Skip to main content

​
Overview

Predicted Race History takes the same gap-to-leader-average chart as the Race History widget and adds one dashed, faded line per driver: the engine’s predicted remainder of the race, joined onto that driver’s actual trace at the playhead.
The engine models no traffic and no lapped cars. A predicted line can cross a real one; that is expected, not a bug.

​
What It Shows

Actual traces joined to the engine's predicted remainder at the playhead (Madrid 2026, lap 43) The base chart (axes, gridlines, actual driver traces) is byte-identical to Race History, built by the same function, and it zooms the same way: right-drag a box to zoom both axes, right-click without moving for a Reset zoom menu, no sliders (see Race History). Each predicted line:
  • starts at the point where that driver’s actual trace currently ends (the join point), so there is no visible gap between actual and predicted;
  • carries the same styling an actual trace would: the driver’s team colour, the per-team solid or dashed line type, an end-of-line driver tag, and a reduced opacity;
  • stops drawing at the last lap the engine judged reachable, if any later lap in that driver’s predicted series carries the engine’s unreachable sentinel;
  • is named by the driver alone, never “predicted”, so the always-on legend can key off a real driver name.
Before lights out. When the playhead sits before lap 1 and the session carries a lap-0 pre-race snapshot, this chart draws the predicted curves alone against a constant reference that puts the predicted winner at zero seconds on the final lap, with a caption naming the loaded race data version. There is no actual trace to join yet. See Pre-Race and What-If Studies. The pre-race view: predicted curves alone, before any lap has been timed (Madrid 2026) The x-axis widens to the full race distance so a predicted point beyond the currently hydrated lap range is never clipped. A caption under the chart states the clear-air assumption plainly, and, when the underlying snapshot is stale, an as-of stamp: “Data as of lap N (playhead lap M)”.

​
Data Behind It

The predicted lines come from the latest strategy_snapshot’s race_history rows, at or before the playhead lap (the as-of rule). model_id strategy_snapshot, model_version the composite engine_version+WRAPPER_VERSION, over vpw:v1:{session_key}:insight:strategy_snapshot, hydrated via REST on session enter and rehydrate, then live. Each predicted point is re-baselined into the same gap frame the actual traces use: y(lap) = cumulative_time(lap) - sum(baseline over laps 1..lap), the identical transform (cumulative lap time minus a rolling baseline) the base Race History chart applies to real laps. Rows for a lap at or before the playhead are skipped: they are already covered by the actual trace up to the join point, and drawing them again would draw a predicted point backwards from it. A row at or above the engine’s 1e300 unreachable threshold truncates that driver’s predicted series at the last reachable lap; this is sentinel decoding, done once in the shared derivation layer, never smoothing.

​
States

Shows the shared empty-state text: the exact-lap status gate and reasons where known, or a generic “No strategy data at lap N.”
When an as-of snapshot exists but the race has not yet produced a completed lap to plot, shows “No completed laps yet.” instead of an empty chart.
Mirrors the base Race History widget’s own honest-empty state: one missing measured lap time anywhere would silently corrupt the running cumulative sum, so the whole widget shows “Laptimes missing” and plots nothing, actual or predicted.
The chart still renders, with an as-of stamp under it: “Data as of lap N (playhead lap M)”.
Outside a race, renders “Session is not a race”; the app never computes a strategy bundle for a non-race session.

​
FAQ

The engine’s remainder prediction assumes clear air: no traffic, no lapped cars. A predicted line crossing another driver’s actual or predicted line is expected under that assumption, not an error.
The engine judged every later lap in that prediction unreachable from the current race state (its DBL_MAX sentinel). The line truncates at the last lap it could still complete.

​
Next Steps

Race Strategy Tool

The shared engine, derivation layer, and sentinel rule behind all six strategy widgets.

Race History

The real, actual-only version of this chart.

Predicted Pit Window and SC

Each driver’s next predicted stop, plus safety-car guidance.

Strategy Status

The gate, reasons, and whole-race health strip that explains a blank or stale chart.
Last modified on October 15, 2026