lapbox.tyres

TyreDegradationModelclass

TyreDegradationModel(*, min_laps=5, age_column='tyre_age', use_fuel_correction=True)

One straight line per compound over every driver: fuel-corrected lap time against tyre age. The slope is the seconds lost per lap of tyre age.

Parameters

min_laps: int = 5

Fewest laps a compound needs to be fitted (and at least two tyre ages).

age_column: str = 'tyre_age'

Column with tyre age in laps; falls back to stint_lap, then FastF1’s TyreLife.

use_fuel_correction: bool = True

Take the fuel effect off each lap when there is a fuel load to read.

Reads LapBox’s engineered laps (compound, lap_time_seconds, tyre_age or stint_lap, fuel_load_est with lap_number) or FastF1’s laps (Compound, LapTime, TyreLife, LapNumber). With FastF1’s laps, pit laps are dropped and the fuel load is the laps still to run; clean the laps first with LapCleaningPipeline. Fit one session at a time.

The fuel effect is fixed at FUEL_EFFECT_S_PER_LAP, not fitted: tyre age and fuel both move with the lap number, so a fitted fuel term couldn’t be told from wear. Only laps within 107 % of each driver’s best are fitted. Fitting each race’s first half and predicting the second half on 65 races of 2024–2026, this cut the median error from 0.79 to 0.64 s and was better in 71 % of races; how lap times change inside a stint was predicted no better.

TyreDegradationModel.fit(laps)

TyreDegradationModel.compounds -> dict[str, CompoundDegradation]

TyreDegradationModel.degradation_rate(compound)

TyreDegradationModel.predict_time_loss(compound, tyre_age)

TyreDegradationModel.summary()

CompoundDegradationdataclass

class CompoundDegradation

The fitted line for one compound.

Fields

compound: str

The compound.

degradation_s_per_lap: float

Seconds lost per lap of tyre age. Can come out negative on low-wear circuits; clamp it before planning with it.

intercept_s: float

Lap time on a new tyre with the fuel all burned (s).

fuel_coef: float | None

The fuel correction applied, in seconds per unit of fuel: per kg with fuel_load_est, per lap still to run with FastF1’s laps. 0 when the load never falls (a fuel_load_est that stays constant, as LapBox’s does in practice and qualifying); None when no correction was made.

r2: float

R² of the fuel-corrected lap times against tyre age. It no longer counts the fuel burn as explained, so it is low; read it as how much tyre age explains.

laps: int

Laps fitted.

stint_degradationfunction

stint_degradation(laps, total_laps=None)

Fuel-corrected pace and degradation for every driver’s every stint.

Parameters

laps: pd.DataFrame

FastF1’s laps of one race; stints are FastF1’s Stint numbers.

total_laps: int | None = None

The race distance in laps; defaults to the highest lap number.

Returns

pd.DataFrame, one row per stint with at least three counted laps.

Columns

driver: str

The driver.

team: str | None

The team.

stint: int

FastF1’s stint number.

compound: str | None

The compound on the stint’s first lap.

laps: int

Laps counted: timed, not pit laps, within 107 % of the stint’s best raw lap.

median_corrected: float

Median fuel-corrected lap (s).

deg_per_lap: float

Slope of the corrected times against lap number (s/lap): the tyre falling away.

compound_summaryfunction

compound_summary(stints)

How each compound behaved across the field, kindest-wearing first.

Parameters

stints: pd.DataFrame

stint_degradation’s rows.

Returns

pd.DataFrame, lowest median degradation first. Stints with no compound are left out.

Columns

compound: str

The compound.

stints: int

Stints on it.

laps: int

Laps counted across them.

median_deg: float

Median of the stints’ slopes (s/lap): one ruined stint doesn’t define a compound.

best_deg: float

Lowest slope.

worst_deg: float

Highest slope.