Cycling race predictor · Vuelta a España 2026

Stage 15 — 2026-09-06

The same production stack priced under 4 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.

4 scenarios 150 riders priced models retrained 8e63bf1bbe book snapshot markets_1720.json

The fork

4 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.

Bunch sprint

The sprint comes home — Visma close it on the flat and Brennan does the rest
sprint1.00
mountain0.00
hilly0.00
rolling0.00
break0.00
TT0.00
keyed: bunch sprint 0.6 / survivor gallop 0.3 / fast-finisher power day 0.1 (board order by rider-type retrieval; prices unchanged)
The base case and the book's card, on the campaign's pure-speed anchor references (flat GT/WT sprints: Xeraco and Lorca both came back on these boards with Brennan #1 at 60%+ and the podium in the top three). A 110 km day whose last summit is 36 km out and whose last 25 km are flat and straight, with a 150-rider field and a leader's team that has no need to race. Inherits the config-root key. The question is composition, not shape: Brennan has been 1st, 1st, 1st and 90th on this race's four sprint days, and the 90th was the hot one.
Top 5 (blend): Brennan Matthew 57.7%; Meeus Jordi 8.4%; van Aert Wout 7.1%; De Schuyteneer Steffen 3.0%; Dainese Alberto 1.4%

Artafi thins it

Artafi thins it — a reduced group in the heat and the fastest man present wins
sprint0.52
mountain0.01
hilly0.46
rolling0.01
break0.00
TT0.00
keyed: survivor gallop 0.5 / bunch sprint 0.3 / split sprint 0.1 (board order by rider-type retrieval; prices unchanged)
The middle case and the operator's read drawn properly. References are the lumpy-finale reduced-sprint archetype (sprint ~0.51, hilly ~0.46, breakaway ~0) — flat stages whose late climb thinned the bunch without truly selecting it. 5.9 km at 5.5% in 39°C on day fifteen is enough to drop the men who suffered on the Pandera, and if Visma ride it for van Aert rather than tempo it for Brennan the group that crests Artafi may be 60 rather than 120. Brennan and van Aert are both riders who win THIS shape as readily as a full bunch, so the board's interest is in who fills the places behind them: Cort, Kron, Laurance, Nys, Albanese, Aular — the survivors' sprint card the market has at 21-41.
Top 5 (blend): Brennan Matthew 60.6%; van Aert Wout 8.8%; Meeus Jordi 6.7%; De Schuyteneer Steffen 2.0%; Laurance Axel 1.9%

Break survives

The escape survives — 39°C, a thin chase, and a break that is never brought back
sprint0.00
mountain0.38
hilly0.00
rolling0.62
break0.00
TT0.00
keyed: hard puncheur day 0.5 / mountain escape 0.3 / rouleur power group 0.2 (board order by rider-type retrieval; prices unchanged)
The branch the market prices at almost nothing and the campaign has learned not to dismiss. Stage 13 — a hillier day, but a break day on a race whose GC has been settled — went to the escape with van Aert in it; stage 14's 44-man move was given seven minutes because nobody had a reason to chase. Here the chase has a reason, but only one team with the legs to act on it, and 39°C after the Pandera is when trains fall apart. References are hilly and medium-mountain GT stages won from an escape solo or two-up with the peloton minutes back (front hilly/rolling/TT-mixed, no sprinters, winner groups of 1-2): van Aert's Vuelta 2024 stage 10, Izagirre's TdF 2023 stage 12, Cort's TdF 2022 stage 10 at Megève, Vendrame's Giro 2024 stage 19 and Bettiol's Giro 2026 stage 13. The GC top 10 is cut on standings — everyone inside 7:01, through Cristian Rodriguez — which frees the whole book tail: Leknessund, Vermaerke, Küng, Romeo, Camprubí, Oliveira, Armirail, Berrade, Bilbao, Dversnes.
Top 5 if the break wins (top 10 on GC removed): Tejada Harold 5.9%; van Aert Wout 5.9%; Romeo Ivan 4.2%; Bisiaux Léo 4.1%; Widar Jarno 3.5%

Hindsight (post-hoc)

Hindsight (post-hoc) — the day as raced: a five-man escape holds a minute, van Aert wins the sprint
sprint0.01
mountain0.32
hilly0.00
rolling0.67
break0.00
TT0.00
keyed: hard puncheur day (board order by rider-type retrieval; prices unchanged)
Post-race reconstruction, NOT a prediction — run with --no-log and never written to prediction_log. Visma rode the Puerto Artafi — Kruijswijk, Nordhagen and Armirail cut the peloton to about forty and dropped Coquard — then Leknessund attacked, Debruyne bridged, and van Aert went across at the Trassierra sprint with Romele on his wheel and Nys closing from behind. The five held 58 seconds on a peloton that stopped chasing; van Aert covered every late move and won the sprint from Romele and Nys, Debruyne and Leknessund two seconds back, Covi leading the peloton home 6th. The oracle labelled the day res_4 (tiny group, winner group of three, a punchy front rather than a sprinters' one). References are small escapes that survived by a minute on short or rolling days and were won by a fast or punchy man: Dversnes's Giro 2026 stage 15, Würtz Schmidt's Tirreno 2021 stage 6, Kron's Suisse 2021 stage 6, Williams's Tour of Britain 2024 stage 2 and Christen's Tour of Slovenia 2025 stage 2 — a set disjoint from every prior board on this race. The question is the usual one: does a profile built from days the escape actually won recover Romele (#34-88 on the morning boards) or Leknessund (#23-92)?
Top 5 (blend): van Aert Wout 4.5%; Widar Jarno 2.7%; Bisiaux Léo 2.4%; Onley Oscar 2.4%; Tejada Harold 2.3%

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 Bunch sprint 87%, Artafi thins it 93%, Hindsight (post-hoc) 71% — 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.

RiderTeamBunch sprintArtafi thins itBreak survivesHindsight (post-hoc)Book oddsBook devigMax edge
Brennan MatthewTeam Visma | Lease a Bike57.7%60.6%0.6%1.5%2.5025.4%+35.2pp
van Aert WoutTeam Visma | Lease a Bike7.1%8.8%5.9%4.5%4.5014.1%-5.3pp
Meeus JordiRed Bull-BORA-hansgrohe8.4%6.7%0.0%0.2%23.002.8%+5.6pp
De Schuyteneer SteffenLotto-Intermarché3.0%2.0%0.1%0.2%251.000.3%+2.8pp
Widar JarnoLotto-Intermarché0.0%0.1%3.5%2.7%67.000.9%+1.7pp
Bisiaux LéoDecathlon CMA CGM Team0.0%0.0%4.1%2.4%101.000.6%+1.8pp
Onley OscarNetcompany INEOS0.0%0.1%–2.4%501.000.1%+2.2pp
Tejada HaroldXDS Astana Team0.0%0.0%5.9%2.3%251.000.3%+2.0pp
Laurance AxelNetcompany INEOS1.1%1.9%2.3%2.2%34.001.9%+0.4pp
Roglic PrimozRed Bull-BORA-hansgrohe0.0%0.0%–1.9%501.000.1%+1.8pp
Romeo IvanMovistar Team0.0%0.0%4.2%1.8%51.001.2%+0.6pp
Johannessen Tobias HallandUno-X Mobility0.1%0.1%3.5%1.7%41.001.5%+0.1pp
Berthet ClémentGroupama-FDJ United0.0%0.1%–1.5%301.000.2%+1.3pp
Bilbao PelloBahrain Victorious0.0%0.0%3.4%1.4%81.000.8%+0.6pp
Dainese AlbertoSoudal Quick-Step1.4%1.2%0.0%0.1%251.000.3%+1.2pp
Fisher-Black FinnRed Bull-BORA-hansgrohe0.0%0.3%2.5%1.4%51.001.2%+0.1pp
Nordhagen JørgenTeam Visma | Lease a Bike0.0%0.0%2.8%1.3%401.000.2%+1.2pp
Castrillo PabloMovistar Team0.0%0.0%2.6%1.3%151.000.4%+0.9pp
Debruyne RamsesAlpecin-Premier Tech0.0%0.1%1.6%1.3%126.000.5%+0.8pp
Vansevenant MauriSoudal Quick-Step0.0%0.0%2.4%1.2%301.000.2%+1.0pp
Kron AndreasUno-X Mobility0.0%0.1%1.9%1.2%29.002.2%-0.9pp
Almeida JoaoUAE Team Emirates-XRG0.0%0.0%3.3%1.2%151.000.4%+0.8pp
Zana FilippoSoudal Quick-Step0.0%0.0%2.2%1.1%126.000.5%+0.6pp
García RaúlMovistar Team0.0%0.0%1.5%1.1%126.000.5%+0.5pp
Break survives 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 sprint comes home — Visma close it on the flat and Brennan does the rest — top 30, all heads
#RiderBlendCascade LambdaRankPodiumTop 15 Book odds
1Brennan Matthew57.7%46.5%69.0%68.4%88.7%2.50
2Meeus Jordi8.4%11.1%8.9%41.0%77.7%23.00
3van Aert Wout7.1%7.5%8.0%26.1%61.0%4.50
4De Schuyteneer Steffen3.0%2.2%3.7%9.7%68.6%251.00
5Dainese Alberto1.4%1.4%1.6%8.5%54.7%251.00
6Laurance Axel1.1%1.8%1.1%8.8%62.4%34.00
7Macias Cesar0.9%1.2%0.9%6.3%62.9%501.00
8Hofstetter Hugo0.8%1.2%0.8%7.1%79.0%101.00
9Brustenga Marc0.7%1.0%0.7%4.8%57.2%401.00
10Coquard Bryan0.6%0.8%0.7%5.5%57.8%29.00
11Govekar Matevz0.6%1.3%0.6%8.8%62.4%251.00
12Aular Orluis0.5%0.9%0.5%5.2%52.5%41.00
13Miquel Delgado Pau0.5%0.8%0.5%4.3%60.3%41.00
14Oliveira Ivo0.4%1.0%0.3%4.5%28.1%67.00
15Renard Alexis0.3%0.8%0.3%4.1%49.5%501.00
16Cort Magnus0.3%1.4%0.1%2.6%17.1%21.00
17Canal Carlos0.3%0.3%0.3%2.5%27.7%251.00
18Braet Vito0.2%0.5%0.2%4.0%57.5%201.00
19Tronchon Bastien0.1%0.5%0.1%3.1%34.1%41.00
20Camprubí Marcel0.1%0.2%0.1%2.4%18.5%51.00
21Nys Thibau0.1%0.3%0.1%1.3%27.1%34.00
22Thornley Callum0.1%0.5%0.0%1.3%8.4%601.00
23Brenner Marco0.1%0.3%0.1%2.2%9.5%101.00
24Garofoli Gianmarco0.1%0.2%0.1%1.2%19.1%201.00
25Albanese Vincenzo0.1%0.2%0.1%1.2%31.2%34.00
26Hayter Ethan0.1%0.2%0.1%2.7%15.8%151.00
27Mulubrhan Henok0.1%0.1%0.1%0.5%9.5%126.00
28Vader Milan0.1%0.1%0.1%1.2%15.1%301.00
29Johannessen Tobias Halland0.1%0.2%0.0%0.5%2.1%41.00
30Alleno Clément0.1%0.1%0.1%0.7%27.4%501.00
Artafi thins it — a reduced group in the heat and the fastest man present wins — top 30, all heads
#RiderBlendCascade LambdaRankPodiumTop 15 Book odds
1Brennan Matthew60.6%47.6%68.6%66.1%84.5%2.50
2van Aert Wout8.8%6.8%10.0%25.8%57.2%4.50
3Meeus Jordi6.7%9.5%6.6%32.5%72.7%23.00
4De Schuyteneer Steffen2.0%1.9%2.2%7.5%54.6%251.00
5Laurance Axel1.9%3.1%1.7%10.6%61.3%34.00
6Dainese Alberto1.2%1.1%1.3%7.5%47.3%251.00
7Hofstetter Hugo1.0%1.5%1.0%8.3%71.1%101.00
8Govekar Matevz0.9%1.6%0.9%8.8%57.9%251.00
9Oliveira Ivo0.7%2.4%0.3%6.0%22.5%67.00
10Coquard Bryan0.7%0.8%0.7%5.2%47.7%29.00
11Miquel Delgado Pau0.6%1.2%0.6%7.2%60.4%41.00
12Macias Cesar0.6%1.6%0.4%6.6%43.6%501.00
13Aular Orluis0.6%0.9%0.6%5.5%49.2%41.00
14Renard Alexis0.6%1.3%0.4%4.4%48.8%501.00
15Canal Carlos0.4%0.4%0.5%3.0%33.9%251.00
16Brustenga Marc0.4%1.0%0.3%4.4%41.9%401.00
17Fisher-Black Finn0.3%0.9%0.2%4.1%17.6%51.00
18Cort Magnus0.3%1.1%0.1%5.4%40.0%21.00
19Braet Vito0.3%0.6%0.2%4.5%58.5%201.00
20Mulubrhan Henok0.2%0.2%0.2%0.9%15.9%126.00
21Brenner Marco0.2%0.4%0.1%1.3%14.1%101.00
22Garofoli Gianmarco0.2%0.1%0.2%1.4%27.2%201.00
23Thornley Callum0.2%0.6%0.0%5.0%11.1%601.00
24Cobo Ivan0.1%0.3%0.1%1.5%22.9%151.00
25Camprubí Marcel0.1%0.1%0.1%1.4%23.0%51.00
26Johannessen Tobias Halland0.1%0.4%0.1%1.1%5.2%41.00
27Albanese Vincenzo0.1%0.2%0.1%1.8%41.0%34.00
28Hayter Ethan0.1%0.3%0.1%3.4%18.5%151.00
29Tronchon Bastien0.1%0.3%0.1%2.7%27.9%41.00
30Nys Thibau0.1%0.2%0.1%1.1%27.0%34.00
The escape survives — 39°C, a thin chase, and a break that is never brought back — top 30, all heads · top 10 on GC removed
#RiderBlendCascade LambdaRankPodiumTop 15 Book odds
1Tejada Harold2.9%2.5%4.1%7.0%27.5%251.00
2van Aert Wout2.8%2.1%4.1%5.9%15.3%4.50
3Romeo Ivan2.0%3.7%2.4%5.9%18.9%51.00
4Bisiaux Léo2.0%3.1%2.4%7.0%33.1%101.00
5Widar Jarno1.7%2.0%2.3%8.2%32.3%67.00
6Johannessen Tobias Halland1.7%1.6%2.4%7.7%25.2%41.00
7Bilbao Pello1.6%1.9%2.2%6.3%21.2%81.00
8Almeida Joao1.6%1.7%2.2%5.9%20.1%151.00
9Nordhagen Jørgen1.3%1.9%1.7%5.2%19.6%401.00
10Riccitello Matthew1.3%0.8%1.9%3.5%32.2%601.00
11Paret-Peintre Valentin1.3%2.1%1.6%8.4%21.1%41.00
12Castrillo Pablo1.3%1.6%1.6%2.5%10.0%151.00
13Fortunato Lorenzo1.2%1.7%1.6%7.2%26.0%401.00
14Fisher-Black Finn1.2%1.8%1.5%4.5%13.8%51.00
15Vansevenant Mauri1.2%1.8%1.4%5.5%23.8%301.00
16Buitrago Santiago1.2%1.0%1.6%3.8%21.8%151.00
17Laurance Axel1.1%1.0%1.5%3.2%17.1%34.00
18Zana Filippo1.1%1.6%1.3%5.7%22.6%126.00
19Vermaerke Kevin1.0%1.7%1.2%5.0%23.9%41.00
20Svestad-Bårdseng Embret0.9%0.7%1.3%2.1%20.0%501.00
21Kron Andreas0.9%1.7%1.0%4.6%23.6%29.00
22Rodriguez Carlos0.9%0.8%1.2%2.2%10.5%n/a
23Brenner Marco0.8%1.5%0.9%6.3%18.0%101.00
24Debruyne Ramses0.8%2.5%0.7%8.3%24.9%126.00
25Leknessund Andreas0.7%0.8%1.0%1.8%14.6%29.00
26García Raúl0.7%1.0%0.9%2.2%10.8%126.00
27Double Paul0.6%0.2%0.9%1.0%5.9%301.00
28Martin Guillaume0.4%0.2%0.7%1.0%14.6%251.00
29Sosa Ivan0.4%0.4%0.6%1.3%7.0%n/a
30Pickering Finlay0.4%0.2%0.6%1.3%12.4%601.00
Hindsight (post-hoc) — the day as raced: a five-man escape holds a minute, van Aert wins the sprint — top 30, all heads
#RiderBlendCascade LambdaRankPodiumTop 15 Book odds
1van Aert Wout4.5%3.8%6.5%10.5%26.4%4.50
2Widar Jarno2.7%1.7%4.0%6.3%32.4%67.00
3Bisiaux Léo2.4%2.5%3.4%5.5%26.7%101.00
4Onley Oscar2.4%0.8%3.7%12.1%27.5%501.00
5Tejada Harold2.3%1.8%3.3%5.1%22.4%251.00
6Laurance Axel2.2%2.5%3.1%7.8%29.6%34.00
7Roglic Primoz1.9%2.6%2.5%5.0%18.5%501.00
8Romeo Ivan1.8%2.1%2.5%4.8%16.5%51.00
9Johannessen Tobias Halland1.7%1.0%2.5%4.4%23.8%41.00
10Berthet Clément1.5%0.4%2.4%2.0%16.3%301.00
11Brennan Matthew1.5%3.9%1.5%8.3%17.2%2.50
12Bilbao Pello1.4%1.3%2.0%4.3%21.9%81.00
13Fisher-Black Finn1.4%2.6%1.7%6.1%18.1%51.00
14Nordhagen Jørgen1.3%1.7%1.8%4.8%19.3%401.00
15Castrillo Pablo1.3%1.2%1.9%2.3%7.1%151.00
16Debruyne Ramses1.3%3.0%1.3%7.9%27.8%126.00
17Vansevenant Mauri1.2%1.7%1.6%5.1%28.1%301.00
18Kron Andreas1.2%2.4%1.4%6.2%32.3%29.00
19Almeida Joao1.2%1.1%1.7%4.6%18.7%151.00
20Zana Filippo1.1%1.6%1.5%4.9%26.3%126.00
21García Raúl1.1%1.5%1.4%2.5%10.7%126.00
22Vermaerke Kevin1.0%1.0%1.5%1.9%26.7%41.00
23Paret-Peintre Valentin1.0%1.7%1.2%8.7%24.2%41.00
24Gall Felix1.0%0.8%1.4%6.6%25.6%n/a
25Fortunato Lorenzo0.9%0.5%1.4%2.4%8.2%401.00
26Buitrago Santiago0.9%0.5%1.4%2.1%12.2%151.00
27Carapaz Richard0.9%0.8%1.3%3.7%15.1%301.00
28Mas Enric0.9%0.8%1.3%4.9%18.9%n/a
29Omrzel Jakob0.9%0.2%1.4%3.4%15.5%n/a
30Skjelmose Mattias0.9%0.8%1.2%3.2%16.9%301.00

After the stage

Results — won by van Aert Wout

PosRiderTeamBunch sprint win pArtafi thins it win pBreak survives win pHindsight (post-hoc) win p
1van Aert WoutTeam Visma | Lease a Bike7.1%8.8%2.8%4.5%
2Romele AlessandroXDS Astana Team0.0%0.1%0.1%0.2%
3Nys ThibauLidl-Trek0.1%0.1%0.1%0.7%
4Debruyne RamsesAlpecin-Premier Tech0.0%0.1%0.8%1.3%
5Leknessund AndreasUno-X Mobility0.0%0.0%0.7%0.7%
6Covi AlessandroTeam Jayco-AlUla0.0%0.0%0.1%0.2%
7Tronchon BastienGroupama-FDJ United0.1%0.1%0.1%0.5%
8Mulubrhan HenokXDS Astana Team0.1%0.2%0.4%0.4%
9Küng StefanTudor Pro Cycling Team0.0%0.1%0.2%0.4%
10Oliveira IvoUAE Team Emirates-XRG0.4%0.7%0.0%0.2%

Scenario scoring — probability given to the actual winner

ScenarioBlend win pBook devig
Bunch sprint7.07%14.08%
Artafi thins it8.79%14.08%
Break survives2.82%14.08%
Hindsight (post-hoc)4.51%14.08%

Reflection

What we make of it

The break survived and the official read did not see it coming. Wout van Aert won his second stage in three days from a five-man escape — Romele second, Nys third, Debruyne and Leknessund two seconds back — with the peloton at 58 seconds behind Covi, Cort and Brennan at 1:47, and the GC group together at 2:44. The oracle labelled the day res_4, a punchy front from a tiny group, and the key of bunch 0.55 / res_3 0.30 / res_1 0.15 was not covered at all: the first family miss of the campaign since stage 7. The tier-3 record stands at 6/13 exact, 11/13 covered, mean weight 0.48.

The same-race sprint ledger — five of five sprint-flavoured stages had come back `bunch` — was the argument, and it was falsified on the sixth. What the road book comparison could not see was who would ride Artafi and why. The morning note named the mechanism almost exactly ('a 5.5% climb 36 km out on a short, hot stage is exactly where a fast climber's team might ride to shed the last of the pure speed') and then weighted it into res_3, a reduced sprint, rather than an escape. Visma did ride it, but not to set up a smaller sprint for Brennan: they rode it to launch van Aert, who bridged to Leknessund and Debruyne at the intermediate sprint and never looked back. The same team held the sprint favourite and the strongest escape candidate, so the chase-incentive test returned 'controlled' — Visma twice over — when the honest answer was that Visma could win either shape and would choose the harder one on a 40°C day fifteen stages in. When one team owns both the fastest man and the best puncheur, the incentive test does not decide the family; which rider the team prefers does, and a team riding the last climb hard is telling you.

The boards told the story in the order the retrieval channel keeps telling it. The official 'bunch sprint' board had Brennan #1 at 57.7% against a 25.4% book, and he finished 24th; van Aert sat #3 at 7.1% against the book's 14.1%, and the market had the winner closer than the model did. The 'break survives' branch board, keyed res_4 0.45, had van Aert #1 and pulled Debruyne (#27) and Leknessund (#23) inside its pool of fifteen — three of the actual top fifteen against the production board's one — while the two sprint boards managed one each. Romele, a 23.0 shot on the book at 2.8%, was #34 to #88 across the boards: the same break-membership wall as Valdelinares, Calar Alto and the Pandera, this time on a flat day.

Van Aert's form signal, flagged after stage 7, has now paid four times: third at Valdelinares from the move, fifth at Lorca in the bunch, a win at stage 13 from a two-up, and a win here from a five-up. The model prices him as a fast man who survives lumps, which is why he was third on the sprint boards; what it cannot see is that on this race he has become the rider Visma choose over their own sprinter when the road is hard enough.

The hindsight board, built on five small escapes that survived by a minute and keyed res_4 outright, makes van Aert #1 at 4.5%, Nys #9, Debruyne #19 and Leknessund #29, and still leaves Romele at #65. It retrieves four of the top fifteen against the production board's four — no better than the morning's break branch, which is the eighth consecutive post-hoc reconstruction to fail to beat a prospective read. The winner was findable; his four companions were not, because escape membership is not a thing the model sees.

The classification did not move. Mas leads Gall by 2:06 and Roglic by 2:10, Carapaz 4:24, Onley 5:44; the book has Mas at 69.5%. The two GC heads remain split the same way as after the Pandera — the state-aware outlook at Mas 36.6% / Roglic 24.3% / Gall 2.9%, the blend at Roglic 30.0% / Gall 16.9% / Onley 16.2% / Mas 8.7% — neither within twenty points of the market on the leader. Five more riders left the race — Houle, Kretschy, Labrosse, Laporte and Peace — leaving 150. Monday is the second rest day; stage 16 is Tuesday 8 September.