CMJ Numbers Dropped But the Athlete Doesn't Feel Fatigued? Here's How to Read It

Monday morning: an athlete jumps 34.2 cm on the force plate. Wednesday, same warm-up, same shoes, same plate — 32.1 cm. Roughly a 6% drop. The athlete says they feel great, slept fine, nothing unusual in the last 48 hours of load. This is exactly where a lot of monitoring programs split into two bad habits: wave the number off because the athlete feels fine, or treat it as proof of fatigue and pull the session back before anyone checks what actually moved underneath it.

Jump height by itself can't settle that argument. It's an output number, and the same centimeter reading can come from a fast, efficient movement or a slow, compensated one — plus a chunk of any single-session change is plain measurement noise rather than anything physiological. The way to actually answer the question is to stop staring at height and look at movement time, eccentric duration, and force development on the same trial. This guide walks through separating real neuromuscular fatigue from measurement error and ordinary day-to-day variation, using the strategy metrics underneath the jump instead of the jump height itself.

Height Is an Output, Not a Process

A countermovement jump is a chain: eccentric braking, a transition at the bottom of the dip, then concentric drive to takeoff. An athlete carrying genuine neuromuscular fatigue often compensates by dipping deeper and taking longer to develop force, and can still land close to their usual height by trading time for output. From the outside, the number on the screen barely moves. Underneath it, movement time can be up 40-70 milliseconds and eccentric duration up by a similar margin — and that's the part that actually reflects the athlete's condition, not the centimeters.

The Reverse Problem Is Just as Common

A single jump height reading also carries real trial-to-trial noise. Footwear, plate temperature, verbal cueing, warm-up length, even which foot contacts the plate a fraction of a second first, can shift height by a few percent with no change in the athlete's underlying state. Chase every dip in height without checking anything else, and a program ends up reacting to measurement noise about as often as it reacts to real fatigue — cutting load an athlete didn't need cut, or worse, missing genuine fatigue because the height held up while the strategy underneath it changed.

1. Is It Inside Normal Measurement Noise?

Every CMJ metric carries a baseline of trial-to-trial variability from testing alone. Jump height typically shows a coefficient of variation in the range of 4-6% across sessions even under well-controlled conditions. A drop sitting inside that band isn't a finding, it's noise. The standard approach in applied sports science is to treat a real change as roughly double the typical error — for jump height that puts a believable threshold closer to 9-10%, not 3-5%.

2. Did Anything About the Test Itself Change?

Before touching the training plan, walk back through what actually happened at testing: different shoes, a colder gym, a rushed warm-up, hands on hips versus free arm swing, a new staff member cueing the jump differently, or a software update that shifted the movement-onset threshold. Any one of these can move movement time by 20-40 milliseconds on its own, which is enough to look like fatigue on paper.

3. What Do the Strategy Metrics Actually Say?

If the test conditions were clean and the drop still holds, look at movement time, eccentric duration, time to peak force, and RSImod together. One of these shifting slightly proves little. All of them moving in the same direction, past their own noise band, on a jump height that barely changed, is a much stronger case for real fatigue than height ever is on its own.

Gathercole et al. (2015): Height Missed It, the Timing Variables Didn't

Gathercole, Sporer, Stellingwerff, and Cronin (2015, International Journal of Sports Physiology and Performance) tested a battery of CMJ variables in team-sport and combat-sport athletes before and after a fatiguing training protocol. Jump height changed only slightly pre-to-post — a small, largely trivial effect that on its own would not have flagged meaningful fatigue in most monitoring setups. Timing-based variables told a different story: eccentric duration and total movement duration lengthened with moderate effect sizes, a clearer and more consistent signal of the fatigued state than height produced. The authors' core conclusion was that a battery built around jump height risks missing fatigue that timing-based variables pick up. The limitation worth flagging: the fatiguing protocol was a controlled training bout, not accumulated in-season fatigue across a competitive block, so the size of the effect may not transfer directly to a long in-season monitoring program.

Claudino et al. (2017): The Pattern Holds Across Studies, With Caveats

Claudino and colleagues (2017, Journal of Science and Medicine in Sport) pooled results across multiple studies examining which CMJ variables actually track neuromuscular fatigue. Jump height showed only a small-to-moderate pooled effect for detecting fatigue, while variables built around movement strategy and flight-time-to-contraction-time ratios showed larger, more consistent effects across the pooled studies. The meta-analysis also flagged substantial heterogeneity between the underlying studies — different fatigue protocols, different populations, and different testing equipment — so the exact effect size shouldn't be treated as a fixed number to apply to every team. The direction of the finding is the reliable part: height alone underperforms as a fatigue indicator relative to the metrics built around timing and force development.

Cormack et al. (2008): Why the Noise Threshold Matters

Cormack, Newton, McGuigan, and Doyle (2008, International Journal of Sports Physiology and Performance) established reliability data for CMJ variables across single and repeated jumps, reporting the coefficient of variation figures that underpin the noise thresholds used above. Their work is the reason a 5% height drop shouldn't automatically be read as fatigue — it's the reason applied practitioners compare any change against a measurement-error baseline before acting on it. The limitation here is straightforward: reliability figures come from healthy, rested athletes performing familiar protocols, so day-to-day noise in a fatigued or unfamiliar athlete may run higher than these baseline figures suggest.

What to Pull Off the Same Trial

None of this requires a separate test. A force plate sampling at 1000 Hz, or a validated wearable, already captures everything needed from the same countermovement jump used to measure height:

• Movement time — from the first drop in vertical force below bodyweight to takeoff. A lengthening here with stable height is the earliest and most common sign of a compensated jump.
• Eccentric duration — the braking phase alone, from the countermovement's onset to the lowest point of the dip. Fatigue tends to show up here before it shows up anywhere else in the trace.
• Time to peak force — how quickly the athlete reaches maximal force production during the concentric phase. A delay of more than roughly 15-20% versus baseline, on a jump height that hasn't moved much, is a stronger fatigue signal than the height itself.
• RSImod — jump height divided by movement time. Because it combines both numbers into one ratio, it tends to move before either number alone looks alarming on its own. See the full RSI-modified breakdown for the formula and normative ranges.

Build a Baseline Before You Judge a Session

None of these thresholds mean much against a single prior session. Run 3 trials per test, take the best jump by height and note its full profile, and build each athlete's own baseline over 4-6 sessions before comparing a new number against it. A published normative range tells you where an athlete sits in a population; it doesn't tell you whether today's number represents a real change for that specific athlete.

The Numbers

Reading It

Height alone would sit right at the edge of noise, easy to explain away, especially with an athlete who reports feeling fine. But movement time, eccentric duration, and RSImod all moved together, all past their own noise thresholds, all in the same direction. That's the pattern the research above points to: the athlete braked slower and took longer to build force, and largely made up the difference on the way up, which is exactly why the height held up better than the underlying mechanics did. Subjective readiness — how an athlete feels — and objective neuromuscular output can genuinely disagree, particularly with accumulated fatigue that hasn't yet produced soreness or perceived tiredness. This is a case worth trimming plyometric or high-velocity volume for that session, not because the height dropped, but because three independent strategy metrics agree it should.

Five Steps Before You Act on a CMJ Change

1. Confirm the test was clean. Same surface, same footwear, same threshold settings, same hand position as baseline. If any of these changed, re-test before drawing a conclusion.
2. Compare height against the athlete's own noise band, not a published norm. A change under roughly 8-10% for height is not yet a finding.
3. Pull movement time, eccentric duration, and RSImod from the same trial. Check whether they moved in the same direction as height, or independently of it.
4. Require agreement across at least two strategy metrics past their own noise thresholds before flagging fatigue. One metric moving alone is more often noise than signal.
5. Weigh it against subjective report, but don't let it overrule the data outright. An athlete feeling fine with three metrics in agreement is still a fatigue flag worth a lighter session; it's a case for a conversation, not a dismissal.

Where Monitoring Programs Actually Go Wrong

• Reacting to height alone. A 5-6% single-session drop sits inside typical noise for most testing setups. Reacting to it every time trains the coaching staff to ignore the metric within a few weeks.
• Ignoring a drop because the athlete says they feel fine. Subjective readiness and objective neuromuscular fatigue overlap only partially. Athletes routinely underreport accumulated fatigue that hasn't yet turned into soreness.
• Comparing against a population norm instead of the athlete's own baseline. A movement time of 0.58 seconds might be completely normal for one athlete and already a 10% jump for another. Norms tell you where someone sits in a population, not whether today's number changed for them specifically.
• Skipping the equipment check. A software update, a new pair of shoes, or a colder testing room can move timing variables by 10-20% on their own, with zero physiological cause behind it.
• Acting on one session. Even a genuine multi-metric flag deserves a same-week re-test before it drives a real programming decision, unless the pattern is severe or paired with other red flags like disrupted sleep or unusually heavy recent load.