The Ledger

The Ghost's public accuracy record. Every rhyme the engine names is an implicit claim about the rest of the day; every close is a verdict. We score every one and publish the running total — the misses as loudly as the hits. Honesty is the product.

Across 720 scored sessions (2026-01-12 to 2026-07-06), the rhymes' median finish called the direction of the close correctly 49.9% of the time (359 of 720).

Over the most recent 20 sessions the hit rate is 40.0% (8 of 20). The median tracking error — how far the rhymes' median path sat from the actual path between the anchor and the close — is 0.27% of price.

By ticker

  • SPY55.0% (66/120)
  • QQQ44.2% (53/120)
  • NVDA50.8% (61/120)
  • TSLA48.3% (58/120)
  • AAPL49.2% (59/120)
  • META51.7% (62/120)

What is actually being claimed

At an anchor of roughly 12:30 ET on each session, the engine is handed only the bars printed so far — never anything after the anchor — and names its top 6 rhymes: the historical sessions whose charts look most like today's up to that moment. Each rhyme is rescaled so its price at the anchor equals today's, and the median of where those rhymes finished their days is the claim. A hit means the sign of that median finish matched the sign of the actual finish — did the day close above or below where it stood at the anchor. A coin flips 50%, so read every number here against that bar. And read it for what it is: this measures the visual-similarity memory — whether days that looked alike ended alike — not a trading system, and nothing on this page is a forecast or advice.

Recent sessions (newest first)

DateTickerTop rhymeMedian rhyme finishActual finishVerdict
2026-07-06TSLAEOG 2020-11-04 (A+)+0.65%+0.99%hit
2026-07-06SPYXLV 2017-05-25 (B)+0.04%+0.13%hit
2026-07-06QQQAMT 2019-12-09 (C)-0.21%+0.05%miss
2026-07-06NVDAABBV 2023-09-14 (C)-0.20%-0.16%hit
2026-07-06METACMCSA 2022-01-19 (B)+0.24%+1.17%hit
2026-07-06AAPLKMB 2019-04-30 (A+)+0.54%+0.15%hit
2026-07-02TSLAGEV 2025-12-17 (A+)-0.55%+0.33%miss
2026-07-02SPYXLU 2021-06-29 (A+)-0.17%+0.46%miss
2026-07-02QQQCTSH 2011-06-13 (A+)-0.24%+0.39%miss
2026-07-02NVDAPH 2016-06-24 (A+)+0.44%+0.81%hit
2026-07-02METAREGN 2021-06-28 (A)+0.06%-0.34%miss
2026-07-02AAPLNVDA 2025-05-13 (A)-0.36%+0.22%miss
2026-07-01TSLAHLT 2023-03-10 (B)-0.65%+0.02%miss
2026-07-01SPYSPY 2014-05-05 (A+)+0.07%-0.28%miss
2026-07-01QQQSLB 2014-05-30 (C)-0.06%-0.40%hit
2026-07-01NVDAINTU 2010-08-18 (A+)-0.25%-0.34%hit
2026-07-01METAAMD 2024-03-25 (B)+0.09%-0.76%miss
2026-07-01AAPLHLT 2021-03-11 (A)-0.17%+0.01%miss
2026-06-30TSLAAMAT 2016-09-12 (B)-0.01%+0.72%miss
2026-06-30SPYTLT 2021-06-04 (A+)+0.08%-0.03%miss
2026-06-30QQQXOM 2011-01-03 (A)+0.00%-0.02%hit
2026-06-30NVDACOF 2026-03-04 (D)-0.14%+0.73%miss
2026-06-30METAPANW 2020-08-06 (C)+0.12%+0.44%hit
2026-06-30AAPLWELL 2026-06-23 (A+)+0.18%+0.24%hit
2026-06-29TSLABKR 2021-05-07 (A+)-0.52%+0.56%miss
2026-06-29SPYXLF 2018-06-15 (B)+0.19%+0.08%hit
2026-06-29QQQNTRA 2025-06-03 (A)-0.25%+0.18%miss
2026-06-29NVDAREGN 2009-07-06 (C)+0.49%+0.28%hit
2026-06-29METAETN 2011-08-15 (D)+0.90%-0.45%miss
2026-06-29AAPLDUK 2009-01-15 (B)+0.56%-0.05%miss

Finishes are % from the anchor close. A miss stays on the record — the ledger is only worth anything if it keeps them.