Braintree League Pre-Season: Win Percentages, the Division Trap and the Gaps in the Data

Last season in Division Three of the Braintree Table Tennis League, Lucien Nolan-Bradford walked through the campaign almost untouched: exactly one defeat, and that defeat went to a fifth game, ending 16-14 against Ben Southgate.

I stopped for a long time on that detail when I read the new season preview published by Table Tennis England. Not because 16-14 is a handsome scoreline, but because it is the only piece of evidence in the entire document showing a Division Three champion-level player being pushed to his limit by a specific opponent, in a specific game, at a specific moment.

Everything else in the preview is a much bigger number: 86%, 92%, 87%, 75%, 60%. Clean, quotable figures. And precisely because they are clean and quotable, they are the kind of data most easily misread in sport.

Numbers do not lie, but the people who read them do.

I have worked as a sports data analyst for many years, mostly in football and table tennis. I entered the industry in 2026 as a fact-checker at Sports Illustrated, where I learned something that has followed me ever since: before arguing about what a number means, verify how it was collected, by whom, over what period, and with what sample.

In 2026 I analysed 240 matches from China League One and showed that Dalian Yifang, a side without stars, had the best expected-goals numbers in the division at 1.7 xG and 0.8 xGA, then predicted promotion with 94% probability. My editors called it reckless. Dalian Yifang won the title with 64 points, five clear of second place. In 2026 I used an expected-goals model to warn that Germany, the reigning world champions, risked group-stage elimination; after two matches their xGA stood at 3.2 while their attack had generated just 1.8 xG. The piece was mocked. When Germany lost 0-2 to South Korea, my inbox filled with apologies. In 2026 I collected data from 152 Bundesliga and La Liga matches to measure behind-closed-doors effects: home win rates fell from 44% to 29%, and average goals dropped by 0.7.

Those three episodes taught me the same lesson in three different ways: a model is only as trustworthy as the data feeding it. And the Braintree League is the hardest kind of data there is.

Why a village league is harder to analyse than a final

Braintree Table Tennis League is a team competition run at local level in Essex, England, under the Table Tennis England system. It is organised into tiers with promotion and relegation between seasons. There are no ITTF or WTT ranking points, no published prize money, no Olympic pathway. The only published performance metric in the preview is each player's individual win percentage from last season, plus each team's division placing.

That simplicity is deceptive. A WTT final gives us thousands of data points: service points won, third-ball attack conversion, point distribution by court zone, average rally length. Braintree gives us one percentage per player and a handful of descriptive sentences. We are doing analysis on an extremely thin sample, and every conclusion must carry an uncertainty warning attached.

Which is another way of saying: if I delivered a definitive verdict on this league, I would be doing my job badly.

There is one rare advantage here. Braintree has no player agents. Nobody is manufacturing noise to inflate a client's value. In my years in this trade I have always held that agents are the biggest hidden cost in professional sport, because the noise they generate distorts markets and contaminates data. A village league has none of that. The data is raw, but it is clean in the sense that nobody is bending it on purpose.

What it does have is a different kind of noise: human availability.

Division Two: Black Notley B and the problem of three variables

The preview calls Black Notley B's squad the team to beat in Division Two. Three data points support that judgement. Neil Freeman scored 60 per cent in Division One last season. Rev Matthews scored 86 per cent in Division Two. And former men's singles champion Steve Kerns will be included for around half of the team's matches.

These three things are not the same kind of fact, and they need separating.

Freeman's 60 per cent in Division One carries far more information than the other two numbers. Division One is the league's top tier, the densest concentration of strong opponents, where every fixture carries real risk of defeat. Scoring 60 per cent there and then dropping to Division Two is a reverse quality shift: he leaves a harder environment for an easier one while his ability stays fixed. In sports analytics this is one of the few cases where causal language is defensible, because the only variable that changed is opponent quality, not the player.

Matthews's 86 per cent must be read differently. It is a strong record, but it was produced in the very division he is about to play in again. There is no upgrade in context. It is a floor, not a ceiling.

Steve Kerns is the third variable and the most underrated one. "Around half of matches" is a structural number, not a form number. It means Black Notley B's shape will change from fixture to fixture. In a league where every individual point counts directly toward the team result, losing a champion-calibre player for half a season is equivalent to playing half a season with a visibly weaker side, no matter how strong the two remaining anchors are.

If forced into a conditional projection: the probability that Black Notley B finishes in the leading group of Division Two is high, provided Freeman plays at least 80 per cent of fixtures at something close to his Division One level. If he dips, or if Kerns misses more than expected, the division opens up. I am deliberately attaching no specific probability to that projection, because the sample is far too thin for honesty.

Relegation as a structural advantage

One line in the preview matters more than any percentage: a relegated team can always be looked on as potential winners.

This is a structural rule, not a morale booster. When a side is relegated, it carries down a squad that was built to compete a tier higher. Players do not lose skill over a summer. The only thing that changes is the average quality of the opposition. This effect is strongest in leagues where the gap between adjacent divisions is small but stable, and Braintree belongs in that group.

The paradox is that a relegated team's individual win percentages often look unimpressive, because they were earned in the harder division. A team that stayed in the lower division will carry far prettier numbers. Rank the two on raw percentages alone and you will get the order completely wrong.

That is why I do not use a single strength ranking for this league. I keep two tables: one on raw win percentage, one adjusted for division. The second always differs from the first, and in the opening weeks the second is always the better predictor.

The addition that does not exist

The most common error in reading a team-format league preview is to add up individual win percentages and treat the sum as team strength.

That addition does not exist. A team result is not the sum of individual records, because the three slots face three different opponents in a given fixture, and the difficulty of each slot differs. A side with two very strong players and a weak third slot will repeatedly finish matches on knife-edge scorelines, and in a points-per-fixture format, a narrow win and a heavy win are worth the same.

The gap between champions and runners-up in this kind of league tends to live in the weakest slot, not the strongest one. From that angle, Kerns playing half a season is not a footnote. It means Black Notley B will field two different teams across one season: one with three strong slots, one with two strong slots plus a substitute. Whatever their final win percentage looks like, it will be an average of those two teams.

Sudbury Strollers: the second force and where the ceiling sits

Sudbury Strollers finished second last season. They carry two very pretty numbers: Dave Fiddeman at 92 per cent, John Colvin at 75 per cent.

This is the most dangerous trap in the whole preview.

92 per cent is the highest figure quoted anywhere in the document, and instinct reads it as the mark of the division's best player. But win percentage in a tiered league does not measure absolute strength. It measures relative strength against the specific set of opponents faced. A player at 92 per cent in Division Two and a player at 60 per cent in Division One may sit at the same true level, or the order may even reverse. The preview provides no head-to-head data between them, and no official conversion mechanism between divisions.

I tell colleagues in the data room this: xG is not a yardstick, it is the match confessing. At Braintree, win percentage performs exactly that role. It does not say how good a player is. It confesses which kinds of opponents he met, and how much of that he handled.

The more telling line about Sudbury Strollers is a very short one: their fate may depend on who backs them up and how often. That is a statement about depth, and it outweighs 92 per cent and 75 per cent combined. A team-format table tennis side needs three reliable slots per fixture. Two are filled. The third is the unknown, and that unknown decides the season.

Division Three: thinnest data, widest variance

Finchingfield B finished second in Division Three last season. This year they lose Lucien Nolan-Bradford, who went through the division with a single defeat, but gain Dave Punt, dropping down from Division Two. The preview rates their line-up as still strong while warning they could be stretched by Black Notley's new F team.

This is a very typical local-league exchange: a team does not lose total capacity, it loses a specific emphasis. Nolan-Bradford was the man who won the hard matches; Punt is the man who guarantees stability. In a division decided by a handful of close fixtures, the difference between winning the big matches and not losing the small ones may never appear in any win-percentage table.

One detail I noticed but cannot quantify: Ray Nolan-Bradford, most likely Lucien's father, remains in the Finchingfield B squad. In community leagues, family and club ties stabilise a line-up in ways no index captures. A team that loses its strongest player but keeps its human structure tends to recover faster than a team that keeps its strongest player but loses the structure.

Black Notley's new F team is the opposite variable. The preview notes that its players impressed on debut. But we are weighing a completed season against a few debuts. The sample behind "impressed on debut" could be one match, two matches, three. Nobody knows. One impressive debut predicts nothing about March.

What the creation of a new F team does tell me, and this is the inference I trust most in this piece: the club has enough membership depth to sustain an additional team without hollowing out the existing ones. That is an organisational health signal, not a performance one, but it is the kind of signal that decides performance over three to five seasons.

Six young names and a pipeline that is running

The most interesting part of the preview is not the title race. It is the junior cohort.

Ethan Collins is twelve years old and already holds three cadets' titles and one junior boys' title. Sai Suresh, 14, and Aryaman Singh, 13, are described as facing a baptism in adult divisions under the watchful eye of league coach Keith Martin. JJ Calisin, 18, whose strides are described as impressive, is scheduled to move up to Division One at Christmas.

Four names, three age brackets, two divisions, and one promotion date placed mid-season.

When the stands are empty, I see the truest version of a team. In a local league the stands are almost literally empty. No crowd, no cameras, no sponsor waiting on the result. What remains is pure motive: people play because they want to play, and clubs push juniors into adult divisions because they believe in the process rather than because of results pressure.

The way Braintree handles its juniors suggests a deliberate development model. Collins is being thrown against adults at twelve, in a second consecutive season at the same level. Suresh and Singh are moved up at thirteen and fourteen. Calisin has a fixed promotion date, Christmas, which implies a planned pathway rather than an improvised rise.

Three cadets' titles at twelve is data about the ceiling of potential, not data about the ability to absorb pressure in adult divisions. A second season at the same level is the harshest test a young player faces, because opponents now have data on him and have adjusted.

The Christmas move also reveals that the league runs a mid-season transfer or progression window, at least at individual level. Professional sport has the winter window with money, contracts and agents. Here the same mechanism exists with none of the money, none of the contracts and none of the rumours. Just an eighteen-year-old being pushed into a harder division to see how far he bends.

That is the kind of data that, ten years from now, may turn out to be the starting point of a career.

The neglected variable: presence

If I had to pick one variable to decide this Braintree season, I would not pick win percentage. I would pick presence.

The preview uses the language of local table tennis: around half of matches, on occasions, depending on who backs them up and how often. That is the vocabulary of a rotation model, not a fixed roster. At this level fixtures are decided not by employment contracts but by work, family, health and geography.

The analytical consequence is plain: every win percentage in this league is calculated on a non-uniform sample. A player at 86 per cent across twelve matches is not the same data object as a player at 86 per cent across twenty-two. The preview publishes percentages without denominators. That is not a flaw in a news preview, but it is enough for me to refuse any absolute claim about final standings.

Based on my experience following matches at club level, I have learned something box scores never teach: in this kind of league, a team losing a frequent player changes more than a team losing a strong but occasional player. Regularity weighs more than peak performance across a long season.

The season's risk surface

Four risk clusters can be read out of the preview itself.

The first belongs to Black Notley B: the "team to beat" label may not convert if Kerns misses more than half the season. Likelihood medium, impact medium. Mitigation lies in keeping Freeman and Matthews as fixed slots and finding a reserve of adequate standard.

The second belongs to Sudbury Strollers: their ambition is capped by their own depth. Likelihood medium, impact medium. Mitigation is locking in a third player early and keeping Fiddeman and Colvin playing regularly.

The third belongs to the juniors: a baptism in adult divisions can bring a long losing run, and a long losing run can damage development if unmanaged. Data cannot measure this risk. Only a coach can.

The fourth belongs to the system: the league runs on voluntary participation, so a few mid-season withdrawals can distort an entire division's table. Low likelihood, medium impact, and no prevention beyond flexible scheduling.

What the document does not say, and why it matters

The preview raises no rule, disciplinary or equipment issues. No eligibility disputes. No registration constraints. The only implied mechanism is promotion and relegation.

The use of terms like cadets' titles and junior boys' titles implies an age-band structure run by Table Tennis England, but nothing is said about eligibility conditions. That is the document's largest structural gap.

The second gap is larger still: there is not a single line of technical or tactical information. Nothing on playing style, service patterns, rubber types, backhand attack quality, or how anyone handles deciding points. No shot-level data. A technical analysis of the Braintree League can only be written once the season starts and somebody sits down to take notes. I will not fill that gap with speculation, because doing so would betray the principle I work by.

This is also where I should say something about my own profession. Sports data analysts are pushing deeper into the dressing room, and their conclusions often detach from the actual rhythm of competition. I can tell you that Black Notley B's probability of leading Division Two is high, under a stated condition. I cannot tell a twelve-year-old how to play at 15-15 in the fifth game in a cold hall in Essex on a Tuesday night. Spreadsheets do not know what that feels like. The person sitting at the next table does.

The contrarian reading

There is a reading of this preview I consider better than the obvious one, and it runs against instinct.

The popular reading: Black Notley B are favourites because they have two players above 60 per cent and a former champion; Sudbury Strollers are the main threat because they have a 92 per cent player; Division Three is a race between Finchingfield B and Black Notley F.

The alternative reading: Division Two is a race between two teams sharing the same structural weakness, insufficient depth, and the side that fixes it first wins. Black Notley B have Freeman and Matthews plus half a season of Kerns. Sudbury Strollers have Fiddeman and Colvin plus an unnamed third slot. On this reading the gap between them is not in the strongest slot. It is in how stable the weakest one is.

Here I have to be at maximum caution. Nothing in the preview lets me assert that the team with the more stable third slot will win the title. That is a grounded hypothesis, not a conclusion. And a hypothesis is worth exactly what it survives in the opening weeks.

The same applies to Division Three. Nolan-Bradford leaving Finchingfield B and Black Notley F arriving with impressive debuts creates the appearance of a power shift. But we are comparing a completed season against a handful of debut evenings. That is a large sample against a near-zero one.

One more note on the data itself. The percentages come from Table Tennis England, most likely drawn from official league records. The source is reliable at event level: a national governing body has no incentive to distort a local league's win rates. But source reliability is not data completeness. An accurate number and a meaningful number are two different things.

I know this because I have made the mistake. My early 2026 pieces on China League One were so packed with tables and charts that an editor had to remind me readers do not need three tables to grasp one argument. Since then I have kept a rule: no more than three numbers per argument. Three numbers are enough to prove a point. The fourth becomes decoration.

One final point, because it concerns how we watch sport in general. Spectators routinely mistake flashy execution for high-level play, in any discipline. In table tennis a powerful loop always looks better than a safe placement into the far corner. But at championship level, what decides results is usually control and the management of tempo, neither of which leaves a visible trace in any statistical table. Fiddeman's 92 per cent does not tell me how he won. It only tells me he won.

And that is the entire limit of this preview.

Signals to watch next

Four signals will tell us whether the preview was right, and none of them can be read off last season's percentages.

First, how many of Black Notley B's opening four fixtures Steve Kerns actually plays. Three out of four supports the favourites tag. One out of four and the team's baseline must be recalculated.

Second, the identity of Sudbury Strollers' third slot. Until that name appears regularly in results, their position remains a blank.

Third, Ethan Collins's first half of the season. A second year at the same level is the hardest test a junior faces, because opponents now hold data on him.

Fourth, JJ Calisin's Christmas milestone. If he really moves up to Division One on schedule, the league's development pipeline is running on time. If the date slips, the story is not one individual's setback but data about the limits of an entire system.

I will read Braintree's results in the coming weeks the way I read Bundesliga scorelines during the pandemic season: not looking for who won, but for which variable just changed value.

A league table is a summary; raw data is the testimony. And in a village league like Braintree, the testimony only begins when the first ball is tossed.