Dashboard / Motorsports / Doc

Motorsports Strategy Spec v2.0

Updated: 2026-04-03 | Status: DEPLOYED — Market Data + F1/NASCAR Scrapers Active

Overview

The Motorsports desk monitors Kalshi prediction markets across 4 series: Formula 1, NASCAR (Cup/Xfinity/Truck), IndyCar, and MotoGP. Edge detection planned via qualifying gap models (F1), track-type split models (NASCAR), and DNF probability adjustments.

Current State

Metric Value
Kalshi snapshots 27,087 across 35 series prefixes
F1 sessions 126 (practice, qualifying, race)
F1 drivers 22 current season
F1 qualifying 22 (latest race weekend)
F1 race results 22 (latest race)
F1 pit stops 29 (latest race)
F1 standings 22 driver + 11 constructor
F1 reliability 24 drivers (DNF rates from 2025-2026)
NASCAR standings 39 drivers
NASCAR race results 335 (10 races)

Market Categories

F1 (9 Kalshi groups)

Race Winner, Podium, Top 5/10, Qualifying/Pole, Fastest Lap, Sprint, DNF/Retirements, Safety Car/Events, Championship

NASCAR (3 Kalshi groups)

Race Winner, Top 3/5/10/20, Championship (Cup/Xfinity/Truck)

IndyCar (2 groups)

Race Winner, Series Champion

MotoGP (2 groups)

Race Winner, Championship (Rider + Teams)

Data Sources

Source Table(s) Rows Schedule Status
Kalshi API sports_odds_snapshots 27,087 Every 30min DEPLOYED
Pinnacle (Odds API) sports_odds_snapshots — Adaptive DEPLOYED
OpenF1 API f1_sessions, f1_drivers, f1_qualifying, f1_race_results, f1_pit_stops 221 Daily 11AM DEPLOYED
Jolpica API f1_standings, f1_constructor_standings, f1_reliability 57 Daily 11AM DEPLOYED
ESPN NASCAR nascar_standings, nascar_race_results 374 Daily 11AM DEPLOYED

Collection Schedule (all ET)

Time What
11:00 AM OpenF1 (sessions, qualifying, results, pits) + Jolpica (standings, reliability) + NASCAR (standings, results)
Every 30min Kalshi price snapshots
Adaptive Pinnacle motorsport odds

De-Vig Methods (Planned)

Market Method Min Edge
Outright winner (20+ field) Power method 5-8%
Head-to-head Multiplicative 3%
Podium/Top 5/10 Monte Carlo (100K sims) 4%
Qualifying/Pole Power method 4%
Fastest Lap Power method 5%
Championship Power method 4%
Sprint Race Power method 4%

Position Sizing (Planned)

Market Kelly Fraction
H2H 1/4
Podium/Top N 1/5
Race winner (F1) 1/6
Race winner (NASCAR superspeedway) 1/8
Championship 1/8 first 5 races, 1/6 after
Fastest lap 1/8
Qualifying 1/5

Key Model Parameters

Parameter Value Source
F1 grid-to-finish R² 0.65 (varies by circuit) Historical
NASCAR qual-to-finish R² <0.20 Historical
F1 pole-to-win rate ~40% Historical
F1 SC probability ~70% of races Historical
DNF base rate 15-20% Rolling 20 races
F1 car vs driver ~80/20 Panel consensus
MotoGP bike vs rider ~50/50 Panel consensus

Future / TODO

  1. Edge scanner — Build motorsport-specific edge scanner using Pinnacle anchor + qualifying model
  2. Circuit database — Track-specific coefficients (overtaking difficulty, SC probability, tire degradation)
  3. Weather model — Dual dry/wet probability system with specialist adjustments
  4. DNF Poisson model — Per-constructor/driver, component mileage tracking
  5. NASCAR track-type models — Separate models for superspeedway/intermediate/short/road
  6. IndyCar scraper — No free API found yet
  7. MotoGP scraper — motogp.com requires paid DataPass
  8. Practice data integration — Long-run pace from FP2/FP3 (OpenF1 laps endpoint)
  9. Lap-by-lap race simulation — F1 with tire degradation + safety car Monte Carlo
  10. WebSocket upgrade — REST to WS for live execution (when desk goes live)
  11. Betfair exchange — Secondary sharp reference (free API, thin liquidity for IndyCar/MotoGP)
Source: ~/edgeclaw/results/spec-panel/sports-desk/motorsports-strategy.md