Cycling race predictor Β· Giro d'Abruzzo

Stage 3 β€” 2026-09-17

The same production stack priced under 1 course readings, against the bookmaker's pre-stage line. Each scenario is one opinion about what kind of day this is; the results section settles the argument.

1 scenarios 123 riders priced models retrained d911de60f7 book snapshot no capture

The fork

1 readings of the same road

Course profile is an input the model can't see from a map β€” it's set from reference races (or overridden by hand). Different readings redistribute the same probability mass across rider types.

Hindsight (post-hoc)

Hindsight (post-hoc) β€” oracle-matched references
sprint0.00
mountain0.18
hilly0.00
rolling0.81
break0.00
TT0.00
keyed: split sprint (res_5) (board order by rider-type retrieval; prices unchanged)
RETROSPECTIVE ONLY β€” no pre-race prediction was made. This board uses the settled result as the oracle and the nearest historical realized race shapes. It is corpus completion, not evidence of forecasting skill.
Top 5 (blend): Godon Dorian 10.8%; Vauquelin KΓ©vin 8.9%; Crescioli Ludovico 8.4%; Berasategi Xabier 6.2%; Fancellu Alessandro 5.0%

The board

Win probabilities, scenario by scenario

Blend win probability per scenario (cascade Γ— LambdaRank, DNF-adjusted at the pricing layer), next to the devigged Kambi (Unibet) win line captured the morning before the stage. "Max edge" compares the most favourable scenario against the book β€” treat it as scenario-conditional, not a free bet.

Read "Max edge" as directional, not as a price. The blend is a set of per-rider marginals and does not sum to one book β€” across this field each reading totals Hindsight (post-hoc) 90% β€” while the devigged book column sums to exactly 100%. Putting the two side by side flatters the model, and Max edge is inflated by about that factor: divide a rider's probability by his reading's total before calling anything an edge. Of our heads only the LambdaRank one is natively book-preserving.

RiderTeamHindsight (post-hoc)Book oddsBook devigMax edge
Godon DorianNetcompany INEOS10.8%n/a––
Vauquelin KΓ©vinNetcompany INEOS8.9%n/a––
Crescioli LudovicoTeam Polti VisitMalta8.4%n/a––
Berasategi XabierEuskaltel-Euskadi6.2%n/a––
Fancellu AlessandroMBH Bank Ballan Telecom Fort5.0%n/a––
Gaviria FernandoCaja Rural-Seguros RGA3.7%n/a––
Castellon JanCaja Rural-Seguros RGA3.4%n/a––
Colnaghi LucaBardiani-CSF 7 Saber3.2%n/a––
Turconi FilippoBardiani-CSF 7 Saber2.7%n/a––
Ulissi DiegoXDS Astana Development Team2.6%n/a––
Rondel MathysTudor Pro Cycling Team2.4%n/a––
Tolio AlexBardiani-CSF 7 Saber2.2%n/a––
Ponomar AndriiGeneral Store-Essegibi-F.Ili Curia1.8%n/a––
Verre AlessandroMBH Bank Ballan Telecom Fort1.7%n/a––
Masnada FaustoMBH Bank Ballan Telecom Fort1.5%n/a––
Dati TommasoTeam UKYO1.5%n/a––
Bessega TommasoTeam Polti VisitMalta1.2%n/a––
Pattinson TomosEF Education-Aevolo1.2%n/a––
BarcelΓ³ FernandoCaja Rural-Seguros RGA1.1%n/a––
Hindsight (post-hoc) β€” oracle-matched references β€” top 30, all heads
#RiderBlendCascade LambdaRankPodiumTop 15 Book odds
1Godon Dorian10.8%25.6%8.4%47.3%74.3%n/a
2Vauquelin KΓ©vin8.9%13.5%8.8%27.6%65.1%n/a
3Crescioli Ludovico8.4%5.4%10.1%16.7%72.1%n/a
4Berasategi Xabier6.2%3.1%7.7%12.8%58.5%n/a
5Fancellu Alessandro5.0%2.6%6.2%13.7%57.5%n/a
6Gaviria Fernando3.7%2.1%4.5%7.1%21.6%n/a
7Castellon Jan3.4%2.5%4.0%7.0%34.1%n/a
8Colnaghi Luca3.2%2.5%3.7%7.9%35.6%n/a
9Turconi Filippo2.7%1.4%3.3%6.0%41.8%n/a
10Ulissi Diego2.6%1.8%3.1%7.4%47.2%n/a
11Rondel Mathys2.4%1.3%3.0%5.3%46.7%n/a
12Tolio Alex2.2%2.4%2.4%5.8%23.8%n/a
13Ponomar Andrii1.8%1.4%2.2%3.2%21.8%n/a
14Verre Alessandro1.7%1.0%2.1%2.1%7.9%n/a
15Masnada Fausto1.5%0.8%1.9%1.8%6.9%n/a
16Dati Tommaso1.5%3.0%1.2%10.1%39.8%n/a
17Bessega Tommaso1.2%1.1%1.4%4.4%36.6%n/a
18Pattinson Tomos1.2%0.9%1.4%4.1%23.9%n/a
19BarcelΓ³ Fernando1.1%1.1%1.3%4.9%27.7%n/a
20Jasch Lennart0.9%1.4%0.9%4.4%39.3%n/a
21Kajamini Florian0.9%0.4%1.1%1.1%3.3%n/a
22Lastra Jonathan0.7%0.2%0.9%1.0%12.1%n/a
23Juaristi Txomin0.7%0.4%0.8%1.7%10.0%n/a
24Tsarenko Kyrylo0.7%0.4%0.8%1.1%9.1%n/a
25Ballabio Giacomo0.7%0.1%0.9%0.7%29.2%n/a
26Poels Wout0.6%0.2%0.8%1.1%12.7%n/a
27Cantoni Andrea0.6%0.4%0.7%3.0%23.1%n/a
28Sutton Louis0.6%0.3%0.7%2.3%21.1%n/a
29Masciarelli Lorenzo0.5%0.2%0.7%0.5%3.6%n/a
30Christophersen Cedrik Bakke0.5%0.6%0.6%1.6%18.7%n/a

After the stage

Results β€” won by D'Aiuto Filippo

PosRiderTeamHindsight (post-hoc) win p
1D'Aiuto FilippoGeneral Store-Essegibi-F.Ili Curia0.5%
2Godon DorianNetcompany INEOS10.8%
3Dati TommasoTeam UKYO1.5%
4Colnaghi LucaBardiani-CSF 7 Saber3.2%
5Lorello RiccardoSC Padovani Polo Cherry Bank0.2%
6Bessega TommasoTeam Polti VisitMalta1.2%
7D'Amato AndreaTeam UKYO0.1%
8Crescioli LudovicoTeam Polti VisitMalta8.4%
9Fancellu AlessandroMBH Bank Ballan Telecom Fort5.0%
10Vauquelin KΓ©vinNetcompany INEOS8.9%

Scenario scoring β€” probability given to the actual winner

ScenarioBlend win pBook devig
Hindsight (post-hoc)0.46%–