Data-driven F1 fantasy team optimizer
Turn Friday practice sessions into race weekend team picks. Our model analyzes FP1, FP2, and FP3 lap times to predict race positions, estimate fantasy points, and build optimal teams within the $100M budget.
Get Started — Free Plan AvailableRace Prediction Model
Optimized session weights trained on every 2025 race. FP3 best lap, FP2 pace, long-run degradation, and compound-adjusted times feed a calibrated prediction engine with driver bias correction.
Fantasy Team Optimizer
Searches 1.4 million team combinations to find the best 5 drivers + 2 constructors within budget. Four modes — Balanced, Safe, Aggressive, Value — with boost chip recommendations.
Season Tracking
Save your team each race weekend, log actual scores, and track your performance across the full 2026 season. Compare your picks against the model's optimal selections.
Chip Strategy Simulator
Replay the 2025 season with different chip timing and transfer strategies. Find the optimal chip deployment across Wildcard, 3X Captain, No Negative, and Limitless.
How It Works
Real-Time Practice Analysis
Every practice session is ingested and analyzed automatically. Raw pace, race-run performance, and tire-adjusted times are all factored in.
Race Simulation
Thousands of race scenarios are simulated incorporating qualifying, strategy, incidents, and weather to build a full picture of likely outcomes.
Calibrated Scoring
Fantasy point estimates are grounded in a full season of historical data, not guesswork. The model learns and adapts as the 2026 season progresses.
Optimized Team Selection
Every valid team combination is evaluated against the $100M budget to find the highest-value lineup for your strategy.
Our Methodology
F1 Pitwall is built on a simple premise: Friday practice sessions contain real signal about Sunday's race. Here's how we turn raw lap times into winning fantasy teams — without human guesswork.
Practice Data Ingestion
As each practice session completes, we automatically ingest every timed lap. Raw times are normalized for tire compound differences and fuel loads, so a lap on mediums can be fairly compared to one on softs. Both single-lap pace and multi-lap race-run degradation are captured.
Session-Weighted Blending
Not all practice sessions are created equal. Later sessions are more representative of race pace than earlier ones. Our model applies separate optimized weights for back-to-back and standalone weekends — covering both best-lap pace and long-run degradation across FP1, FP2, and FP3 — to produce a single blended performance score per driver.
Driver Bias Correction
Some drivers consistently outperform their practice pace on race day, while others do the opposite. Our calibration engine tracks each driver's historical practice-to-race delta and adjusts predictions accordingly. New drivers start with pre-season baselines that adapt quickly once real races begin.
Monte Carlo Simulation
A single predicted finishing order isn't enough. We run thousands of race simulations incorporating qualifying variance, first-lap incidents, safety cars, weather, and strategy divergence. The result: a full probability distribution for each driver's points — not just a point estimate, but a range.
Fantasy Point Calibration
Predicted positions are mapped to expected fantasy points using a calibration table built from historical scoring data. This captures non-obvious dynamics like overtake bonuses, qualifying streaks, and constructor pairing effects that a simple position-to-points lookup would miss.
Budget-Constrained Optimization
Finally, the optimizer evaluates every valid 5-driver + 2-constructor combination within the $100M budget — over a million teams per race — to find the lineup that maximizes expected points for your chosen strategy (Balanced, Safe, Aggressive, or Value).
2025 Season Backtest
Before going live for 2026, we backtested the model against the entire 2025 season. For each race, the model used only practice data available before qualifying — no hindsight — to pick a fantasy team. Here are the results.
| Race | Predicted Team | Pts | Best | Capture |
|---|---|---|---|---|
| Australian GP | Russell, Sainz, Albon, Tsunoda, Hulkenberg Mercedes + Williams | 146 | 212 | 69% |
| Chinese GP S | Albon, Tsunoda, Stroll, Hulkenberg, Ocon McLaren + Mercedes | 197 | 252 | 78% |
| Japanese GP | Gasly, Albon, Bearman, Ocon, Hulkenberg McLaren + Mercedes | 162 | 196 | 82% |
| Bahrain GP | Hadjar, Gasly, Ocon, Hulkenberg, Bearman McLaren + Ferrari | 237 | 264 | 90% |
| Saudi Arabian GP | Antonelli, Doohan, Bearman, Hulkenberg, Ocon McLaren + Ferrari | 172 | 202 | 85% |
| Miami GP S | Antonelli, Hadjar, Hulkenberg, Doohan, Bearman McLaren + Mercedes | 189 | 226 | 84% |
| Emilia Romagna GP | Verstappen, Sainz, Hadjar, Bearman, Hulkenberg McLaren + Williams | 122 | 242 | 50% |
| Monaco GP | Tsunoda, Sainz, Hadjar, Bearman, Hulkenberg McLaren + Red Bull | 208 | 222 | 93% |
| Spanish GP | Alonso, Hadjar, Lawson, Hulkenberg, Bearman McLaren + Red Bull | 244 | 251 | 97% |
| Canadian GP | Albon, Alonso, Hadjar, Bearman, Hulkenberg McLaren + Mercedes | 254 | 260 | 98% |
| Austrian GP | Alonso, Lawson, Albon, Hulkenberg, Bearman McLaren + Ferrari | 208 | 258 | 81% |
| British GP | Albon, Lawson, Bearman, Antonelli, Hulkenberg Red Bull + Ferrari | 202 | 280 | 72% |
| Belgian GP S | Ocon, Lawson, Gasly, Bearman, Hulkenberg McLaren + Red Bull | 192 | 215 | 90% |
| Dutch GP | Piastri, Lawson, Bearman, Ocon, Hulkenberg McLaren + Williams | 180 | 264 | 68% |
| Italian GP | Hamilton, Piastri, Hulkenberg, Bearman, Ocon Ferrari + Williams | 175 | 218 | 80% |
| Azerbaijan GP | Hamilton, Albon, Leclerc, Bearman, Ocon Ferrari + Williams | 178 | 228 | 78% |
| Singapore GP | Hadjar, Ocon, Bearman, Hulkenberg, Albon McLaren + Red Bull | 183 | 230 | 80% |
| United States GP S | Verstappen, Hulkenberg, Albon, Bearman, Ocon McLaren + Audi | 146 | 221 | 66% |
| Mexico City GP | Tsunoda, Stroll, Ocon, Bearman, Hulkenberg McLaren + Ferrari | 196 | 241 | 82% |
| São Paulo GP S | Piastri, Hulkenberg, Stroll, Bearman, Ocon McLaren + Aston Martin | 291 | 316 | 92% |
| Las Vegas GP | Verstappen, Stroll, Bearman, Ocon, Hulkenberg Mercedes + Williams | 185 | 201 | 92% |
| Qatar GP S | Tsunoda, Hulkenberg, Bearman, Stroll, Ocon McLaren + Mercedes | 150 | 232 | 65% |
| Abu Dhabi GP | Verstappen, Bearman, Hulkenberg, Ocon, Tsunoda Mercedes + Haas | 213 | 256 | 83% |
Capture rate = percentage of maximum possible points the model's team earned. Transfer penalties (-50 pts total) are included in actual scores. All 23 races shown including 6 sprint weekends (marked S). The model starts cold at Round 1 and improves as the season progresses — the first-race capture rate of 62.2% reflects the challenge of predicting without any 2025 calibration data.
2026 Season — Weekly Accuracy
Live accuracy tracking for the 2026 season. Each week we publish the model's predicted team before qualifying, then report actual results after the race. Full transparency — no retroactive edits.
Australian Grand Prix
March 8, 2026 — MelbourneModel's Pick (Balanced)
Actual Result
Tough opener — Bottas DNF wiped 40 pts of value (scored -20 vs replacement driver's +20). 4 DNFs total (BOT, HAD, HUL, PIA) made this a chaotic race. Best team: Russell, Antonelli, Bearman, Lindblad, Colapinto + Ferrari + Racing Bulls.
Chinese Grand Prix
March 15, 2026 — Shanghai (Sprint Weekend)Preview
Sprint weekend — model uses FP1 (30%) + Sprint Qualifying (70%) to predict race results. Fewer practice sessions means higher uncertainty but sprint quali is a strong signal. Price changes from AUS: watch for BOT/HAD/HUL/PIA drops after DNFs, and Russell/Antonelli/Bearman rises after strong finishes. Predictions go live after Sprint Qualifying on Friday.
Predictions update automatically after each practice session. Check the Dashboard for live picks once Friday data is in.
Frequently Asked Questions
How does F1 Pitwall predict race results?
F1 Pitwall analyzes FP1, FP2, and FP3 practice session lap times from every race weekend. The model uses optimized session weights — split between back-to-back and standalone weekends — combining best-lap pace and long-run degradation with compound-adjusted times, driver bias correction, and post-qualifying blending to predict race finishing positions and estimate fantasy points.
How accurate is the prediction model?
The model achieved an 80.6% capture rate across all 23 races of the 2025 season backtest, meaning it captured 80.6% of the maximum possible fantasy points using only practice data available before qualifying.
How does the team optimizer work?
The optimizer searches all valid combinations of 5 drivers and 2 constructors within the $100M budget — approximately 1.4 million team configurations — to find the lineup with the highest expected fantasy points. It offers four modes: Balanced, Safe, Aggressive, and Value.
Is F1 Pitwall free?
F1 Pitwall offers a free tier with race predictions and basic team recommendations. Pro ($3.99/mo) unlocks the full team builder, all optimization modes, and season tracking. Elite ($7.99/mo) adds the strategy simulator, Monte Carlo confidence data, and API access. See all plans.
What data sources are used?
Practice session lap times from all FP1, FP2, and FP3 sessions, sourced from publicly available timing data. The model also incorporates tire compound data, long-run pace, qualifying results, and historical driver performance bias calibrated across multiple seasons.
When should I check F1 Pitwall each weekend?
The best time is after FP3 on Saturday (or after FP1 + Sprint Qualifying on sprint weekends). That's when the model has the most data and predictions are most accurate. Team picks are updated as each session completes.
Which F1 fantasy game does this work with?
F1 Pitwall is built for the official F1 Fantasy game (fantasy.formula1.com) using the 2026 scoring rules, budget system, and chip mechanics. The team optimizer respects the $100M budget cap, 5-driver + 2-constructor roster, and all scoring categories.
Ready to win your F1 fantasy league?
Sign in to access race predictions for free, or upgrade for the full competitive edge.
View Plans