Temporal trends and weight-class differences in takedown density in professional mixed martial arts: a retrospective analysis of 8,461 UFC bouts, 1997–2025

Abstract Background Takedowns are a central grappling action in mixed martial arts (MMA), but long-term evidence on time-normalized takedown output across competitive eras and weight classes remains limited. This study examined takedown density (TDpM; successful takedowns per minute) as a descriptive performance indicator in professional MMA. Methods We analyzed publicly available UFC bout-level records from 1997 to 2025, comprising 8,461 bouts and 16,922 fighter–bout observations. TDpM was calculated as successful takedowns divided by bout duration in minutes. Temporal trends were examined separately for men and women, with sensitivity analyses excluding bouts shorter than 90 s and excluding pre-2002 men’s observations. Weight-class differences were evaluated using Kruskal–Wallis tests with Dunn–Holm post hoc comparisons. Division-specific percentile values were calculated as descriptive TDpM benchmarks, and winner–loser differences in TDpM, takedown success rate (TD%), and control time per successful takedown (CTRL/TD) were examined exploratorily. Results Men’s annual mean TDpM declined significantly across 1997–2025 (β = −0.00426 takedowns·min⁻¹·year⁻¹, r = − 0.753, p < 0.001), and the negative direction remained in the 2002–2025 sensitivity analysis (β = −0.00202, r = − 0.712, p < 0.001). Women’s TDpM showed no significant trend across 2013–2025 (β = −0.00252, r = − 0.394, p = 0.183). Men’s TDpM differed significantly across weight classes (Kruskal–Wallis p < 0.001), although effect sizes were small (ε² ≈ 0.005–0.007); women’s between-division differences were not significant. Winners showed higher TDpM (median 0.067 vs. 0.000 takedowns·min⁻¹), TD% (50.0% vs. 20.0%), and CTRL/TD (99.75 vs. 78.88 s) than losers. Conclusions TDpM provides a simple, time-normalized descriptive indicator of takedown-related performance in professional MMA. Men’s TDpM declined over time and differed modestly across weight classes, whereas women’s patterns were comparatively stable in the available sample. Percentile benchmarks and winner–loser contrasts offer descriptive reference values, but TDpM should not be interpreted as a causal determinant of victory or a training-prescription threshold. Future studies incorporating video-coded context and training data are needed to clarify its broader applied value.