Every team has the player everyone side-eyes — the one who wins too much for their rating. So we pulled four seasons of Marin-area USTA results and asked a plain question: how often does a dominant season actually precede a move up a level?
Winning proves you belong at your level. It doesn't prove you've outgrown it: win three-quarters of your matches or better and only 37% move up. That is more than twice the 16% base rate, so the system is plainly responding — but it still leaves most dominant seasons ending exactly where they started.
That 37% averages over two very different worlds. Among players at 3.5 and below, 55% of dominant seasons were followed by a move up (37 of 67). At 4.0 and above, just 19% were (12 of 64).
Some of that is simply room. This sample has 220 teams at 3.5 and 25 at 5.0 — the higher you climb, the less there is above you. The same dominant record means something quite different at 3.5 than at 4.5.
Of the 35 players who won 75%+ of their 2025 rated matches and are back on a roster this year, 20 are still at the same level. These 35 seasons are part of the four-season figure above, not a second check on it.
Follow everyone who put up a 70%-plus season between 2022 and 2024 — 105 players — and by 2026 about half of them, 56, had moved up a level at some point. The other 49 never did.
So a big season is less a trigger than a coin flip spread over several years. Some players move up a season or two after the year that made the case; plenty are simply left where they are. This data can't tell you which is happening to any particular player — only that the move is nowhere near automatic.
I'm in this data, so here is mine, including the parts that don't flatter. In 2023 I went 14–3 in Adult league — 19–8 counting the Combo and Mixed matches this analysis excludes — while self-rated 3.5. My first computer rating came back 3.5. I went 20–5 the next year and the move to 4.0 landed. Then at 4.0 I went 2–5 in Adult league last season and stayed put. The level found me eventually; it took its time, and it has not been kind since.
So the next time someone across the net gets accused of being underrated: maybe they are — or maybe they are rated correctly and playing a lot of tennis. This data can't tell you which. What it does say is that staying put after a big year is the ordinary outcome rather than evidence of anything, and that at 4.0 and above it is what usually happens.
Free, no signup. The Tennis Desk reads the real USTA NorCal match record.
Grounded — every figure above is a count over a stated denominator. No estimates, no model output: just match results and what the rating system did next.
Scope. USTA NorCal league rosters and scorecards for ten Marin-area clubs — BELVEDERE TENNIS CLUB, BAY CLUB ROLLING HILLS, MARIN TENNIS CLUB, MILL VALLEY TENNIS CLUB, MT TAM RACQUET CLUB, RAFAEL RACQUET CLUB, TIBURON PENINSULA CLUB, SCOTT VALLEY SWIM & TENNIS CLUB, NOVATO TENNIS CLUB INC and HARBOR POINT TENNIS CLUB — covering 2022–2026. It is a ten-club sample, not a census of Marin and not a NorCal-wide study: players at other clubs appear here only as opponents, so their own records and rating changes are outside the data.
Method. Win rates count the matches USTA uses to generate a year-end rating, per the USTA League Regulations. That is every Adult Division age group — 18 & Over, 40 & Over, and the 55, 65 and 70 & Over groups, which play at a combined team rating but are Adult Division all the same (Reg 1.04D(5)). Combo is not counted: NorCal runs it as a local league, not one of the two USTA League Divisions that generate ratings. Mixed is not counted either, with one exception the rule states explicitly — for a player who played no Adult Division matches that season, their Mixed play is what generates their rating, so it counts for them (Reg 1.04F(1)b). The rating level itself is read from all of a player's teams. A record is the wins and losses USTA shows on a player's own page, summed across their teams that season. Minimum 15 matches — which in practice means players on two or more teams, the most active in the sample. A lower floor lowers every rate on this page; a higher one raises them.
What "moved up" means. The player's NTRP level rose between seasons. We did not separate an earned year-end computer bump from a self-rate correction or an appeal, and roughly one in seven of the moves counted here began from a self-rated or appealed rating. Bands are left-closed, so exactly 75.0% counts as 75%+. Players who did not return the following season are not counted.