Signal

Measuring directional conviction against market outcomes — 2025-12-15 to 2026-04-09

Our pipeline tracks AI-generated signals against real market outcomes. Here's a preview of what premium members see — updated daily with live data.

120+

Stories Tracked

52.4%

Bullish 7d Hit Rate

Example

38.5%

Bearish 7d Hit Rate

Example

94.2%

Signal Clarity

Example

Example data — premium members see live results updated daily.

Ticker Consensus Accuracy

Deduplicated per-ticker view — one vote per ticker per measurement window. This is the honest number.

Window	Clusters	Hit Rate	Confirmed	Missed	Avg Favorable
1d	108	28.7%	31	77	-0.42%
7d	51	52.4%	27	24	+1.14%
30d	14	58.3%	8	6	+2.86%

Best Calls (7-Day Window)

Ticker	Headline	Direction	Excess Return	Published
████	AI chip manufacturer reports record revenue surpassing...	bullish	+18.4%	████
████	Gene therapy breakthrough enters Phase 2 clinical trial...	bullish	+12.7%	████
████	Nuclear SMR contract awarded for major data center proj...	bullish	+9.2%	████

Does Conviction Predict Accuracy?

When multiple stories reinforce the same ticker direction, does accuracy improve?

Conviction	Clusters	Hit Rate	Confirmed
1 signal	112 clusters	34.8%	39
2 signals	34 clusters	41.2%	14
3+ signals	19 clusters	52.6%	10

🔒

Unlock Signal Analytics

Premium members get full access to live consensus tables, conviction curves, best/worst calls, per-signal breakdowns, and daily-updated performance tracking across all coverage sectors.

Upgrade to Premium — $9.99/month

Already premium? Sign in to unlock.

Stories Tracked

51.61%

Bullish 7d Hit Rate

62 clusters

Bearish 7d Hit Rate

5 clusters

94.36%

Signal Clarity

184/195 clear consensus

Median Time-to-Move

40/342 signals crossed 2%

Ticker Consensus Accuracy

Deduplicated per-ticker view. Multiple signals on the same ticker within a measurement window are clustered and scored once. This is the honest number — one vote per ticker per window.

Consensus Hit Rate by Window

Bullish

Bearish

Bullish Consensus (62 clusters at 7d)

Window	Clusters	Hit Rate	Confirmed	Missed	Avg Favorable
1d	108	25.93%	28	80	-0.54%
7d	62	51.61%	32	30	+1.7%
30d	0	<5 clusters	0	0	0%
60d	0	<5 clusters	0	0	0%
90d	0	<5 clusters	0	0	0%

Bearish Consensus (5 clusters at 7d)

Window	Clusters	Hit Rate	Confirmed	Missed	Avg Favorable
1d	6	16.67%	1	5	-0.6%
7d	5	0%	0	5	-2.89%
30d	1	<5 clusters	0	1	-2.47%
60d	1	<5 clusters	0	1	-3.55%
90d	1	<5 clusters	0	1	-5.65%

Does Conviction Predict Accuracy?

When multiple stories reinforce the same ticker direction, does accuracy improve? High-conviction clusters (3+ signals) should outperform single-signal clusters if our multi-source pipeline adds value.

Conviction	Clusters	Hit Rate	Confirmed
1 signal	120 clusters	36.67%	44
2 signals	40 clusters	27.5%	11
3+ signals	24 clusters	25%	6

Maturity Funnel

Most meaningful data comes from longer windows. Stories need time to mature before we can judge them fairly.

1 day

7 days

30 days

60 days

90 days

Best Calls (7-Day Window)

Ticker	Headline	Direction	Excess	Peak	Published
OKLO	Trend Roundup: Nuclear Energy Revival	bullish	+25.63%	+25.63%	2026-04-08
RLAY	AI-Designed Drug Candidate Enters Phase 2 Trial in Record 18...	bullish	+22.25%	—	2026-03-26
BEAM	Clinical Trial Demonstrates Success of Advanced Gene Editing...	bullish	+18.3%	+18.3%	2026-04-08
ORCL	Trend Roundup: AI Infrastructure Acceleration	bullish	+16.36%	+16.36%	2026-04-08
INTC	Intel Joins Elon Musk's Terafab AI Chip Project with Tesla, ...	bullish	+15.72%	+19.11%	2026-04-07

Worst Calls (7-Day Window)

Ticker	Headline	Direction	Excess	Published
SMR	M&A Activity Heats Up in Nuclear Infrastructure Sector Drive...	bullish	-13.73%	2026-04-06
SMR	NuScale Power Highlighted as Top Nuclear Stock Amid Data Cen...	bullish	-13.73%	2026-04-06
IONQ	IBM Demonstrates Quantum Error Correction Below Threshold on...	bullish	-12.81%	2026-03-25
ORCL	Port Washington, Wisconsin, Passes Nation's First Referendum...	bearish	+12.39%	2026-04-07
RGTI	IBM Demonstrates Quantum Error Correction Below Threshold on...	bullish	-10.59%	2026-03-25

▶

Per-Signal Detail

Raw per-story scores. A single ticker may be counted multiple times if multiple stories mention it. Use the Ticker Consensus tables above for the deduplicated view.

Bullish Signals (287 scores)

Window	Scored	Hit Rate	Confirmed	Refuted	Avg Favorable	Avg Peak
1d	155	25.81%	40	56	-0.39%	—
7d	132	62.88%	83	35	+2.63%	+6.91%
30d	No data yet
60d	No data yet
90d	No data yet

Bearish Signals (55 scores)

Window	Scored	Hit Rate	Confirmed	Refuted	Avg Favorable	Avg Peak
1d	23	30.43%	7	9	-0.06%	—
7d	23	13.04%	3	17	-3.08%	+2.53%
30d	3	0%	0	3	-2.47%	—
60d	3	0%	0	3	-3.55%	+2.04%
90d	3	0%	0	3	-5.65%	+2.04%

Methodology

Two views of accuracy: The headline "Ticker Consensus" stats deduplicate by grouping all signals on the same ticker within a measurement window into a single cluster. Each cluster gets one vote. The "Per-Signal Detail" section shows raw per-story scores for granularity.

Clustering: Signals on the same ticker published within N days of each other (where N = measurement window) form a cluster. If 67%+ of signals agree on direction, the cluster is scored. Mixed signals are flagged as "conflicted" and excluded from hit rate.

Conviction: Cluster size (1, 2, 3+ signals) measures conviction strength. If high-conviction clusters outperform, it validates that multi-source reinforcement predicts better outcomes.

Direction-adjusted returns: All "favorable" returns are direction-adjusted. For bullish calls, positive excess return vs SPY is favorable. For bearish calls, negative excess return vs SPY is favorable. Positive always means correct.

Signal clarity: The percentage of ticker-clusters where our pipeline produced a clear directional consensus (not conflicted). Higher clarity = more consistent signal.

Windows: We measure at 1, 7, 30, 60, and 90 trading days after publication. The Fourth Factor detects signals early — 1-day measures noise, 7-30 day measures signal quality.

Data source: All price data from Yahoo Finance. No LLM in scoring — pure math on real market data. Scores computed daily, cannot be retroactively modified.

Threshold: ≥0.5% favorable excess return = "confirmed." ≥0.5% adverse = "missed." Between = "neutral." Hit rates require 5+ clusters to display.

The Fourth Factor provides technology trend analysis and is not a registered investment advisor. All content is for informational purposes only and does not constitute financial advice. Past tracking accuracy does not guarantee future results. Always conduct your own research before making investment decisions.