The NAS100 short the macro stack pre-cleared in one read

The macro tape was already decided when we opened the chart

The 12:30 UTC NFP release was the catalyst, and the bond market priced it inside the first thirty minutes. Ten-year yields jumped 6.1 basis points off yesterday's close to print 4.546% by the New York morning, a fresh five-day high. The dollar followed. DXY cleared 99.80 and made its own five-day high at 99.867. Gold dumped 89 points, oil softened to 93.81, the $ADD sat at -829 with the day's low at -975. Three of those four cross-asset reads we watch are pure risk-off, and the fourth, gold, is rate-driven liquidation rather than safe-haven bidding. The Macro Agent flagged the read as bearish at -74 with 74% confidence, citing the NFP-driven rate repricing in its own words.

Against that backdrop, the NAS100 had already gapped open ~550 points below the prior cash close and printed a session low of 29,724.7 inside the first hour. By the time we ran our 14:36 evaluation, price had corrected back to 29,876, retracing about half of the opening drop into the 29,860 to 29,907 zone. Discretionary traders would have looked at that bounce and called it a coin flip. The five-minute rejection candle, formed inside the Fibonacci confluence on rising volume, was already on the chart by the time we evaluated.

We took the read at 14:36 UTC. If you want to see the same engine grade a setup on your watchlist, you can [start a 7-day free trial](https://skyanalyst.com/pricing) and run it on a chart of your choosing this afternoon.

What every agent saw, in the same direction

The Macro Agent rated tradeability at 78 out of 100, the highest band we use for index shorts in this regime. The Trend Agent registered NAS100 as bearish at 84% confidence, with the 60-minute EMA stack fully aligned: price 244 points below session VWAP, RSI at 30.5 lifting from oversold, MACD histogram at -30 and expanding. The 15-minute view backed it up at 32.5 RSI with MACD also expanding bearish. The five-minute structure showed the bounce off the session low rolling over inside the 38.2% to 50% Fibonacci band of the opening drop. There was no signal in the system pointing the other way. That is the part worth pausing on.

The setup the Trend Agent flagged at 14:36 has a name among professional traders: a bounce-rejection short in a confirmed downtrend with macro tailwind. It is the mirror image of the patterns we publish more often in this journal, and it is worth a minute on the mechanics. The pattern is teachable, the math behind it is fair, and the way our system reads it is part of why the entry happened on a single evaluation rather than four.

What the pattern is

Price has been falling on the 60-minute chart with momentum. Inside the first hour or two after a high-impact catalyst, the move overextends and prints a session low. From that low a corrective bounce develops, usually retracing 38.2% to 61.8% of the most recent impulsive leg. The bounce-rejection trade is not entered on the touch of the Fibonacci band. It is entered on the rejection: a five-minute candle that prints inside the band, fails to reclaim the broken support, and closes lower with volume above the 60-period average. The pattern names the tape's failure to absorb buyers as the trigger, not the level itself.

How professional traders use it

Discretionary index traders use this pattern as a continuation-trade entry on strong trend days. The math favors it because the corrective bounce after an impulsive move is usually thin: the original sellers are not closing their books because the catalyst that produced the move has not reversed. So the bounce attracts late longs from traders who think the gap will fill, and gets sold by the original participants on the first sign of failure. Rough rule of thumb among the desks that publish on this kind of pattern: a clean rejection inside the 38.2 to 61.8 band with volume confirmation completes in the trend direction 60 to 65% of the time on strong-trend days, and the R:R skews high because the stop sits just above the rejection level while the target sits at the prior low or beyond.

Why it works at the structural level

The corrective bounce after a catalyst-driven move is the market probing for absorption. If real buyers are present they leave footprints: stalled price at the prior breakdown level, contraction of volatility, a clean reclaim of the broken zone. When those footprints do not appear, the bounce is purely positioning. The rejection candle is the visible moment when the late longs are forced to recognize there is no buy-side liquidity defending their entry, and the trend resumes. It fails, like every pattern, in the wrong regime. A risk-on tape with falling yields and a soft dollar turns the same setup into a base-and-rip continuation higher. That is why the macro grade has to clear before the trend agent is allowed to size in at all. On this trade, the macro grade was as clean as we get.

How the system sees it: dynamically, not dogmatically

SkyAnalyst does not favor the bounce-rejection short, or any single pattern. On the same trading day, our agents were scoring a pullback-to-reclaim long on US30 in a different macro window, watching a USDJPY long off intraday support, and grading a US500 short on a structurally distinct setup. Each of those is a different playbook with a different edge profile and a different failure mode. The Trend Agent reads the tape first and fits the available pattern to what is actually there.

That is the single biggest difference between how the system trades and how a retail desk that has fallen in love with one setup tends to lose money. A trader who only runs the bounce-rejection short will eventually try to force it onto a chart where the trend is not confirmed, the macro is not aligned, and the rejection candle is wishful thinking. The system has no favorite. Every evaluation re-reads the regime, re-scores the structure, and lets the confluence math decide which playbook applies, if any. Some sessions, none do, and we sit them out. The four agents running in parallel each contribute a different lens on what kind of market is in front of us and what strategy it rewards right now. When they agree, we trade. When they don't, we don't. The NAS100 short was one shape of that agreement. The next case study will almost certainly look different, because the system reads the tape first and the system is dynamic, not dogmatic.

What this trade teaches and what it does not

We do not publish a case study because a trade printed +8.61R (TP3). We publish a case study because there is something repeatable to point at inside it. The repeatable thing here is not the size of the result. The repeatable thing is the structural clarity that produced a single-evaluation entry, and the discipline of separating that clarity from the outcome.

Confluence math, not conviction

The 14:36 evaluation cleared on the first scan because seven of seven factors were genuinely present at the bar close. Yields, dollar, volatility, breadth, trend structure, lower-timeframe momentum, calendar. That is rare. It is rare because most days at least one input is ambiguous, and ambiguity drops the score below the threshold. When ambiguity is absent, the system does not need multiple evaluations to build a case. The score is the case. Discretionary traders read this kind of moment as "high conviction." The system reads it as math.

The size of the print is the tape, not the system

+8.61R (TP3) is the largest single-trade R-multiple in our quarter-to-date journal. It is not what an average trade looks like. Our quarter-to-date net R is -1.82R across 36 trades. Our win rate is 47.2%. The risk policy is fixed: one R of risk per trade, regardless of confidence. When the market gives an asymmetric payoff like this, the policy lets us collect it. When it does not, the same policy caps the damage at -1R. The expectancy of the system is built across hundreds of trades, not on top of any one of them. Showing readers the +8.61R (TP3) print without the -1.82R running net would be advertising. We are publishing both.

What we are watching going into tomorrow

The setup that produced this trade is rare because the macro stack lined up. If tomorrow's session opens with the same configuration we expect the system to behave the same way, and if it opens with a more ambiguous tape we expect the system to wait. The interesting question is never whether the next trade will work. The interesting question is whether the system continues to act the same way regardless of what just happened. Same threshold, same math, same risk per trade. That is the consistency we are building toward and the consistency the journal is here to document.

A note, before we move on

We almost waited a day before filing this one. A +8.61R (TP3) intraday short on a single-evaluation entry is not the median case study, and we are wary of leading with outliers. The more honest article would have been about one of our 4-evaluation entries that finished at TP1, or one of the stop-outs from earlier in the quarter where the macro read was right but the timing was wrong. Those are coming.

We chose this trade because it is the cleanest example of how the four agents are supposed to coordinate when the tape itself is clean. The Macro Agent had written its bearish read to shared state at 14:30 UTC, six minutes before the Trend Agent's evaluation, and had not revised it. The Trend Agent's 14:36 pass read that value and used it to unlock the size it eventually took. If the Macro Agent had been ambivalent, expressing in prose that "yields are rising but the dollar is mixed and breadth is unclear," the Trend Agent would have had to interpret a tone instead of a value. It does not. So it did not. The coordination between the four agents is the product. It is what a chat interface with one model cannot simulate, and it is what this case study shows in practice.

A reasonable question for any reader is whether they could replicate what the system did here using a chat tool and the same data. They could read the same chart, see the same Fibonacci confluence, look at the same yields and dollar reads. They would still be missing the structural piece: the four agents producing four typed reads on a shared schedule, the rule that all four must clear thresholds before size is allowed, the auditable log that gets published whether the trade wins or loses. That structure is the product. The +8.61R (TP3) print is one shape of what it produces when conditions cooperate.

The next case study will almost certainly be a smaller R-multiple on a noisier tape. That is the median trade, and the median trade is what the system is built to repeat.

The SkyAnalyst Team