Models / gradient-boosted-ensemble

Fastest Lap Probability

P(this driver holds the race's fastest lap at the flag).

gradient-boosted-ensembleanalytics.race-odds

How it works

A field-relative model scores each driver's chance of holding the race's final fastest lap from the laps completed so far, then L1-normalises the scores so the field sums to 1. It is updated every lap. A driver with no green-flag lap yet starts at zero.

When to use it

F1 Fantasy: fastest lap still scores in the official game even though it no longer earns a championship point.
Display or price who is likely to set the fastest lap, including the late-race soft-tyre attempts teams make for it.

Inputs

pace vs field
tyre freshness
laps remaining
position context

Output

drivers[].fastestLapProbabilityField-normalised probability (sums to 1 across the field), rounded to 5 decimals.

Example payload

An excerpt of one analytics.race-odds delivery, trimmed to the fields this model produces. Values are illustrative. The full message shape is on the feed reference.

json

{
  "feed": "analytics.race-odds",
  "lap": 35,
  "drivers": [
    {
      "number": "1",
      "tla": "VER",
      "fastestLapProbability": 0.27841
    },
    {
      "number": "4",
      "tla": "NOR",
      "fastestLapProbability": 0.21307
    }
  ]
}

Use it

Subscribe to analytics.race-odds with a webhook and read this model's fields from each delivery:

bash

curl -X POST https://api.racehooks.io/v1/webhooks \
  -H "Authorization: Bearer $RACEHOOKS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"webhookUrl":"https://your-app.com/f1","feedId":"analytics.race-odds"}'

To try it without waiting for a race weekend, replay the recorded Canadian Grand Prix (2026-canada_r) from the quickstart.

How to read it

The values form one distribution over the field, so compare drivers against each other. Early in the race, whoever currently holds the fastest lap keeps it far less often than late in the race — expect the distribution to be spread out early and to concentrate as laps run out.

Limitations

The current holder of the fastest lap is much less secure early in a race than late, so early values move a lot.
Races shortened by a red flag are handled slightly differently from how the model was trained.
Published in Race and Sprint sessions only; absent for drivers who were not scored.

FAQ

Do fastest lap probabilities add up to 1?

Yes. Scores are normalised across the field so fastestLapProbability sums to 1. If only one driver has been scored, that value is passed through as is.

Does fastest lap still matter without a championship point?

It no longer earns a World Championship point, but it still scores in the official F1 Fantasy game.

Related models

Qualifying Pace — Per-sector pace deltas vs the session median, per driver.
Pace Normalization (LTOE) — Lap Time Over Expected — the normalized pace delta, per lap.
DNF Risk — P(this driver retires before the flag, given still running), per lap.

Get the data

This model is delivered on analytics.race-odds. Subscribe by webhook or stream it over SSE. The quickstart walks through registering an endpoint and replaying a past session against it.

← All models