rán

A lap nobody drove, and it’s quicker than pole.

Every 2025 qualifying session, its fastest pieces sewn into one lap that nobody drove and run as a ghost against pole. Then the check of how small the pieces can get before the stitch is measuring the ruler instead of the drivers.

Telemetry
Signal processing
Python
NumPy
0.16 s
the honest gap
00.5 sthe median session

median that the fastest three official sectors beat pole by. Austria and Mexico: nothing.

1.7 s
mostly error
325100pieces oftelemetrysectors

what the same stitch claims cut into 25 pieces of telemetry.

5 of 8
Verstappen poles
abuazebriitajapmiasauunihatched: someone was quicker in pieces

of Verstappen's poles were beaten, in pieces, by someone's own best sectors.

The idea

Broadcasts love an ideal lap: the best of every sector, summed. Take it further and you get a stitched lap, each corner from whoever took it best, sewn into a lap no car ran, and you can race it against the pole sitter to see where the time was hiding. Rán is Yoruba for sew.

The question that matters is how small the pieces can be. The official timing splits a lap into three sectors with loops in the track, timed to the millisecond. Anything finer has to come from telemetry, the speed and position each car sends about four times a second, and telemetry has an error. Stitch the fastest of many noisy pieces and you collect the error, not the speed.

How it works

  1. Pick 1,360 laps
    Each driver's three fastest clean qualifying laps: timing accurate, not deleted, not rebuilt by FastF1.
  2. Stitch three sectors
    The fastest official S1, S2 and S3 of the session, from the timing loops, summed.
  3. Run it on pole’s line
    Inside each sector the ghost follows the lap that set it, scaled to its official time.
  4. Check the ruler
    Rebuild every official sector from telemetry, then cut finer and watch what the stitch claims.
Qualifying · 2025
Pole against the lap nobody drove
The open ring is the pole lap. The brown dot is the three fastest official sectors of the session, each driven on the trace of the lap that set it and scaled to its timing-loop time, both on the pole lap's line. Drag the bar to scrub, or play it at four times real speed.

Key decisions

The timing loops are the truth

A lap's telemetry adds up to its official time exactly, but where the car is at a given moment is less certain. Rebuilt from telemetry, a sector lands a median 48 ms from the official one, and one in ten is off by more than a tenth (1,336 sectors checked). So the stitch is summed from official sectors only; telemetry gives the shape inside them, never the time.

Line laps up by their own distance

Matching the slowest point of every corner to pole's sounds better and is worse: at four samples a second the slowest point is itself a guess, and against the official sectors it roughly doubled the error. Each lap is placed by its own distance over its own length.

Clean laps only

A sector from a lap deleted for track limits is not one anyone was allowed to drive, so those laps are out, as are laps the timing flagged as inaccurate and three Miami laps whose telemetry ran far longer or shorter than the track. Pole is the pole sitter's Q3 lap; the stitch draws on all of qualifying.

Results

The ruler · 24 qualifying sessions
Cut the lap finer and the stitch gets faster, because it is picking noise
024635102550100pieces the lap is cut intoclaimed over pole, s0.270.451.031.672.743.49median of 24timing loops,0.16 seach sessionevery cut is one morechance to pick noise

Three official sectors put the stitched lap a median 0.16 s ahead of pole. Three equal pieces of telemetry already claim 0.27 s, and 25 claim 1.7.

A pole lap is not hiding a second and a half. Each extra cut is another chance for a few hundredths of measurement error to land in the fast direction, and the minimum takes it every time. The broadcast-style mini-sector ideal lap is mostly that.

Every session · three official sectors
Where the lap nobody drove beat pole
  1. Las Vegas
    pole NOR 1:47.934 · NOR NOR SAI
    0.868
  2. Hungarian
    pole LEC 1:15.372 · PIA PIA PIA
    0.629
  3. Azerbaijan
    pole VER 1:41.117 · PIA SAI RUS
    0.540
  4. Belgian
    pole NOR 1:40.562 · RUS NOR RUS
    0.412
  5. Miami
    pole VER 1:26.204 · VER NOR LEC
    0.331
  6. Emilia Romagna
    pole PIA 1:14.670 · VER PIA NOR
    0.294
  7. Italian
    pole VER 1:18.792 · HUL PIA VER
    0.224
  8. Japanese
    pole VER 1:26.983 · NOR PIA PIA
    0.223
  9. Saudi Arabian
    pole VER 1:27.294 · VER RUS PIA
    0.221
  10. British
    pole VER 1:24.892 · VER VER RUS
    0.204
  11. Singapore
    pole RUS 1:29.158 · RUS RUS VER
    0.194
  12. Canadian
    pole RUS 1:10.899 · VER RUS PIA
    0.185
  1. Monaco
    pole NOR 1:09.954 · PIA LEC NOR
    0.135
  2. Dutch
    pole PIA 1:08.662 · PIA VER PIA
    0.120
  3. Chinese
    pole PIA 1:30.641 · VER PIA PIA
    0.115
  4. United States
    pole VER 1:32.510 · VER ANT NOR
    0.115
  5. Australian
    pole NOR 1:15.096 · NOR VER NOR
    0.082
  6. Abu Dhabi
    pole VER 1:22.207 · VER VER PIA
    0.051
  7. Qatar
    pole PIA 1:19.387 · VER PIA PIA
    0.051
  8. São Paulo
    pole NOR 1:09.511 · NOR NOR NOR
    0.038
  9. Spanish
    pole PIA 1:11.546 · PIA PIA VER
    0.027
  10. Bahrain
    pole PIA 1:29.841 · RUS PIA PIA
    0.013
  11. Austrian
    pole NOR 1:03.971 · NOR NOR NOR
    nothing0.000
  12. Mexico City
    pole NOR 1:15.586 · NOR NOR NOR
    nothing0.000

sorted by how much three official sectors beat pole · after pole: who set S1 S2 S3

Of the 72 sectors, PIA took 22, NOR took 19, VER took 16, RUS took 9. Pole sitters owned all three in Austria, Mexico City, São Paulo, all of them Norris, and none at all in Baku (VER), Hungary (LEC), Japan (VER). In Hungary Piastri set every fastest sector and started second; Q3 ran slower than Q2, and Norris's Q2 lap was half a second quicker than Leclerc's pole. The biggest gaps, Las Vegas, Hungary and Baku, are sessions where the track changed under them.

Give every driver their own stitched lap, their own best three sectors, and in 9 of the 24 sessions someone other than the pole sitter had a quicker one. Five of those were Verstappen poles, out of his eight: Suzuka, Jeddah, Miami, Austin and Baku. The other sixteen poles lost four between them. Verstappen's poles were the ones least likely to have been the fastest car in pieces. They were the laps put together best.

Where it stands

The honest stitched lap is small: a median 0.16 s, and nothing at all in two sessions. The dramatic one is mostly the stitch finding error in telemetry that only knows where a car is to within a few metres. Three sectors is as fine as public data can cut a lap and still be believed.

Two things would sharpen this: the teams' own GPS and timing, which would let mini-sectors mean something, and fuel and tyre state for each lap, since a Q1 sector on older tyres and a Q3 sector on new ones are not the same piece of track.

Everything is computed in research/ran/stitch.py from the 2025 F1 live-timing feed as archived by TracingInsights from FastF1. Sector times are the official ones; the traces inside them are linear interpolations of telemetry in each lap's own distance. The three-lap cap per driver, equal-length pieces and the pole lap as the drawn line are this study's choices.