Opener bids three of a suit. This particular opener always has seven cards when s/he does this. My a priori view of "average expectations" for opener's suit would be something like A-Q-T-8-6-4-2 (every other card for a total of seven). If I held, say, A-8-3, I would subtract the A and 8 from opener' hand, and replace it with a 7 and 5 (conservatively) or more likely, J-5 (a balanced high low), giving Q-T-7-6-5-4-2, or Q-J-T-6-5-4-2 respectively. If I held a K-J-9, I would mentally remove the opener's Q, and credit him/her with A-T-8-6-4-3-2.

Is this a reasonable way to estimate opener's suit holding? Could it actually be stronger (on average) if it is true that "no one" would open with Jxxxxxx or weaker? On the other hand, would I reduce my estimate if the opener would bid one of a suit with the above suit and say, a side ace?

For "guidance," consider the following questions.

  1. Is it true that based on sheer numbers, hands as strong as AKQxxxx or as weak as Jxxxxxx are low probability events?
  2. Do most conventions eliminate prempting with the above hands?
  3. If yes, is the remaining range, (AKJxxxx to Qxxxxxx) reasonably well represented by my hypothesized "average expectations" or do you need further agreements for a narrower range to "bracket" A-Q-T-8-6-4-2?

This question is for first and second positions, third and fourth have their won peculiarities.


Seat, vulnerability, and partnership agreement all affect required strength for preemptive openings.

On seat:

  1. In second seat, you are preempting 1 partner and 1 opponent; this seat has the narrowest range of suit quality and hand strength.
  2. In third seat, you are preempting 0 partners and 1 opponent; this seat has the widest range of suit quality and hand strength.
  3. In first seat, you are preempting 1 partner and 2 opponents. In this seat the hand might belong to your side or opponents'; it requires a narrower range than third seat but allows a wider range than second seat.

On vulnerability:

  1. When you are vulnerable and opponents are not, missing your side's game imposes a higher cost and preventing opponents from reaching their side's game (or preventing them from finding their best strain) has a lesser benefit.
  2. When you are not vulnerable and opponents are, missing your side's game imposes a lower cost while preventing opponents from finding their best contract has a greater benefit.
  3. When vulnerability is equal, the cost/benefit ratio is in between. You should be slightly more conservative when both are vulnerable than when neither is vulnerable, because the penalty for undertricks is higher (though I take this as a lesser concern).

On partnership agreement:

Broadly speaking, preempts have two purposes. The first is constructive: you wish convey to partner your hand type. The second is destructive: you wish to prevent opponents from finding their best contract. Different partnerships will have different philosophies here as to which goal is more important, and this will vary by seat and vulnerability as laid out above.

The more constructive you wish to be, the narrower the range of hand strength and suit quality you should have. The narrowest common range is something like "my opening 3 bids show exactly 7 cards with 2 of the top 3 honors or 3 of the top 5, in a hand with 7-10 points and no void or side 4-card major." This is a very reasonable standard for nonfavorable second seat openings. This will help partner evaluate their hand for a raise to 3NT or 4 of your major.

The more destructive you want to be, the wider your range should be. This increases the frequency of the bid while decreasing the information given to your opponents. This will make it more difficult for them to determine whether 3NT is safe, for example, and make it more difficult for them to place cards in the defenders' hands during the play. The widest "reasonable" standard for 3-bids is something like "a good six-card suit or a bad seven-card suit in a hand with 2-10 points."

Edited to answer the sub-questions:

  1. Given a 7-card suit (and no other information), it will have all 3 top honors 12% of the time, 2 top honors 44%, 1 top honor 37%, and 0 top honors 7% of the time.

  2. With most of my partners, I would open e.g. a 7321 hand with AKQxxxx and no other honors at the one level in most seats, but at the three level in second seat. Other partnerships open it at the three level more often. With a 7321 hand with Jxxxxxx and no outside honors, I would open at the three level not vulnerable in third seat; other partnerships would never open such a hand. As mentioned above, this depends wholly on your partnership's philosophy of constructive vs destructive preemption. There's not a single right answer here.

  3. I don't understand your "average expectations" rubric. Why do you care about the spots? If you have agreed to play sound preempts with your partner, and you hold Axxx AT Axxx Axxx across from your partner's 3H opening, you can be nearly certain that you'll take 10 tricks in a no trump or hearts contract. It's not reasonable to replace the AT in AQT8642 with the 3 and 5, because your partner has agreed to open only if holding two of the top three or three of the top five. They almost certainly hold KQxxxxx or KQJxxxx. If you agreed to play a wider range, you'll often have to lose a trick in hearts, but it's probably worth the gamble when vulnerable -- even if partner has a poor heart suit like Q9xxxxx, they'll generally have an outside king to make the tenth trick.

  • Upvoted for good philosophical discussion. But the other point of the question, is, do the "conventions" basically narrow the range of possible hands. For instance, if I have a 13 card suit, I would have exactly 10 hcp. With seven, I could have AKQJTxx or 8765432, but I would bid one of the suit with first hand, and pass with the second., basically ruling out those hands So is the actual range of hands ikely to be much narrower than the theoretical?I was thinking in terms of "permutations." – Tom Au Aug 8 '17 at 18:45
  • 1
    I don't really understand your question, but maybe this answers it: If you've agreed that opening 3 diamonds shows 2 of the top three honors in a 7 card suit, and your partner opens 3D when you hold KJx in diamonds, you know partner has AQxxxxx. You can't answer the question without knowing your agreements. – ruds Aug 8 '17 at 20:29

Just to elaborate on ruds' excellent answer, there are 12% of seven card hands that are clearly "too good" to pre emmpt" (open with 1 of the suit instead with AKQxxxx), and 7% with Jxxxxxx or worse. which should be passed. Most people would not preempt with these 19% of hands in the "tails."

Most people would preempt with the 44% of hands in the upper middle that had two of the top three honors. That leaves 37% in the lower middle.

Of that 37% of suits, most people would pre empt with QJTxxxx, KJTxxxx or AJTxxxx. Most people would not pre empt with Q9xxxxx or weaker. That leaves QJxxxxx and QTxxxxx as a matter of choice among Q headed suits. My preference is for "light" preempts, and I would open these hands. Likewise, I would pass K7xxxxx and K8xxxxx, but would open the three K-headed hands between K9xxxxx and KJxxxxx. I would also open any hand with a suit headed by an ace, even A765432, because it has one sure top trick and four probable tricks on length if the suit becomes trump.

Adding "three of the top five" to the 44% gives you roughly half of the seven card suits being opened with a preempt by most people. Since I favor "light" preempts, partner will know that I will bid one of the suit with the strongest 12% of hands, pass with about the weakest 12% (after eliminating a few more of the weaker hands slightly stronger than Jxxxxxx), and open the three quarters in the middle unless we agree otherwise. In any event, the partnership discussion will be about these 37% of hands in the lower middle.

Median strengh is roughly hands with "two honors out of top three" with AQ being a bit above, KQ being a bit below, and KQT or AJT being right on. The lower limit of my range (QTxxxxx) is about a king weaker; a king stronger would put me at the top of my range at AKJTxxx. Other partnerships prefer to lower the "variance" to a queen or so.

The addition of side honors would reduce my tendency to preempt. For instance, I would preempt with KQxxxxx and six side suit spot cards, but not with KQxxxxx KQx xx x, or even KQxxxxx Kxxx x x, especially if the other king-headed suit were major. In the first case, I have a genuinely weak hand with 5 high card points; in the second, I have my "fair share" of 10 hcps, in the third, 8 hcps, close enough to average, so I don't feel the need to preempt.

If partner has nothing, the first hand has no defense against a small slam, and limited defense against a grand slam, which is why I am willing to risk a four trick set at any vulnerability except "unfavorable" (vulnerable vs not). In the second case, I should be able to defeat a grand slam, and possibly a small slam, on my own. So I favor a "light" preempt with 3-7 hcp (QTxxxxx barely counts as such, "adjusted") that is, with some offense and no defense.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.