Braintree League preview: 92%, 87% and 86% win rates redraw the map of divisions two and three

In a club-level league, the most interesting thing is rarely the winner. I spent a weekend evening re-reading the win-rate tables of the Braintree Table Tennis League — a local competition in Essex, England — and stopped at a detail any predictive model would flag: the team installed as the number-one contender for division two this season is the very team that has just been pushed down from division one.

Black Notley B.

At the same time, in division three, an 18-year-old named JJ Calisin has been scheduled to move up to division one at Christmas — mid-season, not end-of-season. To me, that schedule carries more weight than any praise. It is a statement about development speed, and it only appears when a coaching setup already has enough data to believe the player does not need another buffer year in division three.

Those two signals — a relegated team installed as favourite, a junior promoted early — are the whole story of this season. The rest is consequence.

In 2026, the press-room door closed in front of me. Today, I read it with data.

The Braintree League runs on the standard local-league model. Teams are split by division, promotion and relegation follow at the end of the season, and individual records are measured by win percentage within the division a player actually competes in. There are no international ranking points, no rolling points system, no prize money. This is the bottom tier of the table tennis pyramid, and it is also the tier with the least public data.

The way the league's own preview is phrased says most of it. One player will feature in around half of the matches. Another team's fate depends on who backs them up and how often. That is the language of a squad-rotation model, where the biggest variable is not ability but availability.

Drawing on my experience of watching matches at club level, I always read this kind of document in two layers. The first layer is the published win rates. The second layer is the frequency words — on occasions, around half of matches — because that is where the prediction is actually decided. A team with three players at 90 per cent who each play six matches will lose to a team with three players at 70 per cent who play all twenty.

This is the methodological problem that costs me the most time: win rates in different divisions do not share a denominator. A player on 92 per cent in division two is not automatically stronger than a player on 86 per cent in the same division, and certainly cannot be compared directly with a player on 60 per cent in division one. Opposition quality differs, squad depth differs, match counts differ.

The league's public documents do not permit cross-division comparison, yet ranking title contenders forces exactly that comparison. The gap between the two requirements is where every local pre-season preview becomes vague.

Black Notley B have Neil Freeman, who scored 60 per cent in division one last season, Rev Matthews, who scored 86 per cent in division two, and Steve Kerns — a former men's singles champion — committed to around half of the matches.

Read crudely, Matthews is 26 percentage points ahead of Freeman. Read carefully, most of that gap is simply the gap between two divisions. A win rate in local table tennis is a division-dependent index, and comparing it across divisions without normalisation is the most common error in pre-season previews.

My approach: build a rough adjustment coefficient based on the quality gap between divisions, then convert every win rate onto a single scale. I do not need absolute precision, only the correct order. On that scale, Freeman's 60 per cent in division one sits in a band equivalent to 80 to 85 per cent in division two. That puts him level with Matthews and turns Black Notley B into a side with two spearheads rather than one star plus a supporting player.

Their status as number-one contenders is more reasonable than it looks.

Sudbury Strollers are the more interesting team. They finished second last season with Dave Fiddeman on 92 per cent and John Colvin on 75 per cent. On raw percentages, Fiddeman is the best-performing player anywhere in the data the preview provides. The preview itself concedes that Sudbury's fate depends on who backs them up and how often.

Here I have to state the limit of my own method. Fiddeman's 92 per cent could reflect three very different things: a player who is genuinely superior; a player who has been matched against weaker opponents in a favourable slot in the line-up; or a player who plays fewer matches than average and therefore has a small denominator. The preview does not give match counts. Without match counts, a percentage is only half the information.

Division three is more open. Finchingfield B finished second last season but lost Lucien Nolan-Bradford — a player who went through division three with a single defeat. In exchange they have Dave Punt dropping down from division two. And they face an entirely new team: Black Notley F, who impressed on debut.

Nolan-Bradford's only defeat came against Ben Southgate, 16-14 in the fifth game. That is the single data point in the whole preview that lets me say anything about clutch-point handling. A game stretched to 16-14 means both players had at least two chances to close the match and neither took them until the last. The sample size is one. With a sample size of one, I note it and stop.

Southgate scored 87 per cent in division three last season and now steps up to division two. That is the earliest and clearest test of the season: how long does a player with a high win rate in a lower division take to adjust to better opposition?

Tactics are what people draw on a blackboard. Data is what they draw on reality.

Most local pre-season previews are written as though the strongest team will win. The data in this preview suggests otherwise: at this level, the winning team is the one with enough players turning up often enough.

Black Notley B have a player available for around half the matches. Sudbury Strollers have a line-up whose fate depends on their backups. Black Notley F are a new team, and new teams mean unstable schedules. For each of these sides, the biggest risk is not opposition quality but appearance rate.

If my cross-division coefficient is misplaced, and 60 per cent in division one is really worth only 75 per cent in division two, Black Notley B drop level with Sudbury and the race becomes a pure depth problem. Under either scenario, the final conclusion is the same. That is a sign the conclusion is solid enough to use.

The remaining data set should be kept separate from the above. Ethan Collins is 12 years old and already has three cadet titles and one junior boys' title. Sai Suresh, 14, and Aryaman Singh, 13, are being introduced at Rayne D, both under the watchful eye of league coach Keith Martin. JJ Calisin, 18, moves up to division one at Christmas.

Reading junior titles as a direct indicator of senior-division performance is a mistake. Cadet titles are decided within an age-banded field. Playing division two and division three means facing men who have played this sport for twenty years and know how to slow a match down. The preview uses the word baptism for Suresh and Singh, and that word is more accurate than it looks.

My prediction model has no heart, and that is why it never gets hurt. But it also cannot see the moment when a 12-year-old realises the opponent across the table does not make mistakes. Those moments are not in the spreadsheet, and they will decide the season for several names on this list.

What I will track over the first six weeks is not the league table. It is the appearance rate.

Can Black Notley B keep Freeman and Matthews for at least four fifths of their matches? Can Sudbury Strollers find a reliable third player? How many senior-pace matches will Ethan Collins face in his first two months, and how does his win rate move after each defeat?

Calisin's scheduled Christmas move to division one shows the league operates a mid-season movement window for individuals. At small scale, that is a miniature transfer market: one player leaves division three, one slot opens in division three, one team has to fill it. In my day job in the transfer market, movements like that are usually ignored because they generate no headlines. They are where the table is actually decided.

Black Notley have enough depth to add an F team. Finchingfield B lose a player and immediately gain one dropping down from above. Those movements never appear in the win-rate tables, but they shape next season's win-rate tables.

Players leave the table, spectators leave the stands, but data never leaves the game.