Live Track Record — US Market Sentiment Watchdog
Every signal we publish, tracked daily for 90 days vs SPY. No hidden losers.
US Market Sentiment Watchdog — Live Track Record
Last updated: May 28, 2026 (auto-generated from the Signal Journal database)
What you're looking at
Every signal we publish in the daily newsletter is registered in a database the moment it goes out. We then track the actual price every day for 90 days and compare it against SPY over the same window. This page is generated directly from that database — no hand-editing, no cherry-picking, no hidden losers.
The metric we report is alpha, not raw return. A 5% gain when SPY is up 6% is a 1% loss to the market. Alpha = (signal return) − (SPY return over the same window). Alpha > 0 means the signal beat the market.
Overview
- Total signals on record: 144
- Currently being tracked (≤90d old): 144
- Closed (>90d): 0
Visual Track Record
The same data as below, in chart form. These regenerate every day from the source database.
Win Rate by Signal Type
Multi-Window Alpha (all signals pooled)
Cumulative Alpha Across All Signals
Recent Signals — Latest Alpha by Type
Performance by Signal Type
Each signal type has a different natural holding period. Reddit sentiment plays decay in days; Edgar insider clusters play out over weeks. We score each type at its native window — and we only count signals that have actually reached that window. Younger signals are listed under In flight and excluded from win rate and alpha.
| Signal Type | Window | Total | Scored | In flight | Win Rate | Mean α | Median α |
|---|---|---|---|---|---|---|---|
| Reddit long calls | T+10 | 53 | 28 | 25 | 54% | +4.18% | +1.16% |
| Reddit short calls | T+10 | 38 | 27 | 11 | 44% | +1.94% | +0.00% |
| Edgar insider buys (HIGH) | T+30 | 12 | 11 | 1 | 27% | -7.14% | -6.74% |
| Edgar insider sells (HIGH) | T+20 | 7 | 7 | 0 | 86% | -3.35% | +3.23% |
| Edgar cluster buys | T+60 | 1 | 0 | 1 | — | — | — |
| Edgar cluster sells | T+30 | 33 | 26 | 7 | 50% | +0.89% | +3.57% |
How to read this: Mean α averages all scored signals; Median α is the middle signal. When mean ≫ median, a few big winners are carrying the average. When mean ≈ median, the edge is consistent across calls. In flight counts signals that haven't yet reached their native window — they'll move into Scored over time.
Multi-Window Alpha (all signal types pooled)
How does the average alpha evolve as we hold longer? This is where you can see when the edge appears and when it decays.
| Window | T+1 | T+3 | T+5 | T+10 | T+20 | T+30 | T+60 | T+90 |
|---|---|---|---|---|---|---|---|---|
| Mean α | +0.6% | +0.3% | -0.1% | -0.4% | +0.2% | +0.0% | — | — |
| N | 156 | 296 | 307 | 295 | 223 | 119 | 0 | 0 |
Most Recent 20 Signals
Every signal, win or lose. No filtering.
| Date | Ticker | Type | Entry | Latest α | Status |
|---|---|---|---|---|---|
| 2026-05-28 | $RKLB | Reddit long calls | $150.23 | no obs yet | open |
| 2026-05-28 | $MU | Reddit long calls | $928.41 | no obs yet | open |
| 2026-05-28 | $CRSR | Reddit long calls | $9.82 | no obs yet | open |
| 2026-05-27 | $RKLB | Reddit short calls | $143.20 | no obs yet | open |
| 2026-05-27 | $MU | Reddit long calls | $895.88 | no obs yet | open |
| 2026-05-26 | $SPY | Reddit short calls | $745.64 | no obs yet | open |
| 2026-05-26 | $QQQ | Reddit short calls | $717.54 | no obs yet | open |
| 2026-05-26 | $NVDA | Reddit long calls | $215.33 | no obs yet | open |
| 2026-05-26 | $NOW | Reddit long calls | $102.13 | no obs yet | open |
| 2026-05-26 | $MSFT | Reddit short calls | $418.57 | no obs yet | open |
| 2026-05-26 | $CRM | Reddit long calls | $180.07 | no obs yet | open |
| 2026-05-26 | $BB | Reddit long calls | $7.91 | no obs yet | open |
| 2026-05-22 | $APP | Reddit long calls | $485.89 | no obs yet | open |
| 2026-05-20 | $PLTR | Reddit long calls | $135.26 | no obs yet | open |
| 2026-05-18 | $NOK | Reddit long calls | $13.95 | no obs yet | open |
| 2026-05-18 | $LULU | Reddit long calls | $119.14 | no obs yet | open |
| 2026-05-15 | $VG | Reddit long calls | $13.01 | no obs yet | open |
| 2026-05-15 | $SPXL | Reddit long calls | $276.90 | no obs yet | open |
| 2026-05-15 | $GOOGL | Reddit short calls | $401.07 | no obs yet | open |
| 2026-05-13 | $UBER | Reddit short calls | $76.36 | no obs yet | open |
Methodology
- Source data: Reddit gap analysis (
processed_data/analysis_*.json) and SEC EDGAR Form 4 insider transactions (edgar_data/edgar_signals_*.json). - Signal trigger: Reddit gaps fire at gap_score ≥ 0.5; Edgar tracks HIGH and CLUSTER tier transactions only. LOW/MEDIUM are noise and not registered.
- Entry price: the close on the day the signal is published.
- Benchmark: SPY close on the same day, tracked at the same windows.
- Alpha: direction-adjusted return minus direction-adjusted SPY return. For short signals, signs are flipped so positive alpha always means we were right.
- Closing: signals are tracked for 90 days, then marked closed.
- No survivorship bias: every signal stays in the database. Removing losing calls is forbidden by the publisher's design.
This page is regenerated from the source database. To audit any single number, query the SQLite file directly. The publisher script lives at agents/scoreboard_publisher.py.