summary: the opening candle continuation report on its own gives you a modest directional read. add a second checkpoint at noon and the days where both agree close in that direction far more often. here are the two checks, how to pull the data yourself, why the red side behaves differently from the green side, and what this bias can't do.

the opening candle continuation report answers one question: does the session close in the same direction as its opening candle? run on a 15-minute opening candle, the answer on NQ is a modest majority — real, but not much to build an afternoon around on its own.

this method adds one more check. at noon, ask whether price is still on the same side of the 9:30 open. on the days those two checks agree, the session closed in that direction a great deal more often than the opening candle alone would suggest. on the days they disagree, there's no measurable bias at all, and the method's answer is "sit on your hands."

be clear about what this is: a directional bias, not a setup. there's no entry, no stop, and no target here. it's context you carry into your afternoon trades, and you still need a setup to execute.

both green, or both red, and you have your afternoon bias. one of each, and you don't.

three reads are worth pulling for your own ticker and session, in this order:

1. the opening candle on its own — how often the session closes in the same direction as the first 15 minutes. on NQ this comes in as a modest majority: better than a coin flip, not by a wide margin.

2. whether the direction still holds at noon. this is where it gets interesting, because the two sides behave completely differently. a green 15-minute open usually is still green at noon, at about the same rate as the full-session read. a red open is a coin flip — it tells you essentially nothing about where price sits at noon.

that asymmetry is the most useful thing in this article. red mornings chop. so on the red side the noon checkpoint isn't a nice-to-have confirmation — it's doing nearly all of the work.

3. the days both checkpoints agree. filter to those and the close follows the same direction a large majority of the time — a substantial jump over the opening candle read on its own, and comparable on both the green and red sides once the noon check has done its filtering.

when they disagree, the day closes green about as often as red. that's a genuine coin flip, and it's exactly what you want a filter to tell you — on those days, this method's honest answer is that it has nothing to say.

this method narrows twice — first to one color of opening candle, then to the days the noon check agrees. that leaves a much smaller sample than the headline report, so the number you read will move around depending on the window you pull.

run it over six months and again over three, and compare. a short window can make one side look dramatically stronger than the other purely because a handful of sessions fell one way — a rate built on a dozen or so days is not the same animal as one built on fifty, even when the percentages look similar. always read the occurrence count next to the rate, and plan against the longer window.

yes, at a similar level, with one difference worth knowing: ES red opens hold their direction into noon more often than NQ's do. ES chops a little less after a red open, which makes its red side hold up better than NQ's before the noon filter is applied.

as always, run it on the ticker you actually trade rather than assuming NQ's behaviour transfers.

on the edgeful API, both checkpoints come from the same endpoint: opening-candle-continuation-standard with candle_size=15min. run it once with end_time=16:00:00 for the close and again with end_time=12:00:00 for the noon check, then join the per-day detailed rows on date. that join is what gives you the both-agree sample, which no single report view will show you. the per-day rows come back on pro and all access keys; essential keys get summary numbers only, and opening candle continuation isn't one of essential's three API reports.