Opening Leads: How to Choose, and How to Check Your Choice

Why the opening lead is the hardest card in the deal

Every later card is played with more information: dummy is on the table, partner has signalled, declarer has revealed a plan. The opening lead has none of that. You see thirteen cards and hear an auction, and the choice is often between four suits with nothing obviously right about any of them.

It is also the play where errors are least recoverable. A lead that gives declarer a trick he could not take himself, or that establishes the suit he needed to develop, usually cannot be undone later in the hand. That asymmetry is why the opening lead repays a disproportionate share of study time.

Read the auction before you read your hand

The auction is the only evidence you have, and it is usually more informative than the cards you are holding. Three questions are worth asking every time.

• Did the opponents stop, or did they drive? A partnership that crawls to 3NT after an invitational sequence has combined values and no source of extra tricks, so you can afford to sit back and make them find their own tricks. A partnership that blasts to game has extra shape or extra strength, and waiting tends to let declarer get discards.
• Did anyone show a long side suit? If dummy has advertised a five-card suit declarer can set up for discards, passive defence loses by default. Take your tricks, or establish them, before the suit is ready.
• What did partner say, and what did partner not say? A partner who could have overcalled and did not is unlikely to hold a good five-card suit. A preemptive raise says something about shape rather than strength.

Against a notrump contract reached by a limit auction — one notrump raised to three, say — the standard answer is your longest and strongest suit, because the contest is a race to establish length and there are no ruffing tricks to look for. Against a suit contract the auction matters far more than any general rule, because trumps change both the race and the risk.

Passive and aggressive

Almost every lead choice is a position on one axis: how much are you willing to give away in order to attack?

• An aggressive lead attacks a suit where you hold honours, hoping to set up tricks quickly. Leading from a king or a broken holding may cost a trick outright if declarer holds the missing honours. That is the price of speed.
• A passive lead — a trump, a sequence, or a suit where you have nothing to lose — gives nothing away and makes declarer do the work. It costs a tempo, which only matters if declarer has somewhere to put losers.

The choice follows from the auction. When declarer has a long, ready side suit, tempo is worth more than the risk and aggression is right. When the opponents have bid to a thin contract with balanced hands, declarer has no fast discards and every trick you hand over is one he could not have taken himself, so passive is right. Against a small slam the usual rule of thumb is passive unless the auction warned you that discards are coming: you only need one trick, and there is rarely time to develop a second.

Which card, once you have picked the suit

Choosing the suit and choosing the card are separate decisions, and the second is largely a matter of partnership agreement. These agreements are worth stating precisely, because the card carries information for partner as well as taking tricks.

• Top of a sequence. From K Q J x lead the king; from Q J T x lead the queen. The lead is safe and describes the holding immediately.
• Fourth best from length. From a broken suit against notrump, lead the fourth highest — from K J 8 6 3, the six. It keeps the honours behind declarer and lets partner apply the rule of eleven: subtract the spot from eleven to get the number of higher cards outside the leader's hand.
• Top of nothing, or middle-up-down. From three small cards, partnerships lead either the highest or the middle card. The point of the agreement is to distinguish a worthless holding from one containing an honour.
• Not away from an unsupported ace against a suit contract. Leading the ace surrenders control, and leading small from A x x risks losing it altogether.

Against notrump, leads tend to be about length; against a suit contract they tend to be about attitude. The reason is structural: in notrump long suits become winners, while against a trump suit the fifth card of a side suit is usually just a card.

A worked example: which card, or which suit?

West holds ♠ K J 8 6 3, ♥ 9 4, ♦ Q 7 2, ♣ 8 6 5, and hears one notrump on the left raised to three notrump — fifteen to seventeen opposite eight to ten, both balanced. The fourth-best spade looks automatic. The Opening Lead Analyzer simulated a thousand deals consistent with that auction for each candidate lead:

Two things stand out, and only one of them is a recommendation.

The first is that the top of the table is a statistical tie. With a thousand deals per candidate, a difference of one or two percentage points sits inside the noise; on this evidence the club, the heart and the fourth-best spade are the same lead. Reporting that a club is "1.5% better" would be claiming precision the simulation does not have.

The second is that the bottom of the table is not a tie at all. Leading the king from K J 8 6 3 beats the contract 18.5% of the time against 27.9% for a small one — a gap of roughly nine percentage points, far outside sampling error. The jack is nearly as bad. Both crash an honour into declarer's holding and hand over a trick that no amount of later defence gets back.

This is the practical way to use simulation on lead problems: not to rank the plausible candidates to the decimal, but to identify which candidates are clearly wrong. The gaps that survive the noise are the ones that change how you defend.

Running your own lead study

The quality of a simulated answer depends entirely on the constraints you supply, because those constraints define what "a deal consistent with the auction" means.

• Constrain the hidden hands the way the auction did. If a one notrump opening shows fifteen to seventeen balanced, set exactly that. An unconstrained hand lets the simulator deal shapes and strengths the bidding has already ruled out, and the answer degrades toward meaninglessness.
• Constrain responder too. A raise to three notrump implies a range, and often denies a four-card major if the partnership uses Stayman. Those restrictions change the hands that appear opposite.
• Use enough deals, and treat small gaps as noise. A thousand deals per candidate gives a standard error near one and a half percentage points on a set rate around thirty percent, so differences under about three points are unresolved.
• Compare the whole list, not just your first choice. The value of the exercise is usually in finding that a lead you would never have considered is as good as the automatic one — or that the automatic one is a disaster.

Frequently asked questions

Should I lead fourth best, or third and fifth?

Either works as long as the partnership agrees, and the difference matters far less than picking the right suit. Fourth best supports the rule of eleven; third-and-fifth gives partner a clearer read on length. Neither will rescue a lead from the wrong suit.

Is leading a trump always passive?

It is usually the most passive lead available, which is exactly why it is right when the auction suggests declarer plans to ruff losers in dummy. It is wrong when you need tricks quickly, and it can cost a trick outright when you hold a trump honour that would otherwise have been protected.

How many simulated deals do I need?

Enough that the gaps you care about are larger than the sampling error. A thousand deals per lead resolves differences of about three percentage points or more; if two leads finish within that, the honest conclusion is that the simulation cannot separate them.

The simulator likes a lead that breaks every rule I was taught. Should I believe it?

Check the constraints first, because an unconstrained hidden hand is the usual explanation for a strange result. If the constraints genuinely reflect the auction and the gap is well outside the noise, the simulation is telling you something real about that specific hand type, which is the whole point of running it.