R03 — JAPANESE GRAND PRIX

Safety car at Suzuka

41%

41% probability that at least one safety car is deployed at Suzuka. High-speed corners increase incident risk. SC deployed in both R01 (3 VSC) and R02 (1 SC).

41%Safety car at Suzuka
46%41%64%

41% [46%64%] (95% CI)

01 / KEY FACTORS

Impact by factor

R01-R02 SC rate (100%)
+10.8%
Suzuka high-speed corners
+7.5%
2026 PU reliability issues
+6.8%
22 cars on the grid
+4.1%
Run-off improvements
-3.5%
Degner curves risk
+3.2%
March weather uncertainty
+1.8%
Impact total+30.7%
02 / MONTE CARLO SIMULATION

Distribution and quantitative analysis

10,000 simulations Monte Carlo

Position Distribution

10,000 simulations

Distribution Statistics

N Simulations10k paths
ExpectedP1.4
MedianP1
ModeP1
Std Dev0.49
Skewness+0.49 →
VaR 95%P2
CVaR 95%P2
Prob(P1)62.0%
Prob(Podium)100.0%
Prob(Points)100.0%
Prob(DNF)3.0%

Convergence

Scenario Analysis — Safety car at Suzuka

ScenarioPositionProb.Conditions
Best case (P5)2+ SC28%Rain or mixed conditions, multiple incidents in heavy braking zones
Optimistic (P25)SC + VSC42%At least one safety car plus virtual safety car from mechanical failure
Base case (P50)1 SC62%Standard safety car deployment from first-lap incident or debris
Conservative (P75)VSC only18%Only virtual safety car from isolated retirement on track
Worst case (P95)No SC8%Clean race, no incidents, all retirements in pit lane or gravel traps

Sensitivity Analysis

03 / FEATURE ANALYSIS

Variable analysis

Base value5.0%
Output41%
5.0%
41%
circuit_sc_rate+12.4%
braking_zone_severity+8.7%
pu_reliability_avg+6.3%
grid_size+4.1%
technical_sections+3.5%
rain_probability-3.8%
track_temperature-2.1%
Feature importancei
circuit_sc_rate
0.195
braking_zone_severity
0.151
pu_reliability_avg
0.122
grid_size
0.098
track_temperature
0.085
rain_probability
0.078
technical_sections
0.065
drs_zone_count
0.052
altitude_correction
0.038
season_round
0.028
Feature correlationsi
sc_rate
braking
pu_reliabil…
grid_size
rain_prob
temp
sc_rate
braking
pu_reliabil…
grid_size
rain_prob
temp
-1.0
0
+1.0
04 / MODEL VALIDATION

Model performance

Model performancei
Accuracy0.710+0.130 vs baseline
Brier Score0.041-0.006 vs baseline
Log Loss0.612
AUC-ROC0.785
Calibration curvei
0%0%20%20%40%40%60%60%80%80%100%100%PredictedActual
Confusion matrix
Predicted vs Actual75% accuracy
SC
No SC
Actual
SC
42
8
No SC
12
38
Pred.
Track characteristicsi
Top SpeedDownforceTyre WearBrakingOvertakingPU Demand

Accuracy

75%

across 127 race winner predictions

Average Brier score

0.041

probability calibration

Last correct

China R2

China R2 — SC deployed (Stroll stranded)

05 / CIRCUIT ANALYSIS

Circuit factors

Suzuka history

Last 4 races

2019

SC

2023

No SC

2024

SC + VSC

2025

SC

Practice pace

Gap to leader (sec)

Tyre degradation

Lap time, laps 1-20

Lap 1Cliff ~L12Lap 20
06 / EVOLUTION & DATA

Probability over time

Training Data Sample4 rows x 8 cols
#racecircuitsc_deployedvsc_deployedretirementsraingrid_sizeprediction
1AUS 2026Albert Park203No2265%
2AUS 2025Albert Park114No2058%
3CHN 2024Shanghai112No2061%
4CHN 2019Shanghai103No2056%