Latest confirmed session
Out-of-sample, slippage charged, same-bar stop-and-target counted as a loss. Daily levels and base rates for NQ, ES and GC — with the sample size on every number, including the ones that embarrass us.
Live ETF proxies.
Not financial advice.
EdgeTape © 2026
Historical rate by regime, VIX, and day-of-week
ORB breakout direction, follow-through, extension
VIX-bucketed performance across all session types
Macro event impact on gap fill and session structure
Replayable archive of historical market sessions
Analog sessions matched by regime, gap, session type
EDGE prediction accuracy vs actual daily outcomes
Daily bias, gap fill odds, expected range before open
Four factors scored at 09:45 ET. When all four agree → HIGH conviction (~68% directional / +1.29R on NQ — preliminary, small out-of-sample sample). Otherwise the honest read is STAND DOWN. A bias, not a win rate.
The Pro Session Bias Map indicator draws this on your chart. Computed from real Globex data. Not financial advice.
The aligned break to watch after 09:30 ET: trade only the side that agrees with the overnight session's direction. Entry at the overnight extreme · stop = ¼ of the range · target 3R.
Computed each morning from real Globex data. Levels are CONTEXT — the naive overnight-break backtest was withdrawn 2026-06 (≈breakeven at realistic fills; see /research). Not financial advice.
NQ breaks its prior-day high/low in the overnight direction while ES does NOT confirm, and the NQ/ES/YM initial-balance triple agrees at 10:30 ET. Both legs traded at half size. Our single out-of-sample-validated lead — small sample, so it is being proven live, in the open.
BETA: positive every calendar year and out-of-sample (IS +0.35R / OOS +0.35R) on 3y of Databento 1-min data with executable fills and slippage. We still call it a lead rather than an edge until the public forward test confirms it (~20–30 trades holding ≥ +0.30R) — and note that the confirmation bar is now close to the backtest itself, because the backtest came down. Integrity note: these figures were revised down on 2026-09-11 (see below), and our earlier overnight-break claim was withdrawn 2026-06 after a fill-model correction — the full record is on the research page.
RE-VERIFIED, AND THE NUMBERS ABOVE CAME DOWN (2026-09-11). Since 2026-09-08 this card carried an open question: research (§55) had found that requiring a second index to confirm a break has the wrong sign once fills are executable — breaks with no index confirming returned +0.57R, breaks with all three confirming −0.12R — and we said our own lead needed re-checking against that. Auditing it turned up something worse and more specific: a lookahead in our own code. The test for “has ES already broken its level?” read the ES bar covering the entry minute — a bar still forming at the moment we decide to trade, so its high and low are not knowable. Restricted to bars that had actually closed, the same rule goes from +0.62R / 65% / PF 3.62 / max-DD 2.4R over 43 trades to +0.35R / 58% / PF 2.13 / max-DD 5.2R over 123. It is still positive every year and still holds out-of-sample, so it survives — as a materially weaker, three-times-more-frequent edge than the one we published. Reproduced three ways, including a clean-room rebuild in another language. The same defect was live in our signal detector, so the forward-test ledger has a seam in it: the rows before 2026-09-11 stay visible and labelled, but the counter restarts at zero from that date — they measured a rule nobody could trade. We did not lower the ≥ +0.30R confirmation bar to suit the weaker backtest; it is now 86% of the corrected estimate and we are publishing that instead. On a $50k / $2,000-trailing-drawdown account at $250/trade the corrected rule still does not breach (+$10,696 over three years) — but it uses 80% of the buffer where the withdrawn version used 30%, and resampling those same days blows the account 12% of the time. See the corrected entry →
NQ futures context before open — Thursday, November 12, 2026
EdgeTape is an independent research site for NQ / ES / GC futures traders. Every setup is tested out-of-sample with costs charged — and we publish the ones that fail.
Built specifically for NQ, ES, and GC — not options, not crypto, not prediction gurus. EDGE focuses on historical probabilities, session replay, and validation-driven accuracy. Every morning read is a function of data — not sentiment, not social media, not narrative.
Designed for daily decision support before the open. Gap fill odds, opening range behavior, volatility regime, macro event impact — structured and repeatable, session after session.
Probability, not prediction
EDGE shows historical base rates. What happened in similar conditions — not what will happen.
Replay + validation
Every session is replayable. Every EDGE prediction is tracked against actual outcomes.
Before the open
The daily read is designed to be consumed in 5 minutes — before 9:30 ET.
Pre-market context, probabilities, and macro setups delivered before the open. Gap fill odds, VIX regime, expected range — all in your inbox.