> ## Documentation Index
> Fetch the complete documentation index at: https://offthepace.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Where did a driver's race actually go?

> The per-lap decomposition summed across an entire race, turning average-per-lap components into total seconds gained or lost to each cause.

A driver finishes ten seconds behind but to what? "How the Race Was Lost" cumulates the per-lap decomposition identity across the whole race for one driver, turning average-per-lap components into total race contributions. The result reads as a budget: across the full race, how many seconds did dirty air cost, how many did the tyres take, and how many were down to the driver?

## What it shows

For each component the chart sums the per-lap value across all classified laps:

* **Components**: fuel, tyres, ambient, constructor, dirty air, and driver skill each summed across the race versus the per-lap field median.
* **Negative total**: the driver/car *gained* time from that factor versus the field median.
* **Positive total**: the driver/car *lost* time from that factor.
* **Driver skill**: the residual after removing all physics components the pure driving contribution.
* **Track noise**: unexplained lap-to-lap variation, shown as informational; it is **not** additive to the seven-term identity and should not be added into the total.

<Note>
  Laps under safety car, virtual safety car and red flag never reach this sum: `fct_lap_residuals` is built from green racing laps only, so the exclusion happens upstream in the warehouse rather than in this query. Major-outlier laps are excluded in the query itself. The totals reflect representative racing laps only.
</Note>

## How to use it

1. **Select a race and driver** from the filter bar; the cumulative budget loads.
2. **Read the largest bars first** they are where the race was won or lost.
3. **Separate car from driver** a large constructor bar is the machinery; the driver-skill bar is the part the driver owns.
4. **Treat track noise as a diagnostic**, not a cause it sits outside the additive identity.

## Reading the signal correctly

These are totals **relative to the per-lap field median**, so a positive dirty-air bar means this driver spent more of the race in traffic than the median car it is a comparison, not an absolute physical loss. The six physics components plus driver skill close the identity; track noise is shown separately on purpose.

<Warning>
  Do not add track noise into the seven-term total it is non-additive and would double-count. See the [seven-term identity](/decomposition/seven-term-identity) for why the identity closes on six environmental terms plus driver skill.
</Warning>

## Data source

The chart queries `fct_lap_residuals` via DuckDB-Wasm in the browser, summing across a single race.

| Column                                        | Meaning                                                  |
| --------------------------------------------- | -------------------------------------------------------- |
| `fuel_component_s`, `compound_component_s`, … | The six environmental components, summed across the race |
| `driver_skill_residual_s`                     | The driver contribution after physics is removed         |
| `track_unexplained_s`                         | Lap-to-lap noise, shown as informational (non-additive)  |

This is a **single-race** view across **2018–2024** data.

For the full model definition, see the [`fct_lap_residuals` reference](/reference/models/fct/fct_lap_residuals).