it had missed.
Over 1♥ and 1♠ the results were opposite.
1♥ – 2♠ gained about 270 points per changed deal (23 deals better, 11 worse). 1♠ – 3♥ was worse on 22 deals and better on 7, and its average was within its margin of error.
The minors gave the same result.
With 2/1 Game Force switched on, dropping the 1♦ – 3♣ jump shift gained about 80 points per changed deal (36 better, 5 worse), with seven slams reached and none lost. After the 3♣ jump the auction had often stopped in five of a minor, where the ordinary 2♣ response reached 3NT or 6NT.

What we decided

Over 1♣ the strong jump shift is used in any suit; over 1♦, hearts or spades; over 1♥, spades only; over 1♠, none. Where the ordinary response is already a two-level game force, the jump uses a level and gains nothing. With 2/1 Game Force switched off, the jump shift is used in every suit, because a two-level response then forces for one round only. The 19 points count length, and the suit must be a good one of five cards or more.

Compared as a whole with the rule the app played before, the new rule gained about 170 points per changed deal over 85 deals (59 better, 16 worse), reaching 23 more slams and missing 2. The bid itself is described on the jump shifts page.

Limits of the study

The study measures the jump shift together with BidMaestro’s own later bidding; a partnership with different continuations could reach a different result. 3,075 deals is a small sample.

← Back to research