When the model called “Move coming,” who actually made the move?
Race:
HUL
Top converter — Move coming
Nico Hulkenberg
50%conversion rate
When the model flagged Nico as “Move coming” — gap under 0.7s with DRS — they turned 1 of 2 opportunities into actual position gains. More than half the time, the model was right.
2
high-conf opps
9
total tracked
“Move coming” conversion rate
Made/Called
1
HUL
Hulkenberg
Nico
50%
1/2
2
HAM
Hamilton
Lewis
40%
2/5
3
BOR
Bortoleto
Gabriel
33%
4/12
4
BOT
Bottas
Valtteri
33%
1/3
5
ANT
Antonelli
Kimi
25%
2/8
6
LEC
Leclerc
Charles
20%
4/20
7
PER
Perez
Sergio
19%
6/32
8
LIN
Lindblad
Arvid
16%
5/32
9
OCO
Ocon
Esteban
13%
4/30
10
ALO
Alonso
Fernando
13%
2/15
11
LAW
Lawson
Liam
12%
3/25
12
COL
Colapinto
Franco
11%
2/19
13
RUS
Russell
George
10%
2/20
14
BEA
Bearman
Oliver
10%
1/10
15
VER
Verstappen
Max
8%
3/40
16
GAS
Gasly
Pierre
7%
2/28
17
HAD
Hadjar
Isack
7%
3/44
18
PIA
Piastri
Oscar
5%
2/39
19
ALB
Albon
Alexander
4%
1/28
20
NOR
Norris
Lando
0%
0/2
“Move coming” = model confidence ≥ 60%. Triggered when gap to car ahead drops below ~0.7s with DRS eligibility — the ingredients the XGBoost model treats as a high-probability attack scenario. Data from Barcelona-Catalunya Grand Prix. Small sample sizes per race are normal — overtake opportunities are rare and the model only fires when conditions are genuinely met.
Your product. Every race. Every driver.
Overtake probability, tire cliff prediction, pit stop timing — all delivered via webhook during every live F1 session.