Methodology

No tipsters, no vibes. ApexIO is a simulation model: it plays every Grand Prix out thousands of times and reports what actually happens across those runs, as probabilities. When it's a coin-flip, it says so.

Ingestion

Pace from every session — plus the facts the data can't see.

We build each car's performance picture from every timed session — practice, qualifying and races, not just Sundays — so even a thin new-regulation season has signal. Each driver is separated from their car, and every circuit's character is modelled. The hard facts — grid penalties, stewards' calls, driver swaps — come straight from official FIA documents, never headlines.

Every session

feeds the pace index

Simulation

The race, run ten thousand times over.

Every Grand Prix is simulated thousands of times. Each run lets the things that actually decide races move the way they do in reality — qualifying spread, reliability, weather and safety cars — with each driver's uncertainty set by how predictable they truly are. You get the full range of how it could go, not a single guess.

10,000+

of race sims

Through the weekend

A forecast that sharpens as the weekend unfolds.

The call is re-issued at four points: Thursday on form and news, after Friday and Saturday practice as real pace lands, and again once qualifying sets the grid. Each reading is stamped the moment it's published — so you can watch confidence build, and see what the grid is worth: the History page shows winner accuracy before the weekend and after qualifying, side by side. On sprint weekends the middle readings come from Sprint Qualifying and the Sprint instead of Friday and Saturday practice.

Thu → Quali

four stamped readings

Measured honestly

Stamped forecasts, scored against reality — and the market.

Nothing is edited after lights out. If a call has to be reissued before the race — a late grid penalty, the FIA's final starting grid — it goes out as a new revision, with its time, its reason and a fingerprint of the numbers, and the earlier version stays on the record. Races are tagged live (posted before they ran) or backtest (walk-forward: predicted using only data from before the race). The benchmark that matters is the betting market's closing line — can we be right more often than the odds say we should be?

Every reissue

on the record

The honesty rule

We publish the forecast before the cars roll out and never touch it after lights out. Every post carries a lock stamp: the time, a SHA-256 fingerprint of each driver's win and podium odds, and a link to the post before it. Change a single number and the fingerprint changes, along with every one after it. Races we forecast live are kept apart from the walk-forward backtest, and there's no invented history: backtest races show only the two stages we actually scored. It's a harder test, and it's the only one worth trusting.

See it in action

Open the live forecast →