Evidence, not promises

Signal follow-through

A transparent measurement of whether meaningful EventAlpha market moves continued, stalled, or reversed over the following 24 hours.

This is not a win rate, forecast score, or trading return. It tests one narrower question: after EventAlpha’s current meaningful-move criteria are met, does probability keep moving in the same direction during the next day?

Evaluated signals

15411

Paired with a follow-up snapshot

Continued

16%

2463 moved another ≥1pp

Reversed

25%

3840 moved back by ≥1pp

Average follow-through

-0.2pp

Direction-adjusted probability points

Recent evaluated signals

Latest independent qualifying move per market, with its next-day observation.

Rolling 90-day window

Market	Signal move	At signal	Next day	Result
Will 2026 be the hottest year on record? Uncategorised · 15 Sept 2026	-1.0pp	76%	76%	flat
Will OpenAI's valuation hit (LOW) $500B by December 31? Finance · 15 Sept 2026	+1.0pp	7%	7%	flat
Will the Republican Party win the MN-01 House seat? Politics · 15 Sept 2026	+1.0pp	81%	82%	continued
Will the Democratic Party win the CA-49 House seat? Politics · 15 Sept 2026	+1.0pp	96%	96%	flat
Will BNB reach $1000 by December 31, 2026? Crypto · 15 Sept 2026	-1.0pp	18%	20%	reversed
Will Trump repeal Presidential term limits in 2026? Politics · 15 Sept 2026	-1.0pp	2%	2%	flat
Will Israel strike 8 countries in 2026? Politics · 15 Sept 2026	+1.1pp	2%	2%	flat
Will the Democratic Party win the IL-04 House seat? Politics · 15 Sept 2026	+1.2pp	97%	96%	flat

How this is measured

• A signal must move by at least 1 probability point over 24 hours and have at least $10,000 of 24h volume or liquidity.
• Near-certain markets at 1% or 99% are excluded, as are inactive snapshots.
• Repeated observations are reduced to at most one signal per market every 24 hours.
• The follow-up is the first stored snapshot 20–36 hours later.
• A move of less than 1 point either way is classified as flat.

Snapshot gaps and changing market liquidity can affect the sample. Historical behaviour does not establish causality or predict future outcomes. See the full methodology.

Signal follow-through

A transparent measurement of whether meaningful EventAlpha market moves continued, stalled, or reversed over the following 24 hours.

Market

Signal move

At signal

Next day

Result

Will 2026 be the hottest year on record?

Uncategorised · 15 Sept 2026

-1.0pp

76%

flat

Will OpenAI's valuation hit (LOW) $500B by December 31?

Finance · 15 Sept 2026

+1.0pp

flat

Will the Republican Party win the MN-01 House seat?

Politics · 15 Sept 2026

+1.0pp

81%

82%

continued

Will the Democratic Party win the CA-49 House seat?

Politics · 15 Sept 2026

+1.0pp

96%

flat

Will BNB reach $1000 by December 31, 2026?

Crypto · 15 Sept 2026

-1.0pp

18%

20%

reversed

Will Trump repeal Presidential term limits in 2026?

Politics · 15 Sept 2026

-1.0pp

flat

Will Israel strike 8 countries in 2026?

Politics · 15 Sept 2026

+1.1pp

flat

Will the Democratic Party win the IL-04 House seat?

Politics · 15 Sept 2026

+1.2pp

97%

96%

flat

How this is measured

• A signal must move by at least 1 probability point over 24 hours and have at least $10,000 of 24h volume or liquidity.

• Near-certain markets at 1% or 99% are excluded, as are inactive snapshots.

• Repeated observations are reduced to at most one signal per market every 24 hours.

• The follow-up is the first stored snapshot 20–36 hours later.

• A move of less than 1 point either way is classified as flat.

Snapshot gaps and changing market liquidity can affect the sample. Historical behaviour does not establish causality or predict future outcomes. See the full methodology.