This browser could not create a WebGL2 context. The engine, the rules and every number in the Lab panel below still work — only the three-dimensional view is missing.
What the engine actually measures
The claim under test
The Fosbury flop is widely said to work because it lets the jumper’s centre of mass pass under the bar — the body draping over in sequence, so the whole body never has to be above the bar at once. This app does not assume that. It gives the jumper published segment masses, lengths, centre-of-mass positions and radii of gyration (de Leva 1996), lets each historical technique adopt its own shape inside plausible joint ranges, and computes where the centre of mass goes.
Both halves of the folklore are wrong in opposite directions. At its anatomical optimum the modelled flop clears a bar 4.6 cm above the peak of its own centre of mass — so the centre of mass really can pass under the bar, and this engine says by about five centimetres. But not one of the eleven world-class floppers World Athletics measured at the 2018 World Indoor Championships did it. All eleven carried the centre of mass above the bar, by 2 cm to 12 cm, mean 7 cm. The folklore describes a geometric possibility, not what elite high jumping does.
Every technique, same centre-of-mass peak
Each style is bisected to the millimetre for the highest bar it can clear from an identical centre-of-mass arc, then checked against its own definition (how horizontal the trunk is, which way it faces, how the legs split) so the search cannot quietly turn a scissors into a flop. Positive efficiency means the bar cleared is above the peak of the centre of mass.
The measured jumps
The eleven finalists of the 2018 World Indoor Championships, as published by World Athletics. H3 is the peak height of the centre of mass; H3 diff is H3 minus the bar they cleared.
The jumper
Provenance
This build declares 77 constants and modelling
choices: 51 DOCUMENTED, 8 DERIVED, 5 MEASURED,
2 CALIBRATED and 11 RECONSTRUCTED — 66.2 % documented,
14.3 % reconstructed. The list is in js/provenance.js and in
CREDITS.txt; the page harness recounts it from the shipped file and fails if this
paragraph disagrees.
The rules this app implements
Quoted from World Athletics, Book C — C2.1 Technical Rules, the official PDF (amendments to 22 December 2021, in force 1 January 2022). Rule numbers are the current TR numbering; the High Jump was Rule 182 before 2019.
- TR26.2 — three consecutive failures
- “Three consecutive failures, regardless of the height at which any of such failures occur, disqualify from further jumping / vaulting except in the case of a tie for first place.” An athlete may pass their second or third trial at a height after failing, and still jump at a later height.
- TR26.6 — how a height is measured
- “measurements shall be made, in whole centimetres, perpendicularly from the ground to the lowest part of the upper side of the bar.” The bar sags, so the announced height is the top of the bar at mid-span. Jump nearer an upright and there is up to 2 cm more bar to clear — the app models that.
- TR26.8 — the count-back
- First, “the athlete with the lowest number of jumps at the height last cleared”. Then “the lowest total of failures throughout the competition up to and including the height last cleared”. Failures at heights above an athlete’s best do not count — a qualifier most popular summaries of this rule leave out. Still level: share the place, unless it is first.
- TR26.9 — the jump-off
- One jump each per height, starting at the next announced height above the height last cleared. If more than one clears, the bar goes up 2 cm; if all fail, it comes down 2 cm.
- TR27.1 / TR27.2 — what is a failure
- Take off from one foot. It is a failure if the bar leaves the supports through the athlete’s action, or if the athlete touches the ground beyond the vertical plane through the nearer edge of the bar without first clearing it, or touches the bar or the uprights while running up without jumping.
- TR26.2 — a pass does not reset the count
- Three consecutive failures eliminate, and a pass taken between them does not break the run. The Rule’s own worked example shows it: athlete A fails at 1.91, passes the rest of that height, then fails twice at 1.94 — three consecutive failures — and takes no further trial. The first version of this app’s test suite asserted the opposite; the engine was right and the test was wrong.
- TR25.17 — the clock
- 1 minute with more than three athletes left, 1.5 minutes with two or three, 3 minutes with one, and 2 minutes for consecutive trials.
Where the published rules do not decide
About this build
High jump is a real athletics event governed by World Athletics; this is an independent, unofficial simulation of it. It is not affiliated with, endorsed by, or connected to World Athletics or any national federation, and no World Athletics material is redistributed here beyond the rule text quoted for reference.
What this model is not
This is a static statement in the page, not something a script writes. If every script on this page failed to load you would still be reading it.
- The flight is solved in one plane — the plane perpendicular to the bar. The jumper’s body is placed in three dimensions and may lie at any angle to the bar, but the rotation in flight is constrained to the bar axis. Real jumpers also twist about their own long axis, and this model cannot.
- The run-up and take-off are not simulated. The jump begins at toe-off with a centre-of-mass velocity and an angular momentum; how a real athlete generates those from a curved approach is a separate and harder problem.
- Every joint-angle trajectory is reconstructed. Published biomechanics gives postures at instants, not the angle-by-angle path between them. The five techniques’ shapes were found by a bounded numerical search, not copied from a motion-capture file.
- Air resistance is ignored, and the landing is not simulated at all — which matters, because the documented reason the flop displaced everything else is that deep foam landing areas arrived, and that is a property of the facility, not of the flight.
- The rival jumpers are invented. Their names, personal bests and consistency are not modelled on anybody.
- The technique comparison holds the centre-of-mass arc fixed. Dapena notes the flop also gives back roughly what it gains through a lower centre of mass at take-off; that trade-off is outside what this model measures.