Racquet Rating Lab

Data-driven tennis analytics

⌕
HomeTop PlayersExplorerCompareMajorsSummariesMethodology
⌕

Data sources

Match data by Jeff Sackmann / Tennis Abstract, licensed CC BY-NC-SA 4.0. Racquet Rating transforms this data into a canonical match schema and derives its own ratings from it. Racquet Rating is operated on a non-commercial basis.

Tennis databases, files, and algorithms by Jeff Sackmann / Tennis Abstract is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. Based on a work at https://github.com/JeffSackmann.

Full data sources, licence and derivation details

Loading player profile…
← Back to players

ATP·ITA

FlavioCobolli

FC

Racquet Rating: 542.6 out of 1000. Status: Active.

Racquet Rating

542.6/ 1000

Status: ACTIVE

Last rated match: May 25, 2026

Rating movement: up 16.8 points over 2 months, since 2026-03-31.

▲+16.8over 2 months

Since the published rating of 525.8 on March 31, 2026.

Official ATP Ranking

#103,540 ranking points

As of 2026-06-08. Source: JeffSackmann. Separate from the Racquet Rating, and not an input to it.

Evidence behind this rating

Qualifying matches
174of 40 required
Record in window
98–76
Rating window
4 years
Distinct opponents
106

Model v1.1.0 · match data through 2026-05-25 · canonical id atp:207925.

Racquet Rating requires at least 40 qualifying matches in the current 4-year rating window. Players without sufficient recent evidence are not assigned a current rating. How ratings are calculated

Match data through 2026-05-25 · model v1.1.0. Data sources · Methodology

Rating history

Every published Racquet Rating observation for Flavio Cobolli, from the canonical rating artifact.

Peak in available history
542.6May 25, 2026 · the current rating
History coverage
Jun 2024 – May 20269 quarterly observations · 1.9 years
Change over 2 years
+56.7over 1.9 years, to 542.6
Racquet Rating historyRacquet Rating rose from 485.9 on June 30, 2024 to 542.6 on May 25, 2026, a gain of 56.7 points. 9 published quarterly Racquet Rating observations. Peak in the available history: 542.6 on May 25, 2026.460480500520540560Jun 24Sep 24Dec 24Mar 25Jun 25Sep 25Dec 25Mar 26May 26

9 published quarterly observations. Hover or tab through the plot for a single observation.

Axis shows 460–560 of the canonical 0–1000 Racquet Rating scale.

View 9 observations as a table
Racquet Rating rose from 485.9 on June 30, 2024 to 542.6 on May 25, 2026, a gain of 56.7 points. 9 published quarterly Racquet Rating observations. Peak in the available history: 542.6 on May 25, 2026.
DateRacquet RatingChangeMatches in window
June 30, 2024485.9—48
September 30, 2024506.6+20.772
December 31, 2024510.9+4.379
March 31, 2025479.6-31.390
June 30, 2025529.7+50.1114
September 30, 2025532.8+3.1126
December 31, 2025524.6-8.3137
March 31, 2026525.8+1.2153
May 25, 2026542.6+16.8—

Published quarterly observations from the canonical rating artifact, model v1.1.0. Never interpolated, smoothed or back-filled.

Why this rating?

The published Racquet Rating is the sum of five weighted model components. This is that sum, for Flavio Cobolli.

Racquet Rating 542.6. Five model components, largest contribution first: Performance Index, 205.1 points; Form Momentum, 131.1 points; Surface Balance, 94.2 points; Era Normalization, 58.8 points; Match Context, 53.3 points. Together they total 542.6 points, which reconciles with the published rating of 542.6. Component scores are 0 to 100; contributions are rating points out of 1000.

The largest contribution is Performance Index, at 205.1 of the 542.6 rating points (38%), followed by Form Momentum at 131.1 (24%). Match Context contributes least, at 53.3 (10%). The rating is computed from 174 qualifying matches over the 4-year rating window.

  1. Performance IndexLargest contribution

    205.1rating points38%

    Score 58.6 out of 100 × weight 35% × 10 = 205.1 rating points

    What Performance Index measures

    How well the player wins points and matches — serve and return points won, break points converted and saved, results against elite opponents, and the strength of the field they have faced.

    It carries the most weight of any component, so it usually accounts for the largest share of a rating.

    Seven inputs are winsorised, standardised against fixed baselines and combined through a logistic map to a 0–100 score. Elo and strength of schedule are two of those seven inputs — they feed this component, they are not components themselves.

    Across the 159 rated players this component contributes between 76 and 328 rating points — a spread of 252.

    Published evidence

    Qualifying matches
    174
    Record in window
    98–76
    Observations with the full feature set
    148 (85.1%)
    Observations against top-ranked opponents
    78
    Elo (an input, not a component)
    1830.59

    These are the observations the component was computed from. A higher share with the full feature set means more of the serve, return and break-point detail was available.

  2. Form Momentum

    131.1rating points24%

    Score 52.5 out of 100 × weight 25% × 10 = 131.1 rating points

    What Form Momentum measures

    Recent trajectory: results over a 26-week window with exponential decay, a fatigue adjustment from minutes and sets played, and an allowance for a player returning after a long absence.

    It is what makes the rating a current-form measure rather than a career résumé.

    The published score is shrunk toward the neutral 50 in proportion to how much evidence supports it, so a short hot streak is damped rather than treated as equal to a long validated record. Both the raw score and the confidence used are published, so the adjustment can be checked.

    Across the 159 rated players this component contributes between 81 and 219 rating points — a spread of 138.

    Published evidence

    Score before evidence damping
    52.46
    Evidence confidence
    1.000
    Evidence span
    1,323 days
    Distinct opponents
    106
    Top-ranked opponents faced
    78

    The published score is 50 + (raw − 50) × confidence. Confidence rises with matches, span, distinct opponents and top-opponent exposure, so a short run is damped toward neutral rather than treated as a settled trajectory.

  3. Surface Balance

    94.2rating points17%

    Score 62.8 out of 100 × weight 15% × 10 = 94.2 rating points

    What Surface Balance measures

    How evenly the player performs across hard, clay and grass — an entropy-like measure of the spread, with a small penalty for extreme specialisation and credit for consistency across surfaces.

    It rewards a game that travels. It is a measure of balance, not of how good the player is on any one surface.

    A high score means performance is spread evenly across surfaces; a low score means it is concentrated. Neither reading says which surface a player is best on — that would need per-surface data the artifact does not publish.

    Across the 159 rated players this component contributes between 15 and 127 rating points — a spread of 112.

    Detailed evidence for this component is not currently exposed. It is computed from per-match detail that the published artifact does not carry.

  4. Era Normalization

    58.8rating points11%

    Score 39.2 out of 100 × weight 15% × 10 = 58.8 rating points

    What Era Normalization measures

    An adjustment against fixed era baselines and tour-depth proxies, so players from different periods remain comparable.

    It is what allows one 0–1000 scale to span the whole dataset rather than only the current season.

    The baselines are frozen, so this component reflects when a player competed and against what depth of field, not how they played.

    Across the 159 rated players this component contributes between 13 and 138 rating points — a spread of 124.

    Detailed evidence for this component is not currently exposed. It is computed from per-match detail that the published artifact does not carry.

  5. Match ContextLowest contribution

    53.3rating points10%

    Score 53.3 out of 100 × weight 10% × 10 = 53.3 rating points

    What Match Context measures

    How results compare with expectation once the round and the event are taken into account, so performance on bigger stages counts for more.

    It carries the smallest weight of the five, so it moves a rating the least.

    The round and event weights are calibrated in the pipeline and published as a single score; this page does not recompute them. Across the rated field this component varies far less between players than the others, so it rarely explains why two players differ.

    Across the 159 rated players this component contributes between 43 and 64 rating points — a spread of 21.

    Detailed evidence for this component is not currently exposed. It is computed from per-match detail that the published artifact does not carry.

Total component contributions
542.6
Published Racquet Rating
542.6 / 1000

The five contributions reconcile with the published rating to within 0.004 rating points — the artifact's own rounding.

Elo 1830.59 (career peak 1838.60) is one of the inputs to Performance Index. It is on its own scale, it is not a component of the Racquet Rating, and it is not comparable with the numbers above.

Component scores are 0–100; contributions are rating points on the canonical 0–1000 scale, model v1.1.0. How Racquet Rating works

Surface performance

Performance over the headline rating window. Rates are shown only where the surface sample provides sufficient evidence.

Surface performance over the rating window: 174 matches, 174 with a recorded surface. Surfaces played: Hard, Clay, Grass. Hard, 47 wins and 46 losses, 50.5 percent, 95 percent interval 41 to 60 percent. Clay, 42 wins and 23 losses, 64.6 percent, 95 percent interval 52 to 75 percent. Grass, 9 wins and 7 losses, 56.3 percent, 95 percent interval 33 to 77 percent. These rates describe this player only and are not used to rank players.

  • Hard

    47–46Hard, 47 wins and 46 losses, 50.5 percent, 95 percent interval 41 to 60 percent.93 matches

    50.5% (41–60%)

  • Clay

    42–23Clay, 42 wins and 23 losses, 64.6 percent, 95 percent interval 52 to 75 percent.65 matches

    64.6% (52–75%)

  • Grass

    9–7Grass, 9 wins and 7 losses, 56.3 percent, 95 percent interval 33 to 77 percent.16 matches

    56.3% (33–77%)

174 matches in rating window · 174 known surface

How surface performance is measured

These records come from the same canonical match observations as the headline Racquet Rating, over the same 174-match rating window. Nothing here is a separate rating.

A win rate appears only where a surface has at least 10 matches. Below that a single result moves the percentage by ten points or more, so the record is shown and the percentage is not.

The range beside each rate is a 95% confidence interval. It is wide on purpose: a surface record of a few dozen matches does not pin a percentage down, and the interval is the honest width of that uncertainty rather than a claim of precision.

Matches whose source records no surface are counted in the totals and assigned to no surface. They are never redistributed.

Surface rates are descriptive. They are not combined into a surface rating, and they are not used to rank or compare players — at these sample sizes most players' rates cannot be told apart.

Recent results

This player's record over their last 10 matches in the rating window. A record of what happened, not a prediction.

Last 10 matches: 6 wins and 4 losses, played Apr 2026 – May 2026.

Last 10 matches

6–46 won, 4 lost

Apr 2026 – May 2026

From 173 matches in the rating window, excluding 1 walkover.

How this record is counted

The last 10 matches this player actually played, from the same canonical match data as the headline Racquet Rating. Retirements count — a match was played and a winner advanced. Walkovers do not, because no match took place.

Match dates in the source are tournament start dates, so several matches can share a date. The order is reconstructed from the tournament and the round, which is why the record is shown as a total and not as a sequence of individual results.

The dates above are the span these 10 matches cover. For a player who has not competed recently that span can be well in the past, which is why it is always shown.

This is a record of what happened. It is not a prediction, and it is not the Form Momentum component of the rating, which measures something different.

Highest-ranked opponents defeated

Victories over the highest-ranked opponents in the published rating window, using the opponent's official rank at the time of the match.

The 5 highest-ranked opponents defeated in the published rating window, out of 97 wins over ranked opponents. Ordered by the opponent's ranking at the time of each match, best first.

  1. Defeated Alexander Zverev, ranked number 3 at the time of the match, in April 2026 at Munich · Semifinal.

    Alexander ZverevNo. 3 at the time

    April 2026 · Munich · Semifinal

    6-3 6-3

  2. Defeated Felix Auger Aliassime, ranked number 6 at the time of the match, in May 2026 at Roland Garros · Quarterfinal.

    Felix Auger AliassimeNo. 6 at the time

    May 2026 · Roland Garros · Quarterfinal

    4-6 6-4 6-4 6-4

  3. Defeated Holger Rune, ranked number 9 at the time of the match, in April 2025 at Madrid Masters · Round of 64, opponent retired.

    Holger RuneNo. 9 at the timeRET

    April 2025 · Madrid Masters · Round of 64

    6-2 0-0 RET

  4. Defeated Daniil Medvedev, ranked number 10 at the time of the match, in April 2026 at Madrid Masters · Round of 16.

    Daniil MedvedevNo. 10 at the time

    April 2026 · Madrid Masters · Round of 16

    6-3 5-7 6-4

  5. Defeated Tommy Paul, ranked number 13 at the time of the match, in August 2024 at Cincinnati Masters · Round of 64.

    Tommy PaulNo. 13 at the time

    August 2024 · Cincinnati Masters · Round of 64

    6-2 4-6 7-5

From 97 wins over ranked opponents in the rating window.

How these are chosen

Every win this player recorded in the rating window against an opponent the source ranks, ordered by that opponent's official ranking in the week of the tournament, and limited to the 5 highest. Each opponent appears once, at that player's highest-ranked victory over them.

The ranking shown is the opponent's ranking at the time of that match, taken from the same match record as the result. It is not their ranking today, and not their career-high.

Retirements are included and marked. The match was played and a winner advanced, and it counts toward the rating this profile explains — but it was not completed, so it is not presented as though it had been.

This is a list ordered by one rule: how highly the opponent was ranked that week. It is not a judgement of which victories were the finest, and a higher-ranked opponent does not by itself mean a better performance.

Not yet available

These sections are published only once the underlying verified data exists. Nothing below is estimated or inferred in the meantime.

Performance signals
Performance signals are not published yet. Each one needs a defined trigger, data window and minimum-evidence rule before it can appear here.
Comparable players
Comparable players are not published yet. Ranking players by raw rating distance alone is not a defensible similarity method, so no comparison is shown.