Research-grade machine learning
Models developed through reproducible experiments, forward-looking evaluation and calibration. Review the evidence relevant to your decisions.
Research discipline & performance intelligenceMotorsport intelligence
Research-grade models.
Enterprise-grade delivery.
Race probabilities, rival pit-stop forecasts and performance models for racing teams and trading desks. Review the outputs on a recorded race, then scope an evaluation for your decisions and systems.
Review model outputs with us. Define the data access, deployment, support and service levels your operation needs.
One race. As it unfolds.Recorded model replay
Across our research archiveDiscuss your series
- Formula 1
- IndyCar
- IMSA
- WEC
- Formula E
- ELMS
- Asian Le Mans Series
- Le Mans Cup
- Indy NXT
- NLS
- European Truck Racing
- MotoGP
- Moto2
- Moto3
- MotoE
- WorldSBK
- WorldSSP
How RaceHooks works
From the race
to your next decision.
Research-grade intelligence and enterprise-grade infrastructure, configured for the way your organization operates.
Enterprise ingestion & delivery
Streaming, signed webhooks, retries and delivery logs. Agree throughput, infrastructure and service levels as part of your production scope.
Built for production integrationData security scoped to you
Review data access, confidentiality, retention and deployment requirements before integration. Confirm the controls and responsibilities in your engagement.
Your requirements, your deploymentFeatured capabilities
A closer look at
what we model.
A selection from our modeling work. Race probabilities, strategy and performance intelligence, built for the questions that matter to your team.
FEATURED / 01
Canadian GP / L22Recorded model replay
Price the next move.
Before the pass.
Before the pass.
91.2%ANT → RUSPass within 5 laps
Pass probability
FORECAST / L22Antonelli on Russell
PASS / L22Overtake confirmed
For books & trading desks
Race & overtake probabilities
Price the whole field, from lights out to the final lap. Win, podium, head-to-head and overtake probabilities translate the developing race into a view your trading desk can act on.
Win · Podium · Head-to-head · Overtakes
The distribution is the product.
Picking a winner is only one part of pricing a race. Calibration, probability scoring and the timing of each forecast determine how useful that distribution is to a trading desk.
Maintain your own view of the race alongside your market line.
FEATURED / 02
Canadian GP / L25Recorded model replay
The pit window
is taking shape.
is taking shape.
41.9%VERStop within 3 laps
ANT41.9%
VER41.9%
HAM24.8%
For racing teams
Rival pit-stop forecasts
Read rival intent before the pit entry. Per-car stop probabilities bring the next lap and the coming pit window into the same strategic picture.
Per-car probabilities · Next-lap & multi-lap horizons
A stop forecast needs a horizon.
A rival stopping next lap and stopping within three laps create different strategic choices. Horizon-specific forecasts help frame the risk of an undercut, a cover stop or staying out.
Weigh your next call against the stops your rivals could make.
FEATURED / 03
Singapore GP / L31Recorded model replay
The gap is visible.
The energy isn’t.
The energy isn’t.
Battery charge53%vs47%GAS / HAD0.49s behind
GAS
53%
HAD · ahead
47%
For racing teams
2026 battery forecasts
See beyond the visible gap. Lap-by-lap forecasts of battery charge across the Formula 1 grid add an energy perspective to rival performance under the 2026 regulations.
Charge forecasts · Per-car state · Lap-by-lap updates
Track position is only part of the picture.
Two cars together on track can have different energy available for what comes next. Grid-wide charge forecasts give analysts another dimension when interpreting closing speed, attack and defence.
Read the energy position behind a rival’s next move.
FEATURED / 04
Canadian GP / L18 / AntonelliRecorded model replay
Same race.
A different points outlook.
A different points outlook.
+1.01PTSExpected points vs baseline
Baseline
16.88
Selected plan
17.89
Lap 18
L23SOFTL43HARD
68Expected finishing position3.65
For racing teams
Race strategy
Compare the race you have with the race your next stop creates. Stop timing, alternative stint plans and rival response belong in one decision.
Stop timing · Stint scenarios · Rival response
The fastest stint is not always the best race.
An early stop can buy pace and cost track position. Extending can preserve options and surrender the undercut. Evaluate those trade-offs alongside traffic and how rivals could respond.
Choose between credible alternatives, with the race context attached.
The breadth behind the models
Motorsport intelligence.
Across your operation.
From underlying pace to the next strategic decision. Specialist models bring different dimensions of the race into your research, strategy and trading systems.
Build your intelligence stackRace outcomes & pricing
Price the race as it develops.
Win, podium and head-to-head probabilities. Finishing distributions and expected championship points.
Pace & performance
Understand what the timing screen cannot tell you.
Lap time over expected, driver pace modes, qualifying pace and fastest-lap probability.
Tyres & energy
Read the resources shaping the next stint.
Tyre health, performance-cliff risk, next-compound predictions and battery charge forecasts.
Pit stops & strategy
Evaluate your options alongside rival intent.
Rival stop forecasts, pit windows, stop-count probabilities, undercut and overcut opportunities.
Race dynamics & risk
Account for the events that change the race.
Safety-car and VSC probabilities, overtaking opportunities and per-driver retirement risk.
Applications
Built for your
side of the race.
Different decisions need different outputs. See how the models fit a trading desk or a race team.
Pricing & market analysis
A model line for
your motorsport markets.
Win, podium and head-to-head probabilities for your desk to evaluate alongside its existing prices.
Pre-raceIn-raceFair odds
Delivery for your trading operation
Integrate model probabilities into your pricing and risk systems. Market coverage, permitted use, delivery speed and service levels are defined for your trading engagement.
Model-implied fair oddsRecorded example
DriverWin chanceFair odds
Antonelli67%1.49
Verstappen19%5.26
Hamilton4%25.00
Canadian GP · lap 31 · selected drivers
Fair odds = 1 ÷ win probability. No bookmaker margin.
Working together
From your first question
to a defined evaluation.
A conversation with the modeling team, focused on your use case and what you need to establish.
- 01
Define the question
Tell us your series, the decision you need to support and the systems your team uses.
- 02
Review the model
Inspect sample outputs and recorded race cases. Agree forecast horizons, coverage and the evidence your team needs.
- 03
Agree the scope
Define the evaluation dataset, success criteria and review. Agree production delivery, support and commercial terms separately.
Let’s talk models
Bring us your next
race decision.
Bring a decision and your existing workflow. We’ll identify relevant outputs, review recorded examples and define what an evaluation needs to establish.
Discuss a model evaluation