Model comparison — 26-event sample-resolved set

Same pipeline (Brave retrieval + market-anchor prompt + 0.10 longshot floor), five different LLMs. Each cell shows p = probability assigned to outcome[0] and brier = Brier loss for the resolved outcome. Hover any cell for the model's rationale on that event. Rows sorted by inter-model spread — most contested events first. Lower Brier = better. Random binary baseline = 0.25, uniform-prior baseline = 0.22.

< 0.05 (excellent) < 0.10 < 0.25 < 0.50 ≥ 0.50 (catastrophic)
Opus 4.7 (production)
0.2329
n=26 resolved
Sonnet 4.6 (prev prod)
0.0879
n=14 resolved
Opus 4.6
0.2264
n=26 resolved
GPT-5.2
0.2584
n=26 resolved
Gemini 3.1 Pro (post-harden)
0.4259
n=26 resolved

Reliability diagram

Production model calibration on resolved rows: Opus 4.7 (production). Small-n view: 26 outcome[0] probabilities, ECE 0.144.

forecast 0.05, observed 0.00, n=7forecast 0.10, observed 0.00, n=6forecast 0.22, observed 0.00, n=1forecast 0.50, observed 1.00, n=1forecast 0.53, observed 1.00, n=2forecast 0.82, observed 1.00, n=3forecast 0.90, observed 1.00, n=6 Mean forecast probability Observed frequency Perfect calibration

26 events × 5 models = 130 predictions in the grid below. Tooltips show rationale.

CategoryEvent (outcomes; resolved winner)Opus 4.7 (production)Sonnet 4.6 (prev prod)Opus 4.6GPT-5.2Gemini 3.1 Pro (post-harden)
SportsWho won the 2025-26 Liga Portugal (Portuguese top-flight football)?
n=18 winner: FC Porto
p=0.05
brier 0.792
p=0.05
brier —
p=0.05
brier 1.789
p=0.05
brier 1.789
p=0.06
brier 0.944
SportsWho won the 2025-26 NHL Calder Memorial Trophy (Rookie of the Year)?
n=30 winner: Matthew Schaefer
p=0.05
brier 0.083
p=0.05
brier —
p=0.05
brier 0.083
p=0.05
brier 0.073
p=0.05
brier 0.975
SportsWho won the 2025-26 La Liga (Spanish top-flight football)?
n=20 winner: Barcelona
p=0.05
brier 0.054
p=0.05
brier —
p=0.05
brier 0.050
p=0.05
brier 0.049
p=0.05
brier 0.950
EntertainmentWhich contestants were eliminated from Survivor Season 50 Episode 7?
n=14 winner: Dee Valladares
p=0.05
brier 0.035
p=0.05
brier —
p=0.05
brier 0.036
p=0.05
brier 0.037
p=0.07
brier 0.929
PoliticsHow many US Senators voted for Trump's Federal Reserve Chair pick (2027 confirmation vote)
n=12 winner: 54
p=0.05
brier 0.369
p=0.05
brier —
p=0.05
brier 0.028
p=0.05
brier 0.028
p=0.08
brier 0.917
SportsWho won the 2025-26 Ligue 1 (French top-flight football)?
n=18 winner: PSG
p=0.85
brier 0.073
p=0.78
brier —
p=0.75
brier 0.105
p=0.78
brier 0.097
p=0.06
brier 0.944
PoliticsHow many Supreme Court justices voted in favor of Louisiana in Louisiana v. Callais (2026)
n=10 winner: 3
p=0.10
brier 0.900
p=0.05
brier —
p=0.05
brier 0.225
p=0.10
brier 0.900
p=0.10
brier 0.900
PoliticsWhich coalition won the 2026 Colombia Senate election?
n=2 winner: Government Alliance
p=0.82
brier 0.032
p=0.75
brier 0.062
p=0.88
brier 0.014
p=0.72
brier 0.078
p=0.35
brier 0.423
SportsWho won the 2025-26 Serie A (Italian top-flight football)?
n=20 winner: Inter
p=0.22
brier 0.611
p=0.05
brier —
p=0.18
brier 0.722
p=0.14
brier 0.761
p=0.05
brier 0.950
EntertainmentWho won The Masked Singer Season 14?
n=18 winner: Galaxy Girl
p=0.05
brier 1.260
p=0.05
brier —
p=0.05
brier 0.952
p=0.06
brier 0.944
p=0.06
brier 0.944
SportsWho won the Sussex vs Leicestershire 2026 County Championship cricket match?
n=2 winner: Sussex
p=0.90
brier 0.010
p=0.75
brier 0.062
p=0.57
brier 0.181
p=0.60
brier 0.160
p=0.50
brier 0.250
ElectionsWho won the 2026 Democratic primary for West Virginia's 1st Congressional District?
n=2 winner: Vince George
p=0.10
brier 0.010
p=0.25
brier 0.062
p=0.10
brier 0.010
p=0.10
brier 0.010
p=0.50
brier 0.250
SportsWho won the Bax vs Arcon ATP Challenger tennis match on May 5, 2026?
n=2 winner: Florent Bax
p=0.55
brier 0.202
p=0.50
brier 0.250
p=0.75
brier 0.062
p=0.52
brier 0.230
p=0.50
brier 0.250
SportsWho won the Watson vs Okamura WTA Challenger tennis match on May 5, 2026?
n=2 winner: Heather Watson
p=0.50
brier 0.250
p=0.52
brier 0.230
p=0.55
brier 0.202
p=0.57
brier 0.185
p=0.73
brier 0.073
SportsWho won the Najzer vs Ebster tennis match in the 2026 W15 Klagenfurt Round of 32?
n=2 winner: Anna Lena Ebster
p=0.10
brier 0.010
p=0.25
brier 0.062
p=0.10
brier 0.010
p=0.10
brier 0.010
p=0.10
brier 0.010
SportsWho won the Rocha vs Johns ATP Challenger tennis match on May 10, 2026?
n=2 winner: Garrett Johns
p=0.10
brier 0.010
p=0.25
brier 0.062
p=0.10
brier 0.010
p=0.10
brier 0.010
p=0.10
brier 0.010
SportsWho won the Perez vs Lalami Laaroussi ATP Challenger tennis match on May 5, 2026?
n=2 winner: Younes Lalami Laaroussi
p=0.10
brier 0.010
p=0.25
brier 0.062
p=0.25
brier 0.062
p=0.10
brier 0.010
p=0.10
brier 0.010
SportsWho won the Worcestershire vs Durham 2026 County Championship cricket match?
n=2 winner: Durham
p=0.10
brier 0.010
p=0.25
brier 0.062
p=0.10
brier 0.010
p=0.10
brier 0.010
p=0.10
brier 0.010
SportsWho won the Bangladesh vs Pakistan men's Test cricket match starting May 8, 2026?
n=2 winner: Bangladesh
p=0.90
brier 0.010
p=0.75
brier 0.062
p=0.90
brier 0.010
p=0.90
brier 0.010
p=0.90
brier 0.010
SportsWho won the Glamorgan vs Somerset 2026 County Championship cricket match?
n=2 winner: Glamorgan
p=0.90
brier 0.010
p=0.75
brier 0.062
p=0.90
brier 0.010
p=0.90
brier 0.010
p=0.90
brier 0.010
SportsWho won the Los Angeles Lakers vs Oklahoma City NBA Round 2 playoff series in 2026?
n=2 winner: Oklahoma City
p=0.90
brier 0.010
p=0.75
brier 0.062
p=0.90
brier 0.010
p=0.90
brier 0.010
p=0.90
brier 0.010
ElectionsWho became Prime Minister of Hungary after the 2026 election?
n=2 winner: Péter Magyar
p=0.90
brier 0.010
p=0.75
brier 0.062
p=0.90
brier 0.010
p=0.90
brier 0.010
p=0.90
brier 0.010
ElectionsWho won the 2026 Democratic primary for Ohio's 15th Congressional District?
n=2 winner: Don Leonard
p=0.90
brier 0.010
p=0.75
brier 0.062
p=0.90
brier 0.010
p=0.90
brier 0.010
p=0.90
brier 0.010
EntertainmentWho was Netflix's next live roast subject after Tom Brady?
n=11 winner: Kevin Hart
p=0.50
brier 0.275
p=0.50
brier —
p=0.50
brier 0.275
p=0.50
brier 0.275
p=0.50
brier 0.275
SportsWho won the Breda vs Heerenveen Eredivisie football match on May 10, 2026?
n=3 winner: Breda
p=0.80
brier 0.060
p=0.67
brier —
p=0.80
brier 0.060
p=0.80
brier 0.060
p=0.80
brier 0.060
EntertainmentWho won Tournament of Champions Season 7 (Food Network)?
n=20 winner: Bryan Voltaggio
p=0.05
brier 0.950
p=0.05
brier —
p=0.05
brier 0.950
p=0.05
brier 0.950
p=0.05
brier 0.950

Source files in data/predictions/ablation_*.json + data/predictions/multi_outcome_retrieval{.phase1_sonnet,}.json. Ground truth in data/resolved.json. Full methodology + per-decision postmortem in docs/DECISIONS.md. ← back to live dashboard