POLYTRACK, MEASURED
PolyTrack performance: frame rate and lap times
PolyTrack is a browser game inside somebody else’s page. The frame it runs in, and whether you are in fullscreen, is not cosmetic: on our machine the same track went from a 1:29 lap to a 0:27 lap once the page chrome was gone. This page is what we can and cannot conclude from that.
The three runs
| Run | Where | Lap | Speed at the line | When |
|---|---|---|---|---|
| #1 | CrazyGames page, game in its frame, page chrome visible | 01:29.748 | 251 km/h | 2026-11-02, 14:34 |
| #2 | fullscreen, no page chrome | 00:27.198 | 251 km/h | 2026-11-02, 14:49 |
| #3 | fullscreen, second attempt on the same track | 00:26.583 | 314 km/h | 2026-11-02, 14:50 |
TESTED
All three are the same track (Summer 1), the same browser and the same machine, with the lap time read off the HUD at the finish line. The first run had just started the session; the other two were within a minute of each other.
What the client shows you
IN-GAME
The HUD gives a lap clock to the millisecond and a speed readout in km/h, both live. It does not print a frame rate, a frame time, or a build number — so a dip in performance has to be inferred from the lap clock, not read directly.
Two explanations, both still open
- The car is a different car. The first run ended at 251 km/h, the third at 314 km/h. A large part of the 3.3× gap could be the racing line rather than the frame rate.
- The simulation is tied to the render rate. If the physics steps once per drawn frame, a page that draws slowly moves the car slowly while the lap clock keeps counting real seconds. That would make frame rate a lap-time factor, which is worth knowing before trusting any published record.
NEEDS CHECK
We have not separated these two. Doing it needs one controlled test: same track, same driver, two runs back to back — one windowed, one fullscreen — with the start and finish of each photographed.
Why it matters for the records on this site
Every lap time we publish is a number somebody else cannot reproduce unless the conditions are stated. If the frame rate moves lap times, then a record set in a windowed tab and one set in fullscreen are not the same measurement, and the site has to say which one it is quoting. That is the whole reason this page exists before the track pages fill up.
The controlled test we owe this page
| Step | What we do | What it settles |
|---|---|---|
| 1 | same track, windowed, two laps | the spread of a single condition |
| 2 | same track, fullscreen, two laps | whether the mode moves the clock |
| 3 | repeat both on a second track | whether it is track-specific |
| 4 | write the frame budget into /tracks/ | so a record carries its context |