October 2026, -9.71R: Our Worst Month, and What the Approvals Don't Explain

October was the worst month the desk has had this year. It lost 9.71R across 37 trades, with 14 winners against 23 losers and a 37.8% win rate, about 19,400 dollars on a simulated 100,000 dollar account at 2% risk. September traded the same 37 times and made +9.59R, as the September recap showed, so one month took back what the last one added and a little more. There is no single bad instrument to point at, and no one direction to blame. The nine longs lost 5.31R. The 28 shorts lost 4.40R. Five of the six instruments we traded finished in the red, and only the S&P book made money. In entry order the month reached its low at -12.01R on Oct 27, and two winners on the final session brought it back to -9.71R. The system stands at +28.63R YTD from Jan 12 inception, down from the +38.35R we printed through the September close, and the same simulated account sits at $157,278.64. That is roughly where the record stood at the end of August. To judge a month like this one on your own instruments, see SkyAnalyst run your markets.

Oct 1 to 9: five straight stops, then a three-book cluster

The month opened with a Dow short that paid +0.87R and then five straight stops, from a Nasdaq short on Oct 1 to a euro short on Oct 5. Payrolls on Oct 2 came in at 29K against 89K expected, the macro read flipped, and two longs that morning both stopped. In the week of Oct 5 the directional read held up on most days and several entries came early. On Oct 7 the Nasdaq, S&P and Dow books went short inside eighteen minutes, into the same bounce, and all three stopped. That left the month at -6.39R in entry order.

Three straight winners on Oct 8 and 9 repaired part of it, each entered on a completed candle. The stretch closed at -3.37R across 15 trades, six winners against nine losers.

Oct 13 to 23: one good lesson, then the week that did the damage

The next week was the thinnest of the month: four trades, +0.09R, and one of the cleanest waits of the month. After a hot CPI print on Oct 14 the S&P book declined six times with price at or near its entry zone, approved only once a candle had closed through the level, though the volume the plan wanted never arrived, and reached its first target. The week ended with a Dow long on Oct 16 that met its trigger and stopped, the first of six straight losses.

The other five came on Oct 19 and 20, across four books and both directions. The week of Oct 19 finished at about -7.7R across 13 trades, three winners against ten losers, which is most of the month's loss in one window. That is one trade and 1R more than we published at the time, because the euro short of Oct 23 was still open and stopped afterward. By the end of that week the month stood at -11.01R in entry order. We covered it in the October 19 weekly recap and in that week's losses report, and both read it as approvals that had stopped holding to the plans.

Oct 27 to 30: fewer fills, and a better finish

The last week filled five trades from 28 plans. A Dow short approved seven seconds after a decline took the month to its low, and a sterling short on Oct 30 approved on a rejection about 14 minutes old cost another 1R. The other three paid. Friday's Dow short waited through three declines for a closed rejection, and it and the S&P short that followed it forty minutes later both reached their third targets. The week made +1.30R, too small to change the month and too small to prove the habit is back.

Pattern of the month

The pattern of October sits in the words of the approvals. Each week's losses report read the losing entries side by side and counted the ones that flagged their own entry as early or marginal: a candle still forming, a close that had not printed, a rejection that never came, a trigger met only at its minimum. Added up across the month, that is 17 of the 22 losses those reports covered. The 23rd, the euro short of Oct 23, called its trigger largely satisfied and noted the last candle showed indecision.

Those 17 are not all the same fault. Some approved before the plan's condition had printed, and some met its letter and said it was thin. What they share is a trend read carrying an entry the approval itself had doubts about. The strongest version arrived on Oct 23, when a Canadian dollar short declined ten times for lack of a rejection candle and the eleventh evaluation approved at the setup's expiration deadline.

The honest version needs the other column, and it changes the conclusion. By our reading of the free text, at least six of the 14 winners also entered with a gap in the trigger and paid, the Dow short on Oct 14 and the S&P short on Oct 30 among them. Held to one standard on both sides, the gap does not separate winners from losers the way our weekly reports implied. Counting only firm gaps, where the approval itself said the plan's condition had not printed, 16 trades won six times and the other 21 won eight, about 38% in both groups. A looser standard that counts every hedge gives much the same picture: 10 wins from 27 against 4 from 10.

That is a judgment applied to prose across 37 trades, and it could move. It still means we cannot say the approvals explain October, after five weeks of writing that they might. They remain worth fixing, because a trade that does not match its plan teaches us little, but the month's losses do not rest on them alone.

The six losses outside that count matter as well. Several met their written triggers and stopped, and a month that cleans up its approvals will still contain those.

Direction gave no shelter either. Our Nasdaq entries went from about 31,300 on Oct 6 to about 29,000 on Oct 23, while Dow entries went from about 51,100 on Oct 5 to about 52,400 on Oct 23, so the index books were reading markets moving in opposite directions. The Dow book still split in a way the headline hides: its eight shorts won four times and finished at +0.01R, and its five longs won once and lost 3.10R in a month when the Dow rose.

Decision highlights

The desk's patience showed up in the count of evaluations. The 14 trades that sat through at least three declines won seven times and finished at -0.05R, close to flat in the worst month of the year. The 12 trades approved on their first evaluation won three times and lost 6.22R. That is one month, and some of those first reads were good ones, so we would not build a rule on it.

Oct 7 and Oct 30 are the same decision with opposite results. On Oct 7 three index books sold one bounce inside eighteen minutes and took -3R together. On Oct 30 two index books sold the same weak tape forty minutes apart and both paid. Two or three index shorts in the same hour carry one view of the market, and that view costs or pays all at once.

Oct 23 had the two approvals we would most like back. The Nasdaq short was approved with a note that price had already left the entry zone and that the reward to the first target was poor. The Canadian dollar short was approved at its expiration deadline. A deadline or a strong trend read gives no new evidence about the candle in front of the plan, and both trades stopped.

Loss worth learning from

The Dow long on Oct 16

The highest-graded loss of the month was the only B+ setup the desk traded in October, and it carried the highest approval confidence of the month at 79. It is the most useful loss of the 23 because there is so little to fault.

The plan asked the Dow to 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, and the long filled at 52,428.1. Macro and trend both read bullish. Price reached 52,467.7, about 40 points our way against roughly 103 to the first target, then reversed through the stop at 52,322 at 15:22 UTC, thirty minutes after the fill.

We would take that trade again. The same is true of the euro short on Oct 5 and the Dow short on Oct 19, both graded B, both met their written triggers, both stopped. Fixing the approvals would have changed none of these three. Some of a losing month is left over even when the process is clean, and we would rather show that than imply every stop has a fix.

From the desk

October lost 9.71R, about 19,400 dollars on the simulated 100,000 dollar account at 2% risk.

The year to date ledger reads +28.63R across 294 trades at a 55.1% win rate from Jan 12 inception, with October now inside it. On a static 100,000 dollar account at 2% risk that is $157,278.64, down from $176,703.28 at the September close. Compounding the same trades, where each position is sized off the balance it actually had, gives $166,650.60, down from $203,855.48. The gap between the two lines was about 27,150 dollars a month ago and is about 9,370 dollars now. It is the same R through fixed fractional sizing. Sizing off a larger balance earns more on the way up and gives back more on the way down, and October was the way down. What the discipline did this month was keep the risk on every entry at a known fraction of the account through two long streaks.

One note on that record. The euro short entered at 14:43 UTC on Oct 23 gapped through its 1.1210 stop on the Sunday reopen and exited at 1.12263, about 2.6 times its planned risk. The ledger books every stopped trade at -1R by convention, so that slippage appears in no R figure in this article, and the same goes for smaller slippage on other stops.

November starts with the same books and the same 2% risk, and each reads its session from scratch. We are not going to promise a rebound. What we can do is test the approval check honestly, keep the planned risk where it is, and publish the next month in the same format as this one.

What we're tuning

We have written about the same check in every weekly recap this month: hold an approval until the condition the plan named has printed on a completed candle. We are still looking at it, and we have no change to report yet. October does not settle it. Losing approvals flagged doubts and so did winners, so any version has to be tested against both sets of trades before it ships.

Three narrower cases came out of the month. An approval seconds after a decline should need a closed candle in between. A plan's expiration time should never count as a reason to enter. And when a third index book is about to take the same direction inside minutes of the other two, that should be a question asked before the entry. We will say so plainly when any of these changes.