Test the ticker.
Then test the strategy.
A stock’s historical opportunity and a strategy’s universe-wide performance answer different questions. Examine them separately.
Each example uses its earliest saved buy/strong-buy appearance from the completed reference study, not the best hindsight signal. Original publication timestamps remain unverified. Prices, peak opportunities, downside and endpoint returns use the retained historical price snapshot.
| Ticker | Matching period | Stock at close | SPY | Difference |
|---|
Read ticker-study methodology
Entry is the open two trading sessions after the saved signal date. Peak is the highest daily high through September 16, including entry day; no minimum hold applies. Downside before peak day excludes the peak day because daily OHLC cannot establish high/low ordering. Peak-day low and full-window downside remain separate. SPY and stock comparison returns use the same entry-date open and peak-date closing endpoint; neither is a dividend-reinvested total return. Do not subtract SPY’s closing return from the stock’s intraday peak. These studies have no predetermined exit rule, so they do not prove a tradable strategy’s return.
Scroll tables horizontally on smaller screens. “Peak gap” is an opportunity comparison, not synchronized excess return.
| Group | Events / stocks | Avg stock peak | Avg SPY | Avg peak gap | Gap vs 1+ | Avg sessions to peak |
|---|
Peak = highest stock daily high through September 16. SPY = entry-day open to the close on that stock’s peak date. The two endpoints share a date, but not necessarily a time. Peaks are chosen in hindsight; these are gross price opportunities, not realized returns, total returns or portfolio performance.
| Marker | Events / stocks | Avg stock peak | Avg SPY | Avg peak gap | Gap vs 1+ | Avg sessions to peak |
|---|
“Gap vs 1+” subtracts the selected sample’s 1+ average peak gap from each row’s average peak gap. It is a descriptive group difference, not a recommendation weight. Nested groups and marker rows overlap; do not add their returns or event counts.
Same-time comparison and downside
For a synchronized endpoint, use each stock’s close and SPY’s close on the stock’s peak date. This removes the intraday-high versus closing-price mismatch, but not the hindsight selection of the date.
| Group | Avg stock at close | Avg SPY | Avg close excess | Close excess vs 1+ | Median peak | Avg pre-peak downside | Avg full-window downside |
|---|
Pre-peak downside uses dates strictly before peak day; entry-day peaks have no prior-day observation and are omitted from that average. Full-window downside uses entry through September 16, so it may occur after the peak. These are adverse price excursions from entry, not portfolio drawdowns.
Inspect individual signal dates
| Stock | Signal date | New / all active | Entry date / price | Peak date / price | Stock peak | SPY same dates |
|---|
Calculation rules and evidence
Signals are evaluated using historical bars available at each cutoff. This is a historical rule-based simulation, not proof that these alerts were delivered at the time. Saved-scan reference cards above are a separate evidence set and are not pooled with these events.
Count one event per ticker/date when at least one A–H marker becomes newly active versus the prior session. A continuously active marker alone does not create repeated events. For combined groups, count all active markers at each new event: 1+, 2+ and 3+ are nested cohorts. For individual groups, require the named marker to be newly active.
Entry is the open two trading sessions after the signal cutoff, leaving one full intervening session. Peak is the highest daily high from entry through September 16, with the earliest date used for ties. No minimum hold or signal-invalidation exit applies. Events without an eligible entry before the endpoint are excluded, not assigned zero returns. One late JPM event on September 15 is excluded for that reason.
Average each event equally: average stock peak return and average matched SPY return; average paired gap equals their difference. This is not a sum of percentages or a compounded portfolio. Simultaneous and overlapping opportunities may describe the same stock move. Earlier signals have longer remaining observation windows, and unequal time-to-peak can influence differences between groups. No fees, slippage, dividends, taxes or cash accounting are applied.
No universal performance result yet.
The ticker studies above are a selected sample. A universal test must also account for losing signals, overlapping positions, costs, and stocks that left the universe.
Review the proposed test specification ↓- Total return vs benchmark
- Not calculated
- Maximum drawdown
- Not calculated
- Win rate and trade count
- Not calculated
- Out-of-sample performance
- Not calculated
Define the rules.
Keep them fixed.
No minimum hold is required. A repeatable exit rule is.
Universe and historical coverage
Define the eligible stocks and dates before testing. Preserve point-in-time membership and delisted names where available. Flag gaps rather than treating today’s survivors as the historical universe. A saved qualifying-results archive alone is not a complete universe-level record.
Signal, entry and exit policy
Freeze the signal-code version and activation rule, set entry latency, and specify invalidation, stop, target or other repeatable exits. There is no minimum hold requirement. A future peak can be reported as an opportunity diagnostic, but cannot serve as an ex-ante exit rule.
Portfolio accounting and benchmark
Set sizing, capital limits, concurrent-position rules, repeat-signal treatment and cash handling. Include fees and slippage. Compare portfolio and benchmark over identical dates and on the same dividend-adjustment basis; do not label an OHLC price series as certified total return.
Validation and evidence
Use a holdout period or walk-forward evaluation and show losses as well as gains. Version the test specification, data snapshot and results. Retain versioned price snapshots, signal logic and run records. This private preview does not launch tests or write to those services.