For each question, BidMaestro bids a large set of deals under every candidate rule. It scores each auction against the best result the cards allow, found by playing the deal with all four hands visible (Double Dummy). The rule with the best score is the one the app plays. Each article gives the question, the deals, the candidate rules, the results and the decision.
How to read the figures. A changed deal is one where two rules produce different auctions. Deals where the auctions are the same are left out, so each per-deal figure is an average over the changed deals only. Each average has a margin of error, shown as ±. A difference smaller than its margin of error is treated as no difference and does not change the app.
- Which conventions to learn first
- A study of 150,000 deals ranking the optional conventions by how much each improves the results: RKC Blackwood first by a wide margin; for bidding to game, negative doubles, support doubles, New Minor Forcing and 2/1 Game Force.
- When to ask for keycards, and when to bid the slam
- A study of 100,550 deals: ask for keycards when the partnership knows of 29 points, and bid the slam with one keycard missing when it knows of 27.
- Counting points for a 1NT or 2NT opening
- A study of 150,000 deals: adding a point for a five-card suit, or not counting short queens and jacks, did not improve on plain high-card points.
- The strong jump shift: how strong, and over which openings?
- A study of 3,075 deals: playing 2/1 Game Force, make the strong jump shift only in a suit whose ordinary response would have been at the one level.
- Counting partner’s likely hand rather than the minimum promised
- A study of 100,550 deals: with a major agreed, ask for keycards when your points plus the points partner’s calls suggest reach 30 and the expected losers between the two hands are 12 or fewer.
- Bidding 6NT on the hand partner usually holds
- Two sets of 200,000 deals: about to settle in 3NT after a one-of-a-suit opening, bid 6NT when your points plus the strength partner’s bidding usually shows reach 31.
- Controls and losers beat points below game
- A study of 100,550 deals: a control bid pays when both hands hold the controls and the losers are few. The rule is not yet in the app.
- Showing a control after partner’s splinter
- A study of 2,825 deals: after opener’s splinter, show a control below game with 13 or more support points.
- Doubling their four-level bid after our redouble
- A study of 200,000 deals, checked on 200,000 more: after partner’s redouble, double the opponents’ bid for penalties when it is at the four level or higher.
- Which controls to show
- A study of 25,380 deals: control bids that show any control, against control bids that show only aces and voids, and two ways of mixing them. The two styles come out about level; a mix in which the first player to bid the trump suit shows any control, and partner only aces and voids, gains on new deals too. The app shows any control.
- Defending against 1NT: natural, Cappelletti, DONT or Multi-Landy?
- A first study of 6,088 deals found Multi-Landy slightly ahead of Cappelletti, too close to decide. A larger study will follow.
← Back to the front page
