#	Question	My Pick	2×	Pts Available	Pts Earned
01	Predict the podium	Antonelli · Norris · Verstappen	2×	54	24
02	Fastest driver in qualifying	Kimi Antonelli	—	10	—
03	Qualifies highest (Bortoleto vs Hülkenberg)?	Nico Hülkenberg	—	15	—
04	How many teams get a car into Q3?	6	—	10	10
05	Fastest lap of the race	Kimi Antonelli	—	10	—
06	Fastest pitstop (team)	Mercedes	—	10	10
07	SC/VSC in the first 15 laps?	No	—	15	15
08	How many classified finishers?	19	—	20	—
09	Finishes ahead (Gasly / Hülkenberg / Ocon / Albon)?	Pierre Gasly	—	10	—
10	Most points (Ferrari vs Red Bull)?	Red Bull Racing	—	15	15
TOTAL				169	74

Final: 74 pts, 40% accuracy. Four questions landed in full (Q4, Q6, Q7, Q10), plus one podium leg. That is 4 up on R14’s 70 and 14.7 below the sheet’s EV of ~88.7. The podium question paid as the multiplier question again, but through the Verstappen leg, not the Antonelli one.

Q1 podium: Russell · Verstappen · Hadjar. Verstappen was on the podium but at P2, not the P3 we picked, so the leg paid its 12 payout with no exact-slot bonus, doubled to 24. Antonelli finished P5 from a Q1 exit, and Norris retired on lap 35.

The platform’s answers for questions our race results file doesn’t record: fastest pitstop was Mercedes, and there was no SC/VSC in the first 15 laps. The rest come from content/results/R15-Azerbaijan.md: pole Russell, Bortoleto out-qualified Hülkenberg (P17 v P18), 6 teams in Q3 (Mercedes, Ferrari, McLaren, Red Bull, Alpine, Williams), fastest lap Russell, 16 classified finishers, Ocon first of the four (P8, with Hülkenberg P11 and Gasly and Albon retired), and Red Bull 33 GP points to Ferrari’s 20.

Pts Available follows the picks file (payout × multiplier, 169 total) and does not include the exact-slot bonus. Q1’s true ceiling was 108.

Notes

Pre-race decisions:

• 2× on the podium (Q1) for the eighth consecutive round, on ~19.0 combined EV against ~8.3 for the best single-answer question (Q7). Sheet EV was ~88.7 with the multiplier, against 169 available.
• A new method: every pick started from a season base rate counted from R01–R14, and narrative had to overturn it. This responded to three post-mortems in which mechanism picks lost to base rates.
• Three declared beliefs: (1) base rates hold unless there is specific evidence against them; (2) Antonelli is the benchmark and Norris was the fastest driver in Madrid; (3) Red Bull is ahead of Ferrari right now, the one deliberate override of a season base rate (Ferrari led 10–4).
• Russell was marked down across the sheet for an end-of-life power unit at a power circuit: left off Q1 in favour of Verstappen at P3, and graded below Antonelli on Q2 and Q5.
• Q6 taken mechanically: the favourite, with no case written, as the R14 post-mortem prescribed.
• Two concentrations named in advance: Antonelli across Q1 (P1), Q2 and Q5 (30 points of availability), and Q7/Q8 as one “cleaner race” belief (35).

Result:

• The base-rate method went 2 from 7, and both hits were the market favourites. Q4 “6” (10) and Q6 Mercedes (10) landed. Everything where the base rate pointed above the favourite lost: Q3 Hülkenberg at 15 (Bortoleto won P17 v P18, now 8–7 on the season), Q8 “19” at 20 (16 classified), and Q9 Gasly (a DNF from P7 on the grid). The two Antonelli picks, Q2 and Q5, lost to the same driver. The method was not wrong so much as unlucky in where it was tested. Q3 was a stated coin flip, and Q9 was lost to a retirement. But one round in, it has earned 20 points from seven questions.

• Belief 3 was the one override of a base rate, and it scored 39 of the 74. Red Bull beat Ferrari 33–20 (Q10, 15), with Verstappen P2 and Hadjar P3, and Verstappen’s podium leg paid 24. Ferrari had both cars in Q3 and no engine penalty, so the “weakest position of the season” case was partly wrong on its facts. The belief still held on pace. The R13 test (put each belief on its highest-payout question) worked for the second round running: the belief carried on Q1’s 2× produced the round’s largest return.

• ⚠ The Antonelli concentration returned zero from 30 points of availability, and one driver took all three. Russell took pole, the win and the fastest lap, on the engine the sheet had marked him down for. Antonelli’s Q1 exit in P16 lost Q2 before the race started. The concentration was named and accepted on base rates (57% of wins, 43% of poles, 50% of fastest laps). Base rates don’t protect against correlated failure: one bad qualifying session lost all three picks together. The “old power unit at a power circuit” mechanism was the costliest belief on the sheet, and it was a narrative override of Russell’s own season rates (7 podiums, 4 poles) that belief 1 told us not to make.

  • The Verstappen-over-Russell call on Q1 still came out ahead. Russell at P3 would have paid 8, doubled to 16, against Verstappen’s 24. The mistake was on Q2 and Q5, where the same engine argument decided close calls against Russell.

• Q7 and Q8 split again, the second round running. Q7 “No” landed (15): no neutralisation in the first 15 laps, even with Stroll retiring on lap 7. Q8 “19” missed: six retirements left 16 classified. R14’s lesson, that retirements and neutralisations do not move together, held exactly, and the sheet still treated them as one belief. Price Q7 and Q8 independently from now on. On Q8 specifically, the distribution is no longer flat: 16 has now happened four times in 15 rounds, more than any other count (18 and 19 three times each). The street-circuit downward shift the picks file acknowledged was right, and the sheet took the 20-point payout on 19 anyway.

• Q6 hit for the first time this season, 1 for 7, with the favourite and no case written. Mercedes at 10. Six rounds of analysis returned nothing, and the first round without any paid. Keep it mechanical.

• Q4: the market’s “6” has now been right six rounds running (R10–R15, Spa to Baku), and this time we took it. Williams were the named risk, and Sainz did make Q3 in P9. But Racing Bulls missed out, and the count stayed at six. Keep taking the favourite.

• Q5 (fastest lap) is now 1 for 7. Russell set it on lap 49 while leading, so neither the leader heuristic nor the out-of-position heuristic predicts which one applies. Treat it like Q6: favourite, no case.

• Season total: 1,256 pts across 15 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 · R15 74). The Q1 multiplier has paid in ten of the eleven rounds it anchored, failing only at R13.