Six Stops in Three Days, and One Index Short Taken Three Times
Six losses at exactly 1R each, all before Thursday. Three were the same index short, filled across three books in eighteen minutes and stopped by one bounce.
Before the trade, meet the system.
SkyAnalyst is not one AI trader. It is four specialist agents, each with its own data pipeline, each maintaining state between evaluations, and each required to agree before a position is sized. They don’t chat in prose. They write structured messages to a shared state object that each reads on every evaluation cycle.
This is the losses report for a week that lost. Eleven trades, five winners, six losers, and every loser cost exactly the 1R it was sized for. All six arrived in the first three days, and by the time the last of them stopped at 17:08 UTC on Wednesday, the window had given back 6.0R in losses against two Cable winners, a simulated 12,000 dollars of losing trades on a 100,000 dollar account at 2% risk. The direction was mostly right. A firm dollar, firm yields and, on Wednesday, breadth that collapsed to -1,684 ahead of the FOMC minutes all pointed the way the desk leaned. The entries were the problem. Three of the six losing approvals said in writing that the trigger had not fully printed, a fourth went in with momentum worsening against it, a fifth met only the minimum version of its trigger, and the sixth met its trigger in full and still stopped. Against the year to date ledger the scale is plain. The system has banked +38.35R YTD from Jan 12 inception, sealed at the September close, which puts the same simulated account at $176,703.28. The desk gave back 6R in losing trades this week against +38.35R YTD, and the week as a whole closed at -1.24R once the winners are counted.
Monday: one early entry, one clean one
At 14:14 UTC the Dow short book sold US30 at 51,099.2 on its fourth evaluation. The three declines before it all said the 5-minute rejection had not closed, and the approval said the same thing, that the signal was workable even if the rejection candle was not fully closed yet. Price went 55 points our way and came back through the 51,208 stop, closing at 51,209.4 ninety-eight minutes later.
Seventeen minutes after that Dow fill, the euro short filled at 1.11942. This one did what its plan asked. It is the highest-grade loss of the week, and one of the two we tear down below. A Cable short later that afternoon reached its first target and was Monday's only winner.
Tuesday: a long at the highs
One trade. The Nasdaq long book bought NAS100 at 31,338.1 with the index at all-time highs. The approval noted that the 5-minute MACD histogram was negative and worsening and that price had not convincingly reclaimed 31,345. The trade saw 31,346.3, eight points of room, and stopped thirty-five minutes after the fill.
Wednesday: three shorts in eighteen minutes
The Cable book opened the afternoon well, waiting out a chase and entering only on a completed rejection candle. Then the index books went short together. The Nasdaq short filled at 31,054.3 at 15:19 UTC, eighty-three seconds after an evaluation declined it. The S&P short filled at 7,785 at 15:37, seventy-nine seconds after a decline that said the rejection candle had not formed. The Dow short filled at 51,054.7 the same minute, twenty-three seconds after its plan was written. The 60-minute RSI on the Nasdaq was 26. All three stopped, and the equity line hit its low for the week at 91,478.54. Thursday and Friday brought three winners and no losses.
Every trade the system took.
| Date | Time | Instrument | Dir | Model | Setup | Grade | R | $ Sim | Result | Details |
|---|---|---|---|---|---|---|---|---|---|---|
| Oct 5 | 14:14 UTC | US30 | Short | US30 SHORT retracement failure | C+ | -1.0R(SL) | -$2,000(SL) | Stop hit | - | |
| Oct 5 | 14:31 UTC | EURUSD | Short | EURUSD Short Sell on Retrace | B | -1.0R(SL) | -$2,000(SL) | Stop hit | - | |
| Oct 6 | 15:55 UTC | NAS100 | Long | Claude Opus 4.7 | NAS100 LONG, Pullback Buy on Trend Continuation | C+ | -1.0R(SL) | -$2,000(SL) | Stop hit | - |
| Oct 7 | 15:19 UTC | NAS100 | Short | NAS100 SHORT, VWAP Rejection / Failed Rally into Resistance | C+ | -1.0R(SL) | -$2,000(SL) | Stop hit | - | |
| Oct 7 | 15:37 UTC | US500 | Short | US500 SHORT, Bearish Trend Continuation at Yesterday's Low Rejection | C+ | -1.0R(SL) | -$2,000(SL) | Stop hit | - | |
| Oct 7 | 15:37 UTC | US30 | Short | US30 sell-the-bounce pullback fade | C+ | -1.0R(SL) | -$2,000(SL) | Stop hit | - |
Dollar figures are simulated on a $100,000 account at 2% risk per trade. Actual subscriber P&L varies with account size. Past performance is not a guarantee of future results.
Pattern of the week
The pattern is the approval, not the instrument and not the direction.
Read the six losing approvals in order and five of them describe a trade that was not quite ready. The Monday Dow short entered while its rejection candle was still open. The Tuesday Nasdaq long entered with momentum worsening and the reclaim of 31,345 unconfirmed. The Wednesday Nasdaq and S&P shorts were each approved within ninety seconds of a decline, with no new closed candle in between. The Wednesday Dow short met only the minimum version of its trigger, at the lower edge of its zone, and said so.
The three winners from Thursday and Friday read the other way. Each entered after a closed 5-minute candle had printed the trigger its plan named, and the Friday euro short waited through six declines in eight minutes to get there.
This is the third losses report in a row that lands on this observation. Three weeks is still a small sample, but the trades that waited for the candle are the ones that keep getting paid.
Decision highlights
The Wednesday cluster is the decision we would take back as a whole. Between 15:19 and 15:37 UTC the Nasdaq, S&P and Dow books all went short, each on its own reading of its own tape, but all three were reading the same broad selloff, and the same 60-minute oversold bounce took out all three for -3R. Nothing at the desk level noticed that the third index book was about to take the same trade as the other two.
The Wednesday S&P short changed its mind in seventy-nine seconds. At 15:35 UTC the evaluation declined because price had not rallied into the 7,781-7,785 pocket to form a rejection, and at 15:37 the approval said the rejection candle had still not formed. The fill came at 7,785, and the trade stopped forty-five minutes later without ever moving in our favor.
The Monday euro short is the decision we would repeat. Four evaluations declined it between 14:03 and 14:09 UTC because the time gate in its plan had not opened and the last candle had closed above 1.11940. It approved only once a completed candle wicked into the zone and closed at 1.11927, under the trigger, and it is the one loss this week we would take again exactly as it was taken.
Six instruments, six stories.
One loss, the Monday short at 1.11942. It met its trigger in full, never got three pips into profit, and stopped late in the day.
All EURUSD this week →No losses this week. Both Cable shorts reached their targets, and the book was the strongest of the window.
All GBPUSD this week →Two losses, a Monday short approved before its rejection candle closed and a Wednesday short filled twenty-three seconds after its plan was written.
All US30 this week →Two losses, a Tuesday long at all-time highs and a Wednesday short into a bounce, approved eighty-three seconds after a decline.
All NAS100 this week →No trades this week. The yen book did not produce a qualifying setup.
All USDJPY this week →One loss, the Wednesday short at 7,785, approved seventy-nine seconds after a decline. It never went a point our way and stopped at 7,797.1.
All US500 this week →No trades this week. USDCAD remains the newest book and has not yet produced a qualifying setup.
All USDCAD this week →Loss of the week: US30 Short · -1R
Losses worth learning from
EURUSD Short, Monday, -1R, grade B
The highest-grade loss of the week, and the cleanest.
What was right
Nearly everything. The plan wanted a 5-minute rejection closing below 1.11940 after a touch of the 1.1195-1.1200 zone, and only after 10:10 AM ET. Four evaluations declined it while the time gate was closed or the candle sat above the level. At 14:10 UTC the approval came at 72% confidence: the 14:05 candle wicked to 1.11972 and closed at 1.11927, price was below both 5-minute EMAs and VWAP, and the dollar and yields were rising.
What was wrong
Only the outcome. The position filled at 1.11942 at 14:31 UTC, saw 1.11915, under three pips of room, and turned. The stop sat at 1.1224 and the trade closed at 1.12257 at 20:50 UTC, a little past the level, and is scored at the 1R it was sized for.
What we'd do the same
All of it. This is the cost of trading a probabilistic edge, and it is what the 2% risk per trade is sized for.
US30 Short, Wednesday, -1R, grade C+
What was right
The read. The Dow sat below its 15-minute EMAs and far below VWAP, breadth was at -1,684, and the latest completed 5-minute candle pushed to 51,045.3 and closed at 51,043.2, back under 51,045. That is the minimum version of a failed-reclaim trigger.
What was wrong
The approval came twenty-three seconds after the plan was written and called the entry valid but not perfect. It noted that price had only tagged the lower edge of the zone instead of rejecting near the 61.8% level around 51,055, that the 5-minute bounce still had positive momentum, and that the trade deserved reduced aggressiveness. The desk sizes every position at 1R, so that note had nowhere to go. It was also the third index short in eighteen minutes. The fill came at 51,054.7, the best price was half a point better, and the 51,120 stop was hit nineteen minutes later.
What we'd do the same
The direction. A Dow short in that structure is a fair trade on a rejection that prints where the plan asked.
On a $100k account at 2.0% risk per trade.
Each trade risks +$2,000 (1R). The system's actual scale-out behavior may differ, see disclaimer.
| Scenario | R-multiple | Profit on $100k |
|---|---|---|
| Window drawdownActual | -6R | −$12,000 |
All six agents combined.
From the desk
The six losses cost 6.0R, or 12,000 dollars on the simulated 100,000 dollar account at 2% risk. Five winners on the other side, three of them after Wednesday, brought the week back to -1.24R, about 2,480 dollars down. The week's own equity curve bottomed at 8.52% below its starting balance on Wednesday.
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. On a static 100,000 dollar account at 2% risk that is $176,703.28. Compounding the same trades, where each position is sized off the balance it actually had, gives $203,855.48. The gap between the two, a little over 27,000 dollars, is not extra edge. It is the same R banked through fixed fractional sizing. This week's 12,000 dollars of losing trades is under 7% of the static figure, and the net 2,480 dollar give-back is under 2%, because no single position was ever allowed to be bigger than 1R.
One note on scope. October is still open, so this week is not yet in the year to date figure. All six losses fell between Oct 5 and Oct 7, and they join the year to date number when the month closes.
What we're tuning
For two weeks we have written that we are looking at making the candle-close conditions in a session plan binding at the approval step, and last week we added requiring an approval that follows a decline within a couple of minutes, with no new closed candle in between, to say what changed. This week added two more of the second kind inside the same twenty minutes, and three losses whose approvals said the trigger was incomplete. The case for both is stronger than it was a week ago.
The new question is correlated exposure across the index books. The Nasdaq, S&P and Dow plans are written separately and each reads its own tape, but on Wednesday they took the same short inside eighteen minutes and lost it together. A check that notices when a third index book is about to enter the same direction within minutes of the other two belongs before the entry, not after the stops. None of these changes ships on one week, but all three now have evidence behind them.
What the numbers actually mean
Six losses in eleven trades, three of them inside eighteen minutes, feels like a system that stopped working. The arithmetic says it is an ordinary week with an ugly middle.
Across 257 trades from Jan 12 inception the system has won 57.59% of the time and banked +38.35R. At that win rate roughly 42% of trades lose, so the chance that any three particular trades in a row all lose is a little under 8%. Over a few hundred trades a year, a three-loss run is something to expect many times over, and this week's longest streak was exactly three. The same arithmetic also says why Wednesday deserves extra attention: three correlated index shorts behave less like three independent draws and more like one larger one, and correlated losses arrive together far more often than independent ones would.
Van Tharp built the R-multiple around the idea that a trading system is a distribution of outcomes measured in units of risk, not a win rate. The average winner in this window's panel is 0.95R, drawn from the week's five winners at their first targets. With a 1R loser, a 0.95R average winner needs a win rate a little above half to break even. That is why the 57.59% record matters more than this week's 45.5%, and why every loss staying at exactly 1R matters as much as any single winner.
Jack Schwager's Market Wizards interviews make the same point from the other side. The traders who last are not the ones who avoid losing runs. They are the ones whose losing runs stay uniform and bounded. Six losses at 1R each, and an 8.52% peak to trough on the simulated account, is that kind of run, even if it is a deeper one than the last two weeks.
One caveat on sample size. A single week is eleven trades, and the drawdown figure in the panel comes from this window's equity walk alone. The 257-trade record and the 57.59% win rate are the numbers with weight behind them. This week is one slice of that record.
At a Glance
Drawdown questions
Six losses in three days. Did something break?
No. The setups, the risk model and the sizing were the same as the week before, and every loss was exactly 1R, the designed size. What the losses share is timing: five of the six were approved with the approval itself flagging the entry as early or marginal. Thursday and Friday, when the entries waited for closed candles, produced three winners and no losses.
Why were three index shorts open at the same time on Wednesday?
Each index book writes its own plan and reads its own tape, and between 15:19 and 15:37 UTC on Wednesday all three tapes read bearish, with breadth at -1,684 and the FOMC minutes ahead. That made it one market view expressed three times. The same oversold bounce stopped all three, and correlated exposure across the index books is one of the things we are now looking at.
The euro short did everything right. Why is it in a losses report?
Because this report covers every loss, not only the avoidable ones. The euro short met its trigger in full, had a strong macro read behind it, and still stopped. That is what a 1R cost of trading looks like, and showing it next to the early entries is the point: a clean loss and an avoidable loss cost the same 1R, but only one of them tells us to change something.
Why does this report say minus 6R when the weekly recap says -1.24R?
They measure different things. This report covers only the losing trades, so it totals 6.0R across six losses. The recap covers all eleven trades in the window and nets those losses against five winners. Neither number is an adjusted version of the other.
Why is this week not in the year to date figure?
Because October is still open. The year to date ledger is sealed at the end of each completed month, so it reads +38.35R across 257 trades through the September close. This week's trades join that figure when October closes. We publish it this way so a number printed in one month still means the same thing later in the year.
Trade with the system that publishes its drawdowns.
Subscribers receive every signal, winners and losers alike, three minutes before entry, with full reasoning.
Dollar figures are simulated on a $100,000 account at 2% risk per trade. Drawdown trajectories shown reflect a small window sample size and are not projections of forward performance. Past performance, including losses, is not a guarantee of future results. Actual subscriber P&L varies with account size and execution. YTD context: +38.35R YTD across 257 trades, see stats strip.
Past performance is not indicative of future results
The trades in this article are taken by the SkyAnalyst agents in live demo accounts at a regulated broker, in a live market environment, so the demo account reflects every transaction SkyAnalyst makes. The app logs each trade and confirms it against the live price feed of the broker; the capital is simulated and no real capital is at risk. The demo account is linked to MyFxbook, which publishes its results publicly. We publish these results to study how the agents trade and reason, for education and trade analysis, not as a recommendation. Trade at your own risk.
From the SkyAnalyst Journal
Down 4.26R by Wednesday, Then Three Winners That Waited for the Candle
Eleven trades, -1.24R. Three Wednesday index shorts stopped together. From Thursday the desk entered only on closed trigger candles and won three straight.
Buy the Retrace, Not the Release: SkyAnalyst's Dow Long After the Data
Six minutes after the 10:00 ET data, the plan passed on buying strength and asked for a 50-61.8% retrace of the Dow's 149.5-point NY impulse. The broker booked +0.90R (TP1); the move ran to +2.18R (TP3).
Right Direction, Wrong Price: The Cable Short SkyAnalyst Refused to Chase
Two reads at 88 and 85 percent said no: price had already left the 1.32130-1.32155 sell zone. The short filled back inside it at 1.32134, and TP1 printed overnight for +1.02R (TP1).