Cycling race predictor · Vuelta a España 2026

Stage 19 — 2026-09-11

The same production stack priced under 3 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.

3 scenarios 147 riders priced models retrained d911de60f7 book snapshot no capture

← System overview, season boards & the full race views

The fork

3 readings of the same road

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

GC showdown

The GC teams control it — Peñas Blancas decides the day and the race

sprint0.00

mountain0.44

hilly0.00

rolling0.55

break0.01

TT0.00

The base case. Stage 18 leaves the race lead close enough to attack, while stage 19 ends with 18.7 km at 6.5% and no descent to undo the selection. References are the campaign's established hard Grand Tour and WorldTour mountain days won by a GC leader from a group of two: Pogacar at Peyragudes in 2022, Vingegaard on Tour 2024 stage 11, Roglic on the Angliru in 2023, Almeida on Vuelta 2025 stage 13 and Roglic on Dauphine 2024 stage 7. The two earlier category-2 climbs add fatigue, but their distance from the finish makes team incentive more important than their gradients: with the red jersey and podium still live, the leading teams have reason to keep the break within reach before Peñas Blancas.

Top 5 (blend): Gall Felix 10.9%; Mas Enric 7.6%; Roglic Primoz 7.4%; Onley Oscar 6.5%; Johannessen Tobias Halland 4.4%

Mountain break

The escape survives — 210 kilometres give the stage hunters their last mountain chance

sprint0.00

mountain0.50

hilly0.00

rolling0.49

break0.01

TT0.00

The alternative, grounded in this Vuelta's repeated success by mountain escapes. The stage is the longest of the race and Puerto de las Abejas and Puerto del Viento offer almost 100 kilometres in which a strong move can build its advantage; after the second summit there is a long run to Estepona before the final climb. References are the campaign's canonical mountain-escape days: Tour 2026 stage 18, Vuelta 2025 stage 14, Pologne 2026 stage 6, Tour 2026 stage 9 and Giro 2026 stage 11. The conditional pool excludes the morning GC top 10 so the board answers the useful question—who wins if the classification teams race Peñas Blancas against each other but concede the stage? The scenario-level day key this branch used to carry has been removed: with no morning read logged, leaving it in place let the tier-3 scorer mistake it for one.

Top 5 if the break wins (top 10 on GC removed): Bisiaux Léo 7.6%; Johannessen Tobias Halland 7.3%; Buitrago Santiago 5.9%; Riccitello Matthew 4.8%; Widar Jarno 3.8%

Summit shatters

Peñas Blancas shatters — the strongest climber finishes alone

sprint0.00

mountain0.51

hilly0.00

rolling0.49

break0.00

TT0.00

The selective branch. A 210.8 km third-week stage followed by an uninterrupted 18.7 km ascent can create much larger gaps than the average gradient suggests, especially if the race is already reduced before Estepona. References are the campaign's established shattering-summit archetype: Tour 2026 stages 3, 10 and 14, Vuelta 2025 stage 20 and Catalunya 2026 stage 6. The comparison with the GC-showdown board isolates riders whose case improves when the finish becomes a long individual climbing test rather than a small-group contest. Its scenario-level day key has been removed for the same reason as the mountain_break branch.

Top 5 (blend): Gall Felix 11.8%; Onley Oscar 7.5%; Mas Enric 7.4%; Roglic Primoz 4.8%; Johannessen Tobias Halland 4.4%

The board

Win probabilities, scenario by scenario

Blend win probability per scenario (cascade × 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 — 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 — across this field each reading totals GC showdown 88%, Summit shatters 93% — 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.

Rider	Team	GC showdown	Mountain break	Summit shatters	Book odds	Book devig	Max edge
Gall Felix	Decathlon CMA CGM Team	10.9%	–	11.8%	n/a	–	–
Mas Enric	Movistar Team	7.6%	–	7.4%	n/a	–	–
Onley Oscar	Netcompany INEOS	6.5%	–	7.5%	n/a	–	–
Roglic Primoz	Red Bull-BORA-hansgrohe	7.4%	–	4.8%	n/a	–	–
Johannessen Tobias Halland	Uno-X Mobility	4.4%	7.3%	4.4%	n/a	–	–
Carapaz Richard	EF Education-EasyPost	3.0%	–	4.1%	n/a	–	–
Bisiaux Léo	Decathlon CMA CGM Team	3.0%	7.6%	3.9%	n/a	–	–
Riccitello Matthew	Decathlon CMA CGM Team	3.5%	4.8%	2.9%	n/a	–	–
Buitrago Santiago	Bahrain Victorious	2.6%	5.9%	2.8%	n/a	–	–
Skjelmose Mattias	Lidl-Trek	2.6%	–	2.0%	n/a	–	–
Kuss Sepp	Team Visma \| Lease a Bike	2.2%	2.7%	2.5%	n/a	–	–
Tejada Harold	XDS Astana Team	1.9%	–	2.5%	n/a	–	–
Widar Jarno	Lotto-Intermarché	2.0%	3.8%	2.2%	n/a	–	–
Almeida Joao	UAE Team Emirates-XRG	1.9%	3.2%	2.1%	n/a	–	–
Omrzel Jakob	Bahrain Victorious	2.0%	–	2.0%	n/a	–	–
Berthet Clément	Groupama-FDJ United	1.3%	–	1.8%	n/a	–	–
Paret-Peintre Valentin	Soudal Quick-Step	1.1%	3.2%	1.5%	n/a	–	–
Tuckwell Luke	Red Bull-BORA-hansgrohe	1.1%	2.8%	1.5%	n/a	–	–
Nordhagen Jørgen	Team Visma \| Lease a Bike	1.2%	3.3%	1.5%	n/a	–	–
Debruyne Ramses	Alpecin-Premier Tech	1.4%	3.2%	1.3%	n/a	–	–
Romeo Ivan	Movistar Team	1.3%	2.5%	1.3%	n/a	–	–
Rodriguez Cristian	XDS Astana Team	0.9%	–	1.3%	n/a	–	–
van Aert Wout	Team Visma \| Lease a Bike	1.2%	3.3%	1.2%	n/a	–	–
Vermaerke Kevin	UAE Team Emirates-XRG	1.0%	1.9%	1.1%	n/a	–	–

Mountain break is shown conditional on the break winning — the top 10 riders on actual GC are removed and the remaining probabilities renormalize within the break pool. That column isn't comparable with the book line or the other readings, so it's left out of Max edge; a "–" means a GC captain we assume won't be up the road.

The GC teams control it — Peñas Blancas decides the day and the race — top 30, all heads

#	Rider	Blend	Cascade	LambdaRank	Podium	Top 15	Book odds
1	Gall Felix	10.9%	9.8%	12.6%	32.2%	84.8%	n/a
2	Mas Enric	7.6%	5.6%	9.1%	27.4%	83.1%	n/a
3	Roglic Primoz	7.4%	7.6%	8.3%	27.9%	77.1%	n/a
4	Onley Oscar	6.5%	5.8%	7.6%	17.9%	77.8%	n/a
5	Johannessen Tobias Halland	4.4%	4.6%	4.9%	17.2%	51.2%	n/a
6	Riccitello Matthew	3.5%	2.2%	4.3%	7.3%	46.4%	n/a
7	Carapaz Richard	3.0%	2.5%	3.6%	10.7%	79.1%	n/a
8	Bisiaux Léo	3.0%	2.0%	3.7%	6.9%	31.5%	n/a
9	Buitrago Santiago	2.6%	1.7%	3.2%	6.0%	26.0%	n/a
10	Skjelmose Mattias	2.6%	2.1%	3.0%	7.7%	65.9%	n/a
11	Kuss Sepp	2.2%	3.7%	2.2%	9.1%	49.9%	n/a
12	Widar Jarno	2.0%	2.6%	2.1%	8.7%	41.7%	n/a
13	Omrzel Jakob	2.0%	1.3%	2.4%	6.3%	60.7%	n/a
14	Almeida Joao	1.9%	1.8%	2.2%	5.9%	21.2%	n/a
15	Tejada Harold	1.9%	3.3%	1.8%	9.8%	62.0%	n/a
16	Debruyne Ramses	1.4%	3.4%	1.1%	9.5%	31.8%	n/a
17	Berthet Clément	1.3%	0.4%	1.7%	4.8%	56.3%	n/a
18	Romeo Ivan	1.3%	2.0%	1.3%	2.9%	7.6%	n/a
19	van Aert Wout	1.2%	3.1%	0.9%	10.1%	19.3%	n/a
20	Nordhagen Jørgen	1.2%	2.3%	1.0%	8.3%	21.5%	n/a
21	Paret-Peintre Valentin	1.1%	3.0%	0.8%	5.4%	22.9%	n/a
22	Tuckwell Luke	1.1%	0.4%	1.4%	2.4%	11.6%	n/a
23	Vermaerke Kevin	1.0%	0.7%	1.2%	3.3%	20.7%	n/a
24	Rodriguez Cristian	0.9%	0.3%	1.2%	3.0%	51.7%	n/a
25	Rodriguez Carlos	0.9%	0.8%	1.1%	2.4%	13.3%	n/a
26	Zana Filippo	0.9%	0.8%	1.0%	2.8%	16.7%	n/a
27	Svestad-Bårdseng Embret	0.8%	0.5%	1.0%	1.9%	12.8%	n/a
28	Berrade Urko	0.8%	1.7%	0.6%	3.5%	16.2%	n/a
29	Leknessund Andreas	0.7%	0.5%	0.9%	3.5%	15.3%	n/a
30	Martin Guillaume	0.6%	0.3%	0.8%	1.6%	19.4%	n/a

The escape survives — 210 kilometres give the stage hunters their last mountain chance — top 30, all heads · top 10 on GC removed

#	Rider	Blend	Cascade	LambdaRank	Podium	Top 15	Book odds
1	Bisiaux Léo	3.6%	3.7%	4.6%	11.4%	48.2%	n/a
2	Johannessen Tobias Halland	3.5%	3.7%	4.4%	12.4%	42.9%	n/a
3	Buitrago Santiago	2.8%	1.9%	3.8%	6.6%	27.5%	n/a
4	Riccitello Matthew	2.3%	2.4%	2.9%	6.7%	40.9%	n/a
5	Widar Jarno	1.8%	2.1%	2.2%	6.9%	37.8%	n/a
6	Nordhagen Jørgen	1.6%	4.1%	1.4%	9.3%	25.3%	n/a
7	van Aert Wout	1.6%	3.3%	1.5%	8.0%	13.9%	n/a
8	Paret-Peintre Valentin	1.5%	2.6%	1.7%	8.0%	30.8%	n/a
9	Debruyne Ramses	1.5%	4.8%	1.1%	11.8%	39.8%	n/a
10	Almeida Joao	1.5%	2.0%	1.8%	8.9%	17.6%	n/a
11	Tuckwell Luke	1.3%	0.5%	1.9%	3.1%	9.8%	n/a
12	Kuss Sepp	1.3%	1.1%	1.7%	5.1%	25.0%	n/a
13	Romeo Ivan	1.2%	1.6%	1.4%	2.2%	5.9%	n/a
14	Zana Filippo	1.1%	0.4%	1.6%	1.5%	20.1%	n/a
15	Brenner Marco	1.1%	2.0%	1.2%	6.7%	17.7%	n/a
16	Vermaerke Kevin	0.9%	0.5%	1.3%	2.4%	18.4%	n/a
17	Sosa Ivan	0.9%	0.6%	1.2%	1.8%	8.2%	n/a
18	Svestad-Bårdseng Embret	0.9%	0.6%	1.1%	1.9%	14.6%	n/a
19	Martin Guillaume	0.8%	0.3%	1.1%	1.5%	24.1%	n/a
20	Rodriguez Carlos	0.8%	1.1%	0.9%	2.6%	12.1%	n/a
21	Leknessund Andreas	0.8%	0.4%	1.0%	2.4%	9.8%	n/a
22	Fortunato Lorenzo	0.7%	0.4%	1.0%	2.1%	17.3%	n/a
23	Berrade Urko	0.7%	0.8%	0.9%	2.0%	11.9%	n/a
24	Vansevenant Mauri	0.7%	0.4%	0.9%	1.1%	6.6%	n/a
25	Pickering Finlay	0.7%	0.4%	0.9%	1.4%	13.8%	n/a
26	Castrillo Pablo	0.6%	0.5%	0.9%	1.1%	6.4%	n/a
27	Dunbar Edward	0.6%	0.5%	0.8%	1.3%	14.1%	n/a
28	García Raúl	0.6%	0.5%	0.8%	2.0%	9.7%	n/a
29	Fisher-Black Finn	0.6%	0.3%	0.8%	1.5%	6.3%	n/a
30	Kron Andreas	0.5%	0.3%	0.7%	0.8%	5.8%	n/a

Peñas Blancas shatters — the strongest climber finishes alone — top 30, all heads

#	Rider	Blend	Cascade	LambdaRank	Podium	Top 15	Book odds
1	Gall Felix	11.8%	10.9%	12.8%	29.1%	81.2%	n/a
2	Onley Oscar	7.5%	5.2%	8.6%	16.8%	74.6%	n/a
3	Mas Enric	7.4%	6.3%	8.2%	26.7%	80.7%	n/a
4	Roglic Primoz	4.8%	4.3%	5.2%	20.6%	69.3%	n/a
5	Johannessen Tobias Halland	4.4%	3.8%	4.8%	14.0%	45.0%	n/a
6	Carapaz Richard	4.1%	3.4%	4.5%	13.4%	77.6%	n/a
7	Bisiaux Léo	3.9%	3.9%	4.2%	10.1%	36.8%	n/a
8	Riccitello Matthew	2.9%	2.8%	3.2%	8.1%	46.9%	n/a
9	Buitrago Santiago	2.8%	3.3%	2.9%	8.5%	30.0%	n/a
10	Kuss Sepp	2.5%	4.2%	2.3%	9.8%	53.5%	n/a
11	Tejada Harold	2.5%	4.3%	2.2%	10.6%	58.5%	n/a
12	Widar Jarno	2.2%	3.0%	2.1%	9.8%	41.4%	n/a
13	Almeida Joao	2.1%	3.4%	1.9%	10.7%	21.2%	n/a
14	Omrzel Jakob	2.0%	2.3%	2.1%	8.2%	69.8%	n/a
15	Skjelmose Mattias	2.0%	1.2%	2.4%	6.5%	58.9%	n/a
16	Berthet Clément	1.8%	0.6%	2.3%	5.0%	62.0%	n/a
17	Paret-Peintre Valentin	1.5%	2.4%	1.3%	7.4%	24.2%	n/a
18	Tuckwell Luke	1.5%	0.9%	1.7%	3.2%	11.4%	n/a
19	Nordhagen Jørgen	1.5%	3.1%	1.1%	7.4%	24.0%	n/a
20	Romeo Ivan	1.3%	1.8%	1.3%	2.5%	6.7%	n/a
21	Debruyne Ramses	1.3%	4.3%	0.6%	9.7%	31.1%	n/a
22	Rodriguez Cristian	1.3%	0.5%	1.6%	4.0%	55.6%	n/a
23	van Aert Wout	1.2%	2.6%	1.0%	6.8%	11.8%	n/a
24	Vermaerke Kevin	1.1%	0.7%	1.3%	2.4%	15.2%	n/a
25	Zana Filippo	1.1%	0.6%	1.2%	1.7%	20.2%	n/a
26	Svestad-Bårdseng Embret	0.9%	0.4%	1.1%	1.7%	15.3%	n/a
27	Berrade Urko	0.8%	1.4%	0.7%	2.7%	13.3%	n/a
28	Sosa Ivan	0.8%	0.5%	1.0%	1.7%	8.1%	n/a
29	Rodriguez Carlos	0.8%	0.6%	0.9%	2.2%	12.2%	n/a
30	Martin Guillaume	0.7%	0.2%	0.9%	1.4%	28.0%	n/a

After the stage

Results — won by Dunbar Edward

Pos	Rider	Team	GC showdown win p	Mountain break win p	Summit shatters win p
1	Dunbar Edward	Pinarello-Q36.5 Pro Cycling Team	0.6%	0.6%	0.6%
2	Buitrago Santiago	Bahrain Victorious	2.6%	2.8%	2.8%
3	Gloag Thomas	Pinarello-Q36.5 Pro Cycling Team	0.1%	0.1%	0.1%
4	Berrade Urko	Equipo Kern Pharma	0.8%	0.7%	0.8%
5	Camprubí Marcel	Pinarello-Q36.5 Pro Cycling Team	0.1%	0.1%	0.0%
6	Leknessund Andreas	Uno-X Mobility	0.7%	0.8%	0.7%
7	Martin-Guyonnet Guillaume	Groupama-FDJ United	–	–	–
8	Hamilton Chris	Team Picnic PostNL	0.0%	0.0%	0.0%
9	Vermaerke Kevin	UAE Team Emirates-XRG	1.0%	0.9%	1.1%
10	Samitier Sergio	Cofidis	0.0%	0.1%	0.0%

Scenario scoring — probability given to the actual winner

Scenario	Blend win p	Book devig
GC showdown	0.55%	–
Mountain break	0.60%	–
Summit shatters	0.56%	–

Reflection

What we make of it

The escape branch was right about the day and the oracle still graded it a miss. Eddie Dunbar won alone in 5:00:54 from a move the classification never tried to hold, with Santiago Buitrago at fourteen seconds and his own teammate Thomas Gloag at twenty-three. Dunbar started the day 35th overall. The front five had a general-classification median of 35th, so nothing about this finish belonged to the men racing for red.

The grade inside the family is the interesting part. The label came back res_4, the hard-puncheur neighbour, rather than res_2. Realised top-15 composition was mountain .40, rolling .267, breakaway .20 — a front of climber-rouleurs rather than pure climbers, which is exactly the signature that separates the two items. That is the second time in five settled stages that a break-won mountain finish at this Vuelta has landed on res_4 rather than res_2, after stage 15.

Nothing moved at the top of the race. Mas, Roglic, Gall, Carapaz and Onley all came in together, holding 1:37, 3:01, 5:17 and 6:03, so the whole classification question rolled into the queen stage unchanged. Mattias Skjelmose was the only man inside the top ten to lose ground, dropping from eighth to below Cristian Rodriguez.

The watch item is the cost of the missed cycle itself, not the read. Two consecutive stages went unkeyed and unpriced, and stage 19 then contaminated the tier-3 card by letting a scenario-level key stand in for a morning read. The guard that stops that exists and works; it only fires when the config is marked post_hoc, which a backfilled config has to be from the moment it is written, not after someone notices.