Game Performance Dashboard

What the evidence says

In professional football, aerobic fitness is the main determinant of match-related physical decline. After the 60th minute nearly all running, acceleration and deceleration zones plus total distance fall by 4–27% — with one striking exception: sprint distance / peak speed does not drop. A player's field lactate-threshold and peak incremental speed correlate moderately-to-strongly (r 0.46–0.67) with how well total distance, HSR and accel/decel are maintained (Aerobic Fitness Is the Main Determinant of Match-Related Decline, Part 2, J Strength Cond Res 2026, PMID 42351367, n=19, one club/season, observational).

The retention index

For each player we take the mean per-minute output in the opening blocks (0–30 min) and the closing blocks (60–90 min) of the same match, and express the late value as a percentage of the early value:

Term	Meaning
HSR retention	closing HSR/min ÷ opening HSR/min
A+D retention	closing (accel+decel)/min ÷ opening
Fatigue-Resistance Index	average of the two, in % retained

100% = output held; below 100% = it faded. Team median is 89%. Peak speed is shown as a separate flat reference — it is not part of the index because it does not decline.

How to use it

Primary (here, Game / post-match): the within-match 15-min data lives here, so the decline is computed here. Secondary (Daily Report, MD+1): show only the resulting flag — a low fatigue-resistance player is a candidate for an aerobic / threshold top-up in the week. Reuse the player baseline from xPO (own baseline > position norm); do not recompute it.

Two cautions

Substitutes look artificially resistant. A player who comes on after minute 60 is fresh, so their "retention" can exceed 100% (e.g. late subs here). Read the index together with minutes / substitution status — it shares that dependency with the substitute-deficit view. Deceleration capacity is not ranked here. Raw max-deceleration comparisons between players are biased by momentum (body mass × approach speed), and body mass is not yet in the export — so this panel ranks a player against themselves over the match, where mass cancels out, not player-vs-player deceleration ability (Deceleration Paradox, Lin et al., J Strength Cond Res 2026, n=80).

What is real vs demo here

Everything shown is real real: computed from the StatSports Game export, Period15 blocks (33 matches, 31 players, 1,856 blocks of ≥10 min, peak speed passed through the >36 km/h plausibility filter). No demo values.