Nordic Hamstring Asymmetry Did Not Track Landing Force Asymmetry in 31 Pro Players

Thirty-one professional soccer players were tested on the Nordic hamstring device and on force plates during jump landings. The two measures of left-right imbalance barely correlated, which means a clean Nordic score says almost nothing about landing asymmetry.

Landing asymmetry and eccentric hamstring imbalance are routinely treated as two readings of the same underlying problem, and strength staff often screen one as a proxy for the other. A study published in Frontiers in Physiology on 6 November 2026 put that assumption to a direct test in a professional squad, and found the two numbers were effectively unrelated.

The research team, drawn from four universities in Türkiye, recruited 31 outfield players from a professional league. Goalkeepers were excluded. The group averaged 23.13 years of age, 178.61 cm in height and 76.11 kg in body mass, and comprised six center backs, five fullbacks, twelve midfielders, four wingers and four forwards.

Each player completed two assessments. Peak eccentric hamstring force was captured on a NordBord sampling at 500 Hz, and peak landing vertical ground reaction force was captured on dual ForceDecks platforms sampling at 1000 Hz after a countermovement jump. Both measures were normalized to body mass, and both showed strong test reliability: an intraclass correlation of 0.91 for Nordic peak force and 0.86 for landing force.

The design is cross-sectional, with no prospective injury tracking. What it measures is how load is distributed between limbs on a given day, not who subsequently gets hurt.

Nordic Force and Landing Asymmetry, Side by Side

On the Nordic device, the dominant limb produced 4.59 N/kg against 4.45 N/kg for the non-dominant limb, a difference that did not reach significance (p = 0.056) and carried a trivial effect size of 0.168. On landing, the figures were 37.49 N/kg and 36.13 N/kg respectively, again short of significance (p = 0.068) with an effect size of 0.135.

Averaged across the squad, Nordic imbalance came in at 8.95 percent with a standard deviation of 6.21, while landing asymmetry averaged 11.50 percent with a standard deviation of 7.84. Vertical stiffness averaged 31.42 kN/m, countermovement jump height 38.45 cm, and landing rate of force development 101.32 N/s.

Those group means look unremarkable. The problem appears only when the two measures are placed side by side at the level of the individual player.

How Landing Asymmetry Is Calculated

Both imbalances were expressed through a bilateral asymmetry index using absolute magnitude, which converts the gap between limbs into a single percentage without recording which side was stronger. For landing asymmetry, the inputs are the peak vertical ground reaction forces registered by the left and right force plates as the player absorbs a jump.

Absolute indexing has a known trade-off. It makes squad-level summaries tidy, because a 12 percent imbalance reads the same whether the left or the right limb dominates, but it discards direction. The authors addressed that by repeating the analysis with signed values, which preserves side, and the landing asymmetry correlation remained flat at rho = 0.041.

Where the Landing Asymmetry Correlation Collapsed

The relationship between Nordic imbalance and landing asymmetry was rho = 0.028, with a confidence interval spanning -0.330 to 0.378 and a p-value of 0.881. Using signed rather than absolute values changed nothing: rho = 0.041, p = 0.826. Nordic imbalance against vertical stiffness fared no better, at rho = 0.185, p = 0.320.

A correlation of 0.028 is not a weak relationship. It is the absence of one. Knowing a player sits at 15 percent on the Nordic device gives a practitioner no usable information about where that same player sits on landing asymmetry, and the reverse is equally true.

Risk Categories Disagreed Just as Badly

The authors also compared the two tests the way a practitioner would actually use them, as pass or fail screens. The Nordic cut-off was set at 10 percent, drawn from prospective strain-injury research. Landing asymmetry was banded as normal at or below 10 percent, moderate risk between 10 and 15 percent, and high risk above 15 percent.

Of the 13 players flagged as high risk by the Nordic threshold, seven landed in the normal band for landing asymmetry. Running the comparison the other way, four of the 18 players who passed the Nordic screen recorded landing asymmetry above 15 percent.

Formal agreement statistics confirmed the mismatch. Fisher exact test returned p = 0.682, and weighted kappa came in at -0.042, a value fractionally below the zero that represents pure chance. This is a different kind of result from the graded imbalance findings reported in work on isokinetic strength imbalance and hamstring injury in rugby, where a single modality was tracked against outcomes.

Why the Two Tests Should Not Be Expected to Agree

Mechanically, the mismatch is less surprising than it first appears. The Nordic test isolates one muscle group through a slow, controlled eccentric action with the hip and knee held in a fixed relationship. A jump landing is a fast, multi-joint event in which the ankle, knee and hip share load, and in which the nervous system has milliseconds to distribute force.

A player can compensate for a weaker hamstring by shifting work to the quadriceps or the calf on landing, and nothing in a Nordic score would reveal that. The reverse holds too: a player with balanced hamstrings can still show marked landing asymmetry for reasons of technique, prior injury or simple preference. Context on the strength side of this question appears in a trial comparing Nordic hamstring work against razor curls in footballers.

What 31 Players Can and Cannot Settle

The sample is the clearest constraint, and the authors say so directly. No a priori sample size calculation was performed. Post-hoc analysis indicates the study had 80 percent power to detect correlations of 0.48 or larger, which means a genuine small-to-moderate association could have been missed.

That caveat cuts both ways, though. The confidence interval around the headline correlation runs from -0.330 to 0.378, so while a modest true relationship cannot be excluded, a strong one can. Whatever connects Nordic imbalance to landing asymmetry, it is not large enough to justify treating one as a stand-in for the other. A squad screening pathway built on that substitution would misclassify a meaningful share of players in both directions.

What This Means for Screening Practice

The practical reading is narrow but useful. If a program screens limb balance at all, this study suggests that one test does not substitute for another, and that a clean result on the Nordic device should not be read as clearance on landing asymmetry.

It is worth being equally clear about what the study does not establish. It did not follow players forward, so it cannot say which measure predicts injury, or whether either does. It is a statement about agreement between tests, not about risk.

General, non-individualized points for coaches and athletes:

• Treat isolated strength testing and dynamic landing testing as separate pieces of information rather than two views of the same imbalance.
• A single screening number, from any device, is a snapshot of one day and one task, not a standing property of an athlete.
• Pain, swelling or a change in movement quality is a matter for a qualified clinician, regardless of what a force plate reports.

Readers interested in how dynamic jump testing behaves under fatigue will find relevant context in a scoping review of countermovement jump fatigue monitoring, and the rehabilitation angle is covered in a meta-analysis of hamstring rehabilitation and return to play.