The Morning Line

The Record

Was the market right?

When the consensus said 80%, did it happen 80% of the time? This page is the aggregate scorecard, graded automatically every morning.

Looking for what happened on a specific question? That lives on Completed Questions, one row per resolved question.

18,861

Graded forecasts

Every morning number, graded against reality.

11,151 settled questions so far. On the average morning the market said 35% — outcomes landed at 34%. The closer those two numbers, the more honest the odds.

Calibration

Each dot is a band of forecasts. On the dashed line, the market was exactly right; above it, outcomes happened more often than the odds implied; below, less.

0%0%25%25%50%50%75%75%100%100%The market said 4% on average (0–10%): happened 3% of 5,030 forecastsThe market said 15% on average (10–20%): happened 14% of 2,496 forecastsThe market said 25% on average (20–30%): happened 24% of 1,909 forecastsThe market said 35% on average (30–40%): happened 32% of 1,710 forecastsThe market said 45% on average (40–50%): happened 42% of 1,961 forecastsThe market said 55% on average (50–60%): happened 51% of 1,827 forecastsThe market said 64% on average (60–70%): happened 63% of 1,102 forecastsThe market said 75% on average (70–80%): happened 75% of 887 forecastsThe market said 85% on average (80–90%): happened 83% of 864 forecastsThe market said 95% on average (90–100%): happened 95% of 1,075 forecastsWhat the market said (morning odds)How often it happened

The numbers

The market saidForecastsHappenedHit rateAvg. said
0–10%5,0301653%4%
10–20%2,49633714%15%
20–30%1,90945224%25%
30–40%1,71054132%35%
40–50%1,96183142%45%
50–60%1,82792751%55%
60–70%1,10269063%64%
70–80%88766375%75%
80–90%86471383%85%
90–100%1,0751,01895%95%

How the grading works

  • We grade the morning number. Every question is scored against the market-consensus odds in our morning snapshot — the number published before the day happened. We never grade the price from the night before settlement: markets are near-certain by then, and grading those numbers would produce a flattering, meaningless curve.
  • Every morning is a forecast. A question tracked for five mornings contributes five graded forecasts, each scored against the final outcome. Probabilities are the stated outcome’s (yes-side) chance.
  • This is the market’s report card, not ours. We publish the consensus; we don’t forecast. When a 92% misses, that is the market being wrong — and it lands on this page either way.
  • A 70% forecast should fail three times in ten. Perfect calibration doesn’t mean never missing; it means missing exactly as often as the odds said.
The Morning Line publishes probability forecasts built from market-consensus odds — the aggregated judgment of thousands of people with real money on the line, refreshed every morning. Forecasts are for news and informational purposes only and are not financial or investment advice. Probabilities move; the numbers on this page were current when the edition was published.