43.08%
7-Day Accuracy
441 signals
+9.06%
Avg Alpha When Right
vs SPY benchmark
1.42x
Win/Loss Ratio
Premium
+4.2pp
Consensus Edge
Premium
Methodology
How we measure: Every story generates a directional signal (bullish or bearish) on a specific ticker. We cluster signals on the same ticker within each measurement window and score the consensus direction against actual market outcomes, measured as excess return vs SPY.
What counts: ≥0.5% favorable excess return = outperformed. ≥0.5% adverse = underperformed. We require 10+ signals per window before reporting accuracy rates.
No LLM in scoring: All price data from Yahoo Finance. Scoring is pure math on real market data, computed daily, and cannot be retroactively modified.
Sample: Best Call
| Ticker | Headline | Direction | Alpha | Sources | Published |
|---|
| INTC |
Commerce Department Proposes Expanded AI Chip Export Fr... |
bullish |
+168.02% |
1 |
2026-03-27 |
Performance by Horizon
| Window | Signals | Accuracy | Avg Alpha Win | Avg Alpha Loss | W/L Ratio | Expected Alpha |
|---|
| 1d |
660 |
36.21% |
+5.23% |
-3.49% |
1.5x |
-0.33% |
| 7d |
441 |
43.08% |
+9.06% |
-6.5% |
1.39x |
+0.2% |
| 30d |
187 |
39.57% |
+20.27% |
-14.5% |
1.4x |
-0.74% |
| 60d |
28 |
42.86% |
+45.47% |
-10.92% |
4.16x |
+13.25% |
🔒
Unlock Full Signal Analytics
Premium members get complete performance tables, win/loss ratios, consensus edge analysis, directional breakdowns, notable calls, and weekly analyst commentary.
Upgrade to Premium — $9.99/month
Already premium? Sign in to unlock.
43.08%
7-Day Accuracy
441 signals
+9.06%
Avg Alpha When Right
vs SPY benchmark
1.39x
Win/Loss Ratio
Positive expected value
+8.86pp
Consensus Edge
Multi-source vs single
Performance by Horizon
Accuracy and alpha metrics across measurement windows. Only windows with 10+ signals shown.
| Window | Signals | Accuracy | Avg Alpha Win | Avg Alpha Loss | W/L Ratio | Expected Alpha |
|---|
| 1d |
660 |
36.21% |
+5.23% |
-3.49% |
1.5x |
-0.33% |
| 7d |
441 |
43.08% |
+9.06% |
-6.5% |
1.39x |
+0.2% |
| 30d |
187 |
39.57% |
+20.27% |
-14.5% |
1.4x |
-0.74% |
| 60d |
28 |
42.86% |
+45.47% |
-10.92% |
4.16x |
+13.25% |
Consensus Strength
Does coverage from multiple independent sources improve signal accuracy?
Multi-source signals outperform single-source by 8.86 percentage points.
1322 total signals scored. This metric stabilizes as more data accumulates.
1 source
36.41%
931 signals
2 sources
38.5%
200 signals
3+ sources
52.36%
191 signals
Directional Breakdown
Bullish and bearish signal accuracy, reported separately. Only windows with 10+ signals shown.
Bullish (2496 raw scores)
| Window | Signals | Accuracy | Avg Alpha |
|---|
| 1d |
605 |
36.03% |
+0.59% |
| 7d |
409 |
43.28% |
+0.73% |
| 30d |
174 |
41.38% |
+0.63% |
| 60d |
27 |
44.44% |
+13.87% |
Bearish (389 raw scores)
| Window | Signals | Accuracy | Avg Alpha |
|---|
| 1d |
55 |
38.18% |
+1.56% |
| 7d |
32 |
40.63% |
-0.68% |
| 30d |
13 |
15.38% |
-13.56% |
Notable Calls
Each call is evaluated using its longest matured window (1d / 7d / 30d / 60d / 90d). Window shown next to each alpha.
Top Performers
| Ticker | Headline | Direction | Alpha | Sources | Published |
|---|
| INTC |
Commerce Department Proposes Expanded AI Chip Export Fr... |
bullish |
+168.02% 60d |
1 |
2026-03-27 |
| 066570.KS |
LG Electronics in Talks with Nvidia on Robotics, AI Dat... |
bullish |
+109.45% 30d |
1 |
2026-04-29 |
| RXT |
Rackspace Technology and AMD Sign MOU to Deliver Govern... |
bullish |
+104.83% 7d |
1 |
2026-05-07 |
| ANTH |
Anthropic in Talks with Trump Administration About Next... |
bullish |
+91.81% 30d |
1 |
2026-04-13 |
| RIOT |
Sovereign Wealth Fund Disclosure Reveals $800M Bitcoin ... |
bullish |
+85.38% 60d |
1 |
2026-03-27 |
Worst Performers
| Ticker | Headline | Direction | Alpha | Sources | Published |
|---|
| BIRD |
Allbirds Shares Jump Over 400% on Plans to Pivot from S... |
bullish |
-79.65% 30d |
3 |
2026-04-15 |
| GIG |
Hadron Energy Secures Strategic Investment and Multi-Pr... |
bullish |
-77.85% 30d |
1 |
2026-04-28 |
| ONCO |
Realbotix Launches Vinci AI Vision System with First De... |
bullish |
-75.78% 30d |
1 |
2026-04-09 |
| ARM |
TSMC Fully Exits Arm Holdings Investment with $231 Mill... |
bearish |
-68.85% 30d |
1 |
2026-04-29 |
| XNDU |
Nvidia Backing Propels Xanadu Quantum Technologies Stoc... |
bullish |
-63.75% 30d |
2 |
2026-04-17 |
Weekly Commentary
This week's 41.69% accuracy across 427 signals marks another period of net underperformance, with 221 signals trailing their benchmarks against 178 that led. The distribution shows no material improvement from prior intervals, as the pipeline continues to generate more negative outcomes than positive ones even after filtering for volume.
Cumulative results reinforce the pattern. Across all horizons, accuracy sits between 36.79% and 41.94%, with expected alpha remaining negative at every interval from one day through 90 days. The 7-day window's 1.3x win/loss ratio and the 30-day window's 1.33x ratio have not translated into positive expectancy, while the small samples at 60 and 90 days show outright negative skew.
Multi-source signals continue to separate from the rest. Accuracy rises from 35.75% on single-source items to 54.09% when three or more sources align, producing the reported 10.68 percentage point consensus edge. This differential persists even as overall accuracy stays below 50%, indicating the pipeline's primary statistical advantage remains concentrated in the higher-consensus subset.
Signals carrying three or more sources that reach the 7-day and 30-day marks will provide the clearest test of whether that edge compounds or erodes. The data continue to show that detection strength is highest in those consensus cohorts and weakest in low-source or extended-horizon cases.
Methodology
Signal clustering: Signals on the same ticker published within N days of each other (where N = measurement window) form a cluster. If 67%+ agree on direction, the cluster is scored as a single signal. Mixed signals are excluded.
Accuracy rate: The percentage of scored clusters where the consensus direction outperformed SPY by ≥0.5% excess return. Underperformed = ≥0.5% adverse excess return. Between = neutral (excluded from accuracy).
Alpha metrics: Avg Alpha Win is the mean excess return on correct calls. Avg Alpha Loss is the mean excess return on incorrect calls. Win/Loss Ratio = Avg Win / Avg Loss — must exceed 1.0 for positive expected value. Expected Alpha = (accuracy × avg win) − ((1 − accuracy) × avg loss).
Consensus edge: The accuracy improvement when multiple independent sources reinforce the same signal vs single-source calls. A positive edge validates the multi-source pipeline.
Measurement windows: 1, 7, 30, 60, and 90 trading days. Short windows (1d) capture noise; 7-30d windows capture signal quality. Rates require 10+ signals to display.
Data integrity: All price data from Yahoo Finance. No LLM involvement in scoring — pure math on real market data. Scores computed daily and cannot be retroactively modified.