US500 Short: A Bounce Stopped 0.6 Points Short, So the Plan Sold the Break

A selloff that came from earnings

The damage started with mega-cap technology. By the 11:33 AM ET analysis, Alphabet's search revenue miss and negative reactions to Amazon and Intel had turned into broad selling, and the S&P 500 stood at 7605, nearly 1 percent under the prior close of 7678.60. Because those names carry so much index weight, the analysis treated the Macro Agent's bearish call as applying with particular force to this instrument.

Breadth did the confirming. The NYSE advance-decline line sat at -579 against a 5-day EMA of 142, after closing the previous day at +1,052, with a low of -658 on the session. VIX was 19.72 against a prior close of 17.19 and above the prior day's high of 17.36, which the analysis read as aligned with a falling index. The dollar was flat near 101.90 and not a driver. Brent was hot at 113.08, a secondary headwind.

The Macro Agent flagged the snapback

The Macro Agent's row at 11:03 AM ET rated the US indexes group bear at 76 percent. On the S&P 500 itself it was bear at 73 percent, which sat between the Nasdaq 100 at strong bear 81 percent and the Dow at lean bear 59 percent. Tradeability on the S&P 500 was only moderate, 45 out of 100, and the same row warned that the risk of a sharp oversold rally was elevated. The bearish side had the better alignment. It did not have a comfortable entry.

Three plans that never traded

That warning describes the morning's problem. The low at 7595.7 was on the Trend Agent's map by 10:38 AM ET. Selling down there meant selling an extended market, so the automation looked higher. The 10:04 AM ET plan wanted 7631-7638. The 10:34 AM ET plan wanted 7628-7638, and the 11:04 AM ET plan wanted 7628-7632. Price bounced 31.7 points to 7627.4 and rolled over. None of the three zones was touched.

By 11:33 AM ET the bounce was 71 percent gone. The analysis stopped asking where to sell a rally and looked at the level underneath.

Breakdown continuation. This is a pattern professional traders use when a market in a strong trend has already bounced from a low and rolled back over. The earlier low is the line everyone can see. A clean break of it suggests that the buyers who defended it the first time have gone, and the trader sells that failure instead of waiting for a pullback that may not come.

The trigger

The plan asked for "a confirmed 5M close below 7595.7 session low", ideally on volume above the pace of the 15:30 UTC bar. The word confirmed is doing the work. A wick under a session low proves little, since stops tend to rest there and price can trade through them and snap back. A close under it says sellers held the level for the whole bar.

There was a second route. If the break came first, a retest of the broken 7602-7605 area from below that got rejected would also count. The analysis called that the more aggressive confirmation.

The zone and the stop

The sell zone was narrow, 7593-7596, sitting on the level itself. The stop went to 7618, just above a broken swing low at 7617.7, and the analysis asked for a 20 to 25 point stop with VIX near 20. From the fill at 7596, the stop was 22 points away. That put it 9.4 points under the bounce high of 7627.4, so the short would be wrong before price had retraced the whole relief rally.

The targets

TP1 was 7573, TP2 was 7555 and TP3 was 7535, written as about 1.0R, 1.7R and 2.6R. From the actual fill they were worth 1.05R, 1.86R and 2.77R. The plan scored quality at 7.5 out of 10 and wrote its confluence count as 5 of 6, although its own list ticks all six. The entry read called it 6/6. We quote both as written.

Why this one is different

Nearly all of this automation's 21 October setups were written to sell a rally, and our October 22 retracement short is the typical shape. Friday's is one of only two written to sell the break directly. The other is the October 14 bear flag breakdown, the useful comparison later in this article.

SkyAnalyst doesn't favor any single strategy. The same automation that spent Friday morning waiting for a pullback changed to a breakdown once the chart stopped offering one, and the Dow plan running beside it sold a bounce the same morning. The plan is written from the morning's chart and macro read, and it is rewritten when that read stops being true.

What this trade teaches

The plan set a condition and the read set it aside, on the record, before the fill. That is the fact to keep in view, because the outcome makes it easy to forget. At 11:45 AM ET nothing the plan had asked for had happened. The last closed bar was above the level. The current bar was about half a minute old, with four and a half minutes left to run. The alternative, a rejected retest of 7602-7605, needed a break first and could not have occurred yet.

I'd prefer to see a confirmed 5M close below 7595.7 for higher confidence, but current price action at 7595.4 with this momentum provides adequate edge. SkyAnalyst entry evaluation, 15:45 UTC

The read also worked from an older number than the one available. It quoted the Trend Agent at 82 percent, the figure in the plan, from 10:48 AM ET. The Trend Agent's newest row, written three seconds before the read, said 88 percent, which helped the short. The same row said entries should avoid chasing into support and use normal rather than full size. The read does not mention that line, and selling a session low on the touch is close to what it warned against.

What the early entry cost

The fill was 7596, 0.3 points above the level the close was meant to clear. So the short was not yet under the line it was betting on, and price showed why that matters. The worst price logged against the trade was 7606.1, 10.1 points above the fill and 46 percent of the way to the stop. That took price back through the 7600 round number the plan had flagged and past the top of the 7602-7605 retest area. Our same-morning Dow short used 8.2 points of a 119-point stop. This one gave back nearly half of its stop distance, and the record does not say when.

The trade record does not store candles, so we will not claim what the 15:45 UTC bar closed at, or say whether waiting would have produced a better or worse fill. A fill from the retest area might have sat closer to the stop. It also might never have come.

Three early fills out of four

Friday's read was not a one-off. This automation filled four setups in October, and on three of them the entering read said the trigger had not printed. On October 7 it entered without a rejection candle and hit its stop (SL). On October 22 it entered without a clean rejection and reached TP2. On October 30 it entered without the close and reached TP3. The exception was the October 14 breakdown, which took seven reads and 18 minutes from first touch, and entered after a closed candle under its levels, though that read still called the trigger only largely satisfied. It reached TP2.

So the month's record is two wins and a stop from early entries, and one win from a patient one. Friday's read did carry the highest confidence of the four, 74 against 62 on the others, but four trades cannot settle anything. We are watching whether reads that override a written trigger behave differently from reads that do not.

From the desk

We log +1.05R (TP1), booked at 1:32 PM ET, one hour and 46 minutes after the fill. The hero panel shows +2.77R (TP3), where the market had traveled by 2:30 PM ET, two hours and 44 minutes after the fill. The space between those two figures is the cost of closing everything at TP1, and both stay on the page. The lowest price the monitor logged was 7533.9, 1.1 points past TP3, but it finalized the trade a minute later, so that is the end of our record and not the low of the day.

One line in the analysis aged well. Its timing note said the breakdown would work either straight away or in the 2:00-3:30 PM ET power hour, and that price might consolidate through lunch. The fill took the first option. The later targets arrived in the second: TP2 and TP3 both printed after 2:00 PM ET, inside the window the plan named.

Set beside the October 23 Dow short, where the entry turned on one closed candle, Friday shows the same desk making the opposite call on the same kind of question. Both reads explained themselves. We would rather publish an argument you can check than a clean-looking rule the system does not always follow.

This is one trade, not a proof. An entry ahead of the trigger worked on a day when breadth, the agents and the index all pointed the same way. The same decision on a mixed day is how a system pays for a stop, as October 7 showed.