The Gravity Race / Design workshop
Find a fast wire
that fits.
The finish is 6 m across and 2 m down. Set your depth and wire limits. See which checked shape gets there soonest.
Checking the finite shape catalogue. Controls, results and saved links unlock when ready. A linked setup stays in the address if loading fails.
Set your limits
How far the wire may dip. The finish is 2 m down.
Measure the whole curve from start to finish.
Set exact limits in metres
One finite family, checked numerically. This is not a search over every possible wire.
Same endpoints / same physical clock
Your wire and the preview
△ Your wire A● Preview B
Depth is measured down from the start. The shaded area is below your allowed space. This front view shows the ideal wire paths, not a real apparatus.
Playback runs at half speed. The clock and arrivals show physical seconds. Motion starts only when you ask.
Result for these limits
Your comparison track stays yours
Shape wire A.
Switch A to a reference or move the shaping handle. Use the preview button by the limits to try B as A. Undo restores the wire you replaced.
Inspect the preview in the original 3D race ↗These move a shaping handle, not a point the bead passes through. Only the actual wire must fit. Manual controls are finer than the search grid, so your curve can be quicker than its grid result.
Keep the known alternatives in view
Why not just use
the fastest ideal curve?
The cycloid is the least-time curve for these endpoints without these extra limits. It gets the same gravity as A. Here it must also fit the space and available wire. Wider limits can make it the right checked choice.
Speed comes from height lost. An early drop can build speed sooner, but it can require extra depth or wire. At the common finish, every track has lost the same 2 metres and crosses at the same speed. Earlier arrival and higher finishing speed are different claims.
Inspect the search family, numerical margins and physics
A finite design search.
The editable quadratic Bézier curves use the original normalized controls: horizontal 0 to 0.85 of the span, vertical 0.15 to 2.50 of the finishing drop. The search steps each control by 0.05, producing 18 × 48 = 864 curves. Manual controls step by 0.01. All-shapes scope also admits the original straight and cycloid references. Restricting the scope is an explicit design choice.
Times are approximate numerical integrals of path length divided by speed. Speed is √(2gh), with height lost h positive downward. The start-from-rest singularity is removed by the original q = u² substitution. The native inverse-time table supplies positions for the common physical clock. The bead is a constrained point particle: no rolling, friction, drag, loss of contact or real-track safety assessment.
For custom vertical control k in whole percentage points, the deepest point is the 2 m finish when k ≤ 100; otherwise its depth is k²/[100(k−50)] metres. The depth limit uses exact integer cross-products. Length comparisons reserve 10⁻⁷ m; cases within that numerical margin stay unresolved rather than being silently admitted. Cycloid depth uses the same margin. This is a conservative decision convention, not a formal error certificate.
Estimated Earth-like times within 10⁻⁷ seconds of the lowest estimate are grouped as a numerical tie. The fixed catalogue order selects a representative. All times then scale by √(9.81/g), preserving the ranking and geometry. The clock stops each marker at its finish; a finish speed is its speed at crossing.
The unchanged native physics retains its earlier independent reference evidence. A separate 50-digit Decimal/Gauss scalar calculation checks the expanded 866-shape set; that expanded check was written after model inspection and is not a blinded comparison. It checks this declared mathematical family, not an actual device.
Inputs and the inspected clock can be saved. Reopening recalculates the search and starts paused; session-only Undo is not included. JSON results are not trusted on import. The catalogue CSV records all 866 shapes, their numeric metrics, selected scope and limit decisions.
The original race retains its 3D apparatus, detailed equations, primary sources and previous-attempt comparison.