How does seeding work in a tennis draw?

A draw made purely at random can put the two best players against each other in round one. One of them goes home in forty minutes, and the final is contested by somebody who never had to beat either.

Seeding fixes exactly that and nothing more. It is not a reward and it does not give anybody an easier path in principle — every seed still has to win the same number of matches. It only guarantees that the players ranked strongest are kept apart until the round where meeting is the point.

The consequence is worth stating plainly, because it is the objection people raise: seeding does make the early rounds less likely to be close. That is the trade. A tournament that produces one thrilling first-round match and a poor final has served its field worse than one that saves the meeting for Sunday.

This is fixed by the size of the draw, not by how many good players entered. The ITF numbers are:

Draw size	Seeds
8	2
16	4
24 or 32	8
48 or 64	16
96 or 128	16

Notice that the count stops growing at 16. A 128-draw has the same number of seeds as a 64-draw, because beyond a quarter of the field the protection stops being protection.

The general shape is one seed per eighth of the draw up to 32, and it is a useful sanity check: if you find yourself seeding more than a quarter of your entries, you are no longer separating the strong players, you are arranging the result in advance.

Qualifying draws are seeded separately, by section rather than by size — no more than 4 seeds in a draw with 2 sections, 8 with 4 sections, and so on up to 16.

Only the top two seeds have a fixed position. Everyone else is drawn by lot within a fixed set of lines, which is what keeps a draw both protected and genuinely random.

Seed 1 goes on line 1. Seed 2 goes on the last line — line 16 in a 16-draw, line 32 in a 32-draw, line 64 in a 64-draw, line 128 in a 128-draw. That single decision is the whole of the protection: opposite ends means opposite halves means they cannot meet before the final.

The rest are drawn in pairs and groups of four. For the draw sizes a club will actually use:

Seeds	16 draw	32 draw	64 draw
3, 4	lines 5, 12	lines 9, 24	lines 17, 48
5–8	—	lines 8, 16, 17, 25	lines 16, 32, 33, 49
9–12	—	—	lines 9, 25, 40, 56
13–16	—	—	lines 8, 24, 41, 57

Read that as: in a 32-draw, seeds 3 and 4 are drawn by lot between lines 9 and 24, so nobody chooses which of the two gets the friendlier quarter. Seeds 5, 6, 7 and 8 are drawn by lot across lines 8, 16, 17 and 25.

Everybody else goes in at random. Unseeded players have no protection at all, which is why the first round of a strong draw is where the upsets live.

Entry lists rarely arrive as 16, 32 or 64. When they do not, the draw is made at the next size up and the unfilled lines become byes — a free pass into round two.

The allocation surprises people: byes are awarded to the highest seeds as a first priority. Any remaining byes are drawn by lot, spread as evenly as possible across the sections of the draw.

The reason is not favouritism, it is arithmetic. Someone has to get the empty line. If byes were scattered at random, a bye would land next to an unseeded player who then meets a seed in round two having played once, while the seed has played twice. Giving them to the top of the list keeps the shape of the draw consistent with the seeding you just did.

In practice this is also the thing that saves your schedule. Twenty-four entries in a 32-draw is eight byes and 23 matches instead of 31, and those eight missing matches are the first round, the most congested slot of the day.

One detail worth knowing at club level: a bye is not a win. In a round robin that feeds into a bracket, a player has not qualified by taking a bye — they have to play and win a match first.

Beyond the seeds themselves, a draw has a second job: not to waste the first round on two players who see each other every week.

The ITF rule is about national associations. If two seeded players from the same nation are entered, they are drawn into different halves. If three or four are entered, the first two are separated as above and the third and fourth go into the quarters that do not already contain one of them. The principle extends down through all the seeds.

At club level the same instinct applies to a different unit. Two players from the same club, or two who train together every Tuesday, produce a poor first-round match — they know each other's game and one of them was going to lose anyway. Separate them if the draw allows it.

Be honest with yourself about where this stops. Manual adjustment beyond a stated rule is how a draw loses credibility. Announce that you separate club-mates in the first round, apply it consistently, and stop there.

When a tournament runs groups before a knockout, the seeding job splits in two: place the seeds across the groups, then place the qualifiers into the bracket.

Into the groups, one seed per group, in order. Seed 1 heads group A, seed 2 heads group B, seed 3 group C, and so on. The point is that no group should contain two of your protected players.

Out of the groups, the winners are placed so that the two strongest groups' winners are kept apart. The ITF pattern puts the winner of group A on the first line of the bracket and the winner of group B on the last, then draws the winners of groups C and D into the two middle-quarter lines, and the rest at random.

If two players advance from each group, the standard convention is that a group winner never meets the runner-up from their own group before the final possible round. Getting this wrong is the most common flaw in a home-made group-plus-knockout draw: two players who already played each other in the group meet again in the quarter-final, and one of them correctly points out that they drove in for a repeat fixture.

Every rule above assumes a ranking to seed from. Most club tournaments do not have one, and this is where organisers either improvise badly or refuse to seed at all.

Refusing is worse than improvising. An unseeded 32-draw at a club will, more often than people expect, put your two best players in round one. The purpose of seeding does not go away because the data is soft.

Use whatever ordering you actually have, in this order of preference: last year's finishing positions, the club ladder or internal ranking, the league team ordering, and finally the committee's honest opinion. Then say which one you used. A seeding announced as based on last year's results is accepted; the same seeding produced silently is questioned all afternoon.

In JudgeMate, seeding comes from the start number — the order players sit in on your entry list is the order the draw reads as strength. There is no connection to any federation ranking, so the judgement is yours and it is worth writing down next to the draw sheet. Once the order is set, the draw places the seeds on the lines above and puts everyone else in at random.