#	Question	My Pick	2×	Pts Available	Pts Earned
01	Predict the podium	Antonelli · Russell · Piastri	2×	62	20
02	Fastest driver in qualifying	Kimi Antonelli	—	10	—
03	Qualifies highest (Sainz vs Alonso)?	Carlos Sainz	—	10	10
04	How many teams get a car into Q3?	7	—	20	—
05	Fastest lap of the race	Kimi Antonelli	—	10	—
06	Fastest pitstop (team)	Red Bull Racing	—	25	—
07	How many SC/VSC calls during the GP?	3	—	20	—
08	How many classified finishers?	18	—	15	15
09	Finishes ahead (Lindblad / Colapinto / Bortoleto / Bearman)?	Franco Colapinto	—	15	15
10	Most points (McLaren vs Ferrari)?	McLaren	—	10	10
TOTAL				197	70

Final: 70 pts, 40% accuracy. Four questions landed in full (Q3, Q8, Q9, Q10) plus one podium leg. That is up 30 on R13’s 40 and 8.8 below the sheet’s EV of ~78.8, the closest a round has come to its forecast in some time. The podium question scored again as the multiplier question after R13’s zero. All 20 of its points came from the one leg the R13 post-mortem demanded: Antonelli at P1.

Q1 podium: Antonelli · Verstappen · Norris. The Antonelli leg hit the exact slot, 5 payout + 5 exact-slot bonus = 10, doubled to 20. Russell finished P5 and Piastri P8.

The platform’s answers for questions our race results file doesn’t record: fastest pitstop was Mercedes, and there was 1 SC/VSC call. The rest come from content/results/R14-Spain.md: pole Norris, fastest lap Russell, 6 teams in Q3 (McLaren, Mercedes, Red Bull, Ferrari, Alpine, Racing Bulls), 18 classified finishers.

Pts Available follows the picks file (payout × multiplier, 197 total). It does not include the exact-slot bonus, which is why Q1’s Antonelli leg earned 20 against 10 listed. Q1’s true ceiling was 124: each leg can pay double for an exact slot, then the whole question doubles.

Notes

Pre-race decisions:

• 2× on the podium (Q1) for the seventh consecutive round, on ~19.2 combined EV against ~7.2 for the best single-answer question (Q3). Sheet EV was ~78.8 with the multiplier, against 197 available.
• Three declared beliefs: (1) Mercedes are the class of the field and Antonelli is the best asset; (2) a new street circuit is messier and qualifying decides the race; (3) on an unknown circuit, season evidence beats one-race results.
• The new R13 test was applied for the first time: for each belief, name the highest-payout question it could carry and give a reason if it isn’t taken there. Belief 1 went onto Q1’s P1 and P2 legs, belief 3 onto Q6 at 25, and belief 2 was deliberately kept off the extreme buckets.
• Piastri (18) taken over Norris (12) for P3, as “the cheaper of two near-identical assets”, with Norris graded 6.96 against Piastri’s 7.92.
• Q6 given a “mechanical” method after 0-for-5: the highest EV among crews that actually win pit stops, which meant Red Bull at 25 over Mercedes (10) and Ferrari (12).
• Two concentrations named in advance: Antonelli across Q1, Q2 and Q5, and Mercedes on two of the three Q1 legs. Q7 and Q8 were flagged as “the same bet twice”.

Result:

• The R13 correction worked, and it is most of the round. R13’s lesson was that the strongest belief had been bet on the smallest questions and left off the biggest. R14 put Antonelli at P1 on the multiplier question, and he won from P2 by 4.351 for 20 points, the largest single return on the sheet. The concentration named in advance went 1 from 3: Norris took pole (Q2) and Russell took fastest lap (Q5). But the one that paid was the one with the multiplier on it, which is how the new test says the risk should be taken.

• Belief 3 (season evidence) went 3 from 4 and is now the most reliable method on the sheet. Sainz over Alonso (P17 v P18, 10), Colapinto ahead (P7, with Lindblad P9, Bortoleto P13 and Bearman P16, 15) and McLaren over Ferrari (19 to 12, helped by Hamilton’s lap-6 DNF, 10). With R13’s Gasly call, season-evidence picks are 4 from 5 across the last two rounds. The miss was Q6, the only one of the four priced at a long payout.

• ⚠ The runner-up column held the answer again, and it was the biggest one on the sheet. The picks file graded Norris for P3 at 6.96 and took Piastri at 7.92 on the grounds that the two McLarens were “near-identical assets”. Norris took pole and finished P3, while Piastri qualified P7 and finished P8. A Norris P3 leg would have paid 12 + 12 exact-slot bonus = 24, doubled to 48, making the round 118 instead of 70. This is the second round running that a named alternative was the answer (R13: Antonelli in Q1, “Tied” in Q7, “19” in Q10). The near-identical-assets argument was wrong on its own terms. Norris had 57 more championship points and finished ahead at Monza, so he was not the same asset, and the 50% payout gap was the market pricing that correctly. On podium legs, stop buying the cheaper teammate. Buy the one the season column says is better.

• Belief 2 (a messier street circuit) was half right, and the two halves came apart. Four retirements gave 18 classified finishers (Q8, 15), but they came with only 1 SC/VSC call against our 3 (Q7). The advance flag that Q7 and Q8 were “the same bet twice” was wrong in a useful way: retirements and neutralisations do not move together. Hamilton, Stroll, Pérez and Sainz all retired without triggering repeated safety cars. Price the two questions separately from now on.

• Q4 is now 0 for 2 on “7” and the market’s “6” has been right three rounds running (Zandvoort, Monza, Madrid). (⚠ Corrected 2026-09-30: five rounds running, R10–R14. Our results files show six teams in Q3 at Spa and Budapest too.) The street-circuit mechanism (“a single Q2 error frees a slot for a seventh team”) didn’t happen: the top four teams got all eight cars through, and Hülkenberg missed Q3 by one place at P11. That was close, but the favourite won on its merits. Take the favourite on this question unless a specific top-team car is known to be compromised before qualifying.

• Q6 is now 0 for 6, with roughly 117 points of availability returning nothing, and Mercedes, the favourite at 10, won it. The “mechanical” method still wrote a case, choosing Red Bull’s reputation over the favourite. Across six rounds the winners have included Audi and Mercedes, and none of them was our pick. Four files have recommended dropping the question, and every one was overridden. Take the favourite, the lowest payout, and write nothing. The question has not rewarded analysis once.

• Q5 (fastest lap) is now 1 for 6, and this time the retired heuristic would have worked. Russell finished P5, 29.8 s behind the winner with nobody close enough to threaten him, and set the lap on lap 49: a car out of position with a free late stop. That is exactly the R10 heuristic the picks file declared “not load-bearing”. One data point doesn’t bring it back, but it undercuts the reason for dropping it. The R13 hit (Antonelli, leading) and this one (Russell, out of position) point in opposite directions, and the question looks close to random.

• Q2 went to neither Mercedes. Russell was graded marginally higher than Antonelli, but Norris took pole by 0.011 and Antonelli was second. Taking Antonelli for consistency with Q1 was the right process: a Norris pick here would have required a belief the sheet did not hold.

• Season total: 1,182 pts across 14 rounds (R01 97 · R02 99 · R03 87 · R04 90 · R05 28 · R06 88 · R07 56 · R08 78 · R09 93 · R10 163 · R11 65 · R12 128 · R13 40 · R14 70). The Q1 multiplier has now paid in nine of ten rounds it anchored. The only difference between 70 and 118 was the choice of P3 leg.