UFC database coverage and division benchmarks
By Verdikt AI · Published
What can this database support?
The current release contains 3,048 fighter profiles. Of those, 2,775 carry an available UFC striking aggregate; 273 do not. That gap matters: listing a professional MMA athlete does not mean we have UFC performance observations for that person. The database includes historical competitors and athletes from other promotions, rather than only today's UFC roster.
2,985 profiles have histories marked complete in this release; 63 are not marked complete. Completeness is the dataset's reconciliation status, not independent certification of every source. A full career record can include more eligible bouts than a partial table. UFC-only totals and career totals must remain separate.
Browse the underlying fighter profiles · Definitions and correction process
How the benchmark sample is selected
We select profiles with a UFC aggregate marked complete, at least five covered UFC bouts, at least 900 seconds of covered fight time, and available striking data. 1,255 profiles meet those conditions. These thresholds reduce dependence on extremely brief appearances; they do not remove selection bias or make the sample representative of a future fight.
Profiles are grouped by their listed division, with a trailing “Division” label normalized. The label may reflect a later division than some of the bouts behind the career aggregate. Divisions need at least ten eligible profiles to appear below; 12 eligible profiles fall outside those displayed groups. No fighter is moved to a division by guessing from weight or name.
Median rates within the selected sample
Each fighter contributes one available rate to each metric. The median is the middle fighter value, or the mean of the two middle values for an even sample. This is an unweighted median of fighter career aggregates, not a pooled event rate. Each cell shows its own denominator because a missing takedown value must not become zero.
SLpM is significant strikes landed per minute; SApM is significant strikes absorbed per minute. TD/15 is takedowns landed per 15 minutes. Measured zeros are retained. Unavailable, nonnumeric, negative and nonfinite observations are excluded from that metric. Rates use their published units and are rounded to two decimal places only after the median is calculated.
| Listed division | Eligible fighters | SLpM median (n) | SApM median (n) | TD/15 median (n) |
|---|---|---|---|---|
| Bantamweight | 110 | 3.51 (n=109) | 3.54 (n=109) | 1.48 (n=109) |
| Catchweight | 51 | 3.17 (n=51) | 3.05 (n=51) | 1.31 (n=51) |
| Featherweight | 119 | 3.56 (n=119) | 3.31 (n=119) | 1.20 (n=119) |
| Flyweight | 51 | 3.31 (n=51) | 3.24 (n=51) | 1.18 (n=51) |
| Heavyweight | 145 | 3.09 (n=144) | 3.07 (n=144) | 1.08 (n=144) |
| Light Heavyweight | 103 | 3.39 (n=98) | 3.38 (n=98) | 0.98 (n=98) |
| Lightweight | 178 | 3.68 (n=178) | 3.59 (n=178) | 1.32 (n=178) |
| Middleweight | 151 | 3.16 (n=149) | 3.11 (n=149) | 1.27 (n=149) |
| Welterweight | 185 | 3.31 (n=185) | 3.22 (n=185) | 1.17 (n=185) |
| Women's Bantamweight | 39 | 3.78 (n=39) | 3.51 (n=39) | 0.87 (n=39) |
| Women's Flyweight | 55 | 3.88 (n=54) | 3.77 (n=54) | 0.96 (n=54) |
| Women's Strawweight | 56 | 3.77 (n=55) | 4.07 (n=55) | 1.10 (n=55) |
Use a benchmark as context
A fighter above the landed-strike median recorded a higher rate than the middle available fighter in this selected division group. That does not establish superior opposition, greater damage, better defense or an advantage in an upcoming matchup. Absorbed strikes require their own interpretation: a lower rate can reflect style, distance, grappling time or the opponents faced.
Read covered bout count and time beside a rate. A five-bout sample is less informative about consistency than a long history, and both can hide changes in style or competition. Compare the same units and scope, then read individual bouts. Do not combine a professional MMA win total with a UFC-only denominator or compare a full-career aggregate with a single recent fight as if they were equivalent.
Reproduction and limitations
The linked JSON contains the exact aggregate counts, metric denominators, thresholds and SHA-256 digest of the profile inputs used for this table. It is generated from the same release manifest used to publish the fighter profiles. The table recalculates when the dataset is rebuilt; this is a description of published data, not a prospective research experiment.
Published MMA records and UFC bout statistics supply the observations; Verdikt AI computes this sample and its medians. Source corrections, incomplete historical coverage, division changes and survival in the promotion can change or bias the result. This analysis does not adjust for era, opponent strength or schedule, estimate causal effects, or validate model probabilities. Report disputed observations with the fighter page and supporting source so that both the profile and derived table can be checked.
Download the benchmark counts and denominators (JSON) · UFCStats published statistics · Report a sourced correction
