Four Trades, +0.09R: The Desk Waited for the Close, and Mostly Got Paid

This was the thinnest week the desk has traded in some time: four entries across five sessions, all of them on the index books. The week turned on Wednesday, when a hot September CPI print pushed yields higher and breadth down to about -1,600. The S&P and Dow books each sold that tape, the S&P on a closed break of a bear flag and the Dow on a failed retest of the post-CPI bounce, and both reached their first targets. The two losses bracketed them. On Tuesday a Nasdaq short went 75 points our way and turned back 27 points short of its first target. On Friday a Dow long entered on exactly the closed candle its plan asked for and stopped thirty minutes later. The window closed at +0.09R, roughly 180 dollars on a simulated 100,000 dollar account at 2% risk. The year to date ledger reads +38.35R YTD through Sep 30, 2026, from Jan 12 inception, sealed at the September close, which puts the same simulated account at $176,703.28. This whole week belongs to October, which is still open and not yet part of that figure.

Tuesday: close enough to VWAP

The week's first trade was a Nasdaq short at 30,504 at 15:06 UTC on Tuesday. The plan asked for a 5-minute close below VWAP after price had tested it from below, and the last completed candle closed at 30,501.1 with VWAP near 30,504, a pass by about three points. The approval came on the first evaluation and named its own doubts: a transitioning trend regime read only 64% bearish, the 15-minute MACD histogram was still positive, and the forming 5-minute candle was carrying stale data from Oct 7. The short worked for a while. Price reached 30,428.9, 75 points in our favor and 27 short of the 30,402 first target, then turned and went through the stop at 30,583 at 16:30 UTC.

Wednesday: two shorts after a hot CPI

CPI landed at 8:30 ET above forecast on both core and headline, and the indices flushed. The S&P plan, written two hours later, wanted a 5-minute close below 7,754 on rising volume and said not to enter on a wick that reversed. Between 14:38 and 14:49 UTC six evaluations watched price trade at or inside the 7,751-7,753 entry zone and declined each time, because no candle had closed below the level. The 14:45 candle finished at 7,749.9, the seventh read approved at confidence 62 while noting that volume was below average, and the short filled at 7,751.1 at 14:56. It reached its first target at 7,736 and later its second at 7,719, and the ledger banks +0.80R.

The Dow short came forty minutes later. The Dow had flushed to 51,532.9 on the print and bounced back to VWAP, and the plan sold a failed retest of the 51,660-51,690 band with its first target at 51,533, a tenth above the morning low. An evaluation at 15:37 UTC declined because the rejection candle had not closed. Eighty-nine seconds later the approval said the candle was "still forming" and called it "a valid but not perfect trigger", and the short filled at 51,652.1. It reached 51,533 that afternoon for +1.30R, the win of the week.

Friday: the trigger printed, the trade did not

The Dow long book had a pullback reclaim plan for Friday: hold the 52,335.8 New York low and VWAP, then close a 5-minute candle back above 52,425. Two evaluations, at 14:40 and 14:49 UTC, declined because that close had not happened. At 14:52 the last closed candle had finished at 52,447.9, the 9 EMA had turned back up and the MACD histogram had improved, and the long filled at 52,428.1. It reached 52,467.7, about 40 points in profit, then reversed through the stop at 52,322 at 15:22 UTC. That took the window from +1.09R to its final +0.09R.

Pattern of the week

The pattern this week was patience at the trigger. Last week we wrote that the trades which waited for a closed trigger candle paid and the ones which entered on a candle still forming mostly did not. This week three of four entries waited for the close, and the S&P short is the cleanest example the desk has produced in a while: six declines in fourteen minutes with price sitting in the entry zone, and an approval only once a candle had closed through the level.

The honest version of the pattern has two exceptions in it, and they point in opposite directions. The Wednesday Dow short entered on a forming candle eighty-nine seconds after a decline, which is exactly the habit we have been criticizing, and it was the best trade of the week. The Friday Dow long waited for its close in full and lost 1R.

Four trades is not enough to prove anything about a rule. What it does show is that the rule changes which losses we take. A trade that meets its trigger and stops is a cost of trading. A trade that skips its trigger is a question about the approval, even when it wins.

Decision highlights

The Wednesday S&P short is the decision of the week. Six evaluations between 14:38 and 14:49 UTC declined it while price traded inside the 7,751-7,753 entry zone, because the plan had asked for a close below 7,754 and ruled out wicks. When the 14:45 candle closed at 7,749.9 the seventh read approved, said in writing that the volume half of the trigger had not arrived, and the trade reached its first target at 7,736 for +0.80R.

The Wednesday Dow short is the decision we would still flag. At 15:37 UTC an evaluation declined because no 5-minute rejection had closed below 51,660, and eighty-nine seconds later the approval went through on a candle it described as still forming. It reached its first target for +1.30R, and a winner does not make an early approval right, so we count it as a correct call taken a little early.

The Friday Dow long is the decision we would take again. Two evaluations declined it because the 5-minute close above 52,425 had not printed, and the approval came only once a candle closed at 52,447.9 with the 9 EMA turning up. It peaked about 40 points in profit and stopped at 52,322 thirty minutes after the fill, which is a 1R cost of trading and not a lapse in the process.

Loss worth learning from

Tuesday's three points

The loss worth learning from this week is the Nasdaq short, because it is the one where the approval had room to say no. The plan asked for a 5-minute close below VWAP. The candle closed at 30,501.1 with VWAP near 30,504, which met the letter of the trigger by about three points. The approval was honest about the rest of the picture: the Trend Agent read the regime as transitioning at only 64% bearish, the 15-minute MACD histogram had not turned negative, and the live candle was showing stale data. It approved on the first evaluation anyway, at confidence 62.

The trade itself was not far off. Price went 75 points our way to 30,428.9, within 27 points of the 30,402 first target, before it turned and stopped at 30,583 eighty-four minutes after the fill. A marginal trigger in a weak trend read is the kind of entry a second evaluation might have passed on, and with the live candle unreliable, a second look would have cost very little.

The Friday Dow long is the other loss, and it is a different kind. That trigger printed cleanly, two evaluations waited for it, and the trade still stopped. We would take it again.

From the desk

The week made 0.09R, roughly 180 dollars on the simulated 100,000 dollar account at 2% risk. Two winners booked at their first targets and two losses at exactly 1R each, on a week when most books did not trade at all.

The year to date ledger reads +38.35R across 257 trades at a 57.59% win rate from Jan 12 inception, sealed at the September close. Run through a static 100,000 dollar account at 2% risk, that is $176,703.28. Run through the same account compounding, where each trade is sized off the balance it actually had, it is $203,855.48. The gap, a little over 27,000 dollars, is not extra edge. It is the same R banked through fixed fractional sizing, which also means a flat week like this one moves the account by a known, small fraction rather than by a guess.

One piece of housekeeping on the numbers. October is still open, so this week is not in the year to date figure. The ledger has -3.27R recorded for the month so far across 19 trades, and it joins the year to date number when the month closes.

Next week each book reads its session from scratch. The habit we want to keep is the one the S&P book showed on Wednesday: the plan's own condition, closed, before the order goes in.

What we're tuning

For two weeks we have written about making the candle-close conditions in a session plan binding at the approval step, and about approvals that follow a decline with no new closed candle in between. This week had one of the second kind, the Wednesday Dow short, and it won. That does not change the case for the check. A rule that only holds when the trade loses is not a rule.

We are also looking at how an approval handles a data problem on the live candle. On Tuesday the Nasdaq read flagged that the forming candle was stale and approved on its first evaluation regardless. When the freshest data is known to be wrong, the safer default is one more evaluation on a clean candle before the order goes in.