Limb asymmetry: how to calculate it, which threshold to use, and the traps hiding in the data

Left–right asymmetry is the number every testing report carries and the one most often calculated three different ways inside the same department. Before any threshold means anything, the formula has to be fixed, the weaker side has to be named, and the data has to be clean. This guide covers those three things in that order. It applies to any paired measure: force plate jumps, dynamometry, hop tests, range of motion, change of direction.

Three formulas, three different numbers

Take a single-leg test where the right leg produces 400 N and the left 340 N. Depending on the formula, the same two legs are reported as:

The first two are the same information read from opposite ends: 15 % asymmetry is 85 % LSI. The third always gives a larger number than the first for the same legs, and the difference grows with the size of the gap. None is wrong. Mixing them is: a 15 % threshold agreed on one formula silently becomes a different threshold on another.

Name the weaker side, and check which way 'weaker' points

A percentage without a side is half a finding. '15 %' tells the coach how big the gap is; '15 %, left weaker' tells them where to look. The catch is that weaker does not always mean the smaller number.

• For force, power, jump height, strength and range of motion, the weaker side is the one with less.
• For contact time, time to complete a turn, balance sway and knee valgus, lower is better, so the weaker side is the one with more.
• A gap that rounds to 0 % names no weaker side. Printing 'left weaker, 0 %' invites a conversation about nothing.

kinexis · single-leg jump

Click to enlarge

Thresholds: 10 and 15 are conventions, not laws

The familiar bands, under 10 % balanced, 10–15 % worth watching, above 15 % worth acting on, come largely from return-to-play work, where the question is whether an injured limb has caught up with the other one. They are a reasonable starting point and a poor finishing point, for three reasons.

• Asymmetry is metric-specific. A 12 % gap in jump height and a 12 % gap in landing rate of force development are different findings with different noise.
• Healthy athletes are not symmetrical. In a fit squad, gaps of 10 % and more are common on many metrics and often stable for years.
• Noisy metrics produce big gaps by chance. Rate-of-force-development measures can differ by 20 % between legs inside a single test with nothing wrong.

The useful reading is therefore the band, the direction and the trend together: is the gap in the red band, is it always the same leg, and has it grown across the last three sessions? A red gap that flips side between sessions is noise. An amber gap on the same leg three sessions running is a pattern.

Four traps in the data

Most surprising asymmetries are not athletes. They are data. These four are the ones we have had to deal with in a professional squad's testing history.

• The zero leg. A single-leg trial that records exactly 0 on one side is a failed trial, not a leg with no force. Left in, it reads as a 100 % asymmetry and can drag an athlete's whole symmetry score into the red.
• The mis-read trial. When an athlete steps off a force plate instead of landing on it, the software records several seconds of 'flight' and a jump height of metres. Any jump whose flight-time height is above 100 cm should void every number of that trial.
• Values that sit near zero. Knee valgus of 2° on one side and 4° on the other is a 50 % 'asymmetry' and a clinically identical pair of knees. For angles that live near zero, read the absolute value against a fixed band and do not calculate a percentage at all.
• Rounding before dividing. Calculate the gap from the full-precision values and round once at the end. Rounding the legs first can move a 9.6 % gap across the 10 % line.

From many gaps to one symmetry score

An athlete tested across jumps, strength and movement has dozens of left–right pairs. Averaging them lets one noisy metric dominate. The median is the better summary: half the athlete's gaps are below it and half above, and a single wild value cannot move it. Expressed as 100 minus the median gap, it becomes a score where higher is better, which is easier to put beside other scores.

From one athlete to a squad

Ranking a squad by asymmetry is useful for finding who to look at first. It hides one thing: the most balanced athlete in the room can simply have two weak legs. A ranking by gap should always show each leg's own level beside it, so a balanced but weak athlete is not mistaken for a good result.

kinexis · squad rankings · single-leg

Click to enlarge

A checklist

• One formula for the whole department, written on the report.
• The weaker side named, with the direction checked for lower-is-better measures.
• Bands per metric family, not one number for everything.
• Zero legs and mis-read trials removed before any percentage is calculated.
• No percentages for angles that sit near zero.
• Band, direction and trend read together before anyone acts.

How Kinexis handles this

Kinexis calculates every left–right gap the same way across force plate, dynamometer, movement and sprint testing, names the weaker side on each one, and applies the cleaning rules above before a number is shown. The gaps then roll up into one symmetry score per athlete and a squad ranking, with the written interpretation saying which side the findings agree on.

• One pill per pair: the gap and the weaker side, green under 10 %, amber to 15 %, red above.
• Zero legs and mis-read trials left out by rule, so a failed rep never becomes a red flag.
• No percentage for knee valgus or trunk lean; those are judged on the angle itself.
• A symmetry score per athlete (100 minus the median gap) and a squad view ranked by gap, each leg in its own colour.