Daytrade Desk
SPY & QQQ hourly options direction — rules decided the night before, replayed and graded on real sessions.
Replay stat strip standing by — no numbers are estimated in its place.
What am I looking at?
- Top tiles — how the overnight signal has performed when replayed on real past sessions: how often it won, how many trades, the average and compounded result.
- Charts — each signal day's gain or loss and the running total, in percent.
- SPY / QQQ tables — the same record day by day: the date, how the day closed, and when the day's low and high printed (Eastern time).
- Options context — what the options market was doing: dealer positioning (GEX), unusually large trades, and yesterday's total volumes.
- Session context — our own measured odds: how gap-up days, morning dips, and winning streaks have actually behaved.
- Everything here is our own derived record on delayed data — no live prices, no advice.
SPY — primary ticker
T+1 REPLAYSPY session trend standing by — no numbers are estimated in its place.
Options context
DERIVEDGEX regime chips standing by — no numbers are estimated in its place.
Unusual options activity standing by — no numbers are estimated in its place.
T+1 options aggregates standing by — no numbers are estimated in its place.
Session context
OUR OWN STATSCatalyst pulse standing by — no numbers are estimated in its place.
measured over 889 sessions through 2026-08-04 (dailies) / 38 sessions (minutes) — docs/research/2026-08-05-spy-gap-dip-timing-stats.md
| ticker | corr | beta | read |
|---|---|---|---|
| QQQ | 0.93 | 1.41 | the twin — confirm/diverge gauge |
| IWM | 0.81 | 1.22 | breadth — small caps confirm broad rallies |
| XLI | 0.72 | 0.95 | old-economy confirm |
| NVDA | 0.66 | 1.87 | loudest single-name echo |
Confirm = QQQ/IWM moving with SPY; SPY up alone is the narrow tape — historically the weak form of a rally. measured over 889 sessions through 2026-08-04 (dailies) / 38 sessions (minutes) — docs/research/2026-08-05-spy-gap-dip-timing-stats.md
QQQ — secondary ticker
Method & odds — how to read this desk
What this desk does. Three overnight rules — thrust2_up, big_up and run3_up — read the prior session's daily bars and set the next session's bias. They fire on roughly 6% of days. Every other day the desk's answer is no trade: chop bleeds options premium, and abstaining is the edge.
The measured record. The rules are replayed over the stored SPY history and regraded as every T+1 bar lands — the scoreboard above IS the live record, with each basis labeled separately (the conditional close-to-close statistic and the executable next-open-to-close expression). Historically the signal cohort has closed up about two sessions in three, against a weaker all-days base. A replay of the past is not a promise about the future.
The options translation. A signal day is played with near-the-money same-day calls. Winners (historically about two of three) have returned +50% to +300% of premium when the session beats its implied move; losers can burn −40% to −100% of premium. Size every trade to survive a total loss.
Timing. Entries buy weakness between 09:30 and 11:15 ET; exits come after 15:00 ET. The session rows above show each replayed day's actual low → high structure, so you can see where those windows landed.
Derived, delayed data · impersonal · not investment advice · full method: the replay write-up.
Desk analysis
Daily round-table analysis standing by — it arms with the analysis plane and refreshes each session close.