Republished: Feb 23 to Mar 1 Drawdown Report, 3 Losses for -3R, Streak of 1
Three losses, all at minus 1R. Net minus 3R for the loss-side ledger. Longest streak of 1. Original printed 4 losses and a streak of 2; the cancelled-trade fix
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.
Three losses. All at minus 1R. No back-to-back stops; longest streak ran one trade deep. Net for the loss-counting window: minus 3R. The trough sat at $98,580.44 Thursday afternoon, a 5.59 percent intraweek drawdown from the eventual Friday peak of $104,417.08. Through Mar 2, 2026, the cumulative ledger reads +7.43R YTD across 24 trades from Jan 12 inception. The simulated $100,000 account at 2 percent risk per trade sits at $114,860 on the static line — the running total has absorbed this window and stayed well above the starting balance. This is a corrected republish. The original drawdown report printed four losses and a longest streak of two. One of those four rows was a paused-mid-trade NAS100 entry from Feb 26 the broker never filled to completion, and the admin trading dashboard never counted. The cancelled-trade exclusion fix in the data layer dropped it. Corrected loss-side ledger: three losses, minus 3R, longest streak 1. We republish into the same canonical slot; slug, URL, and publish date are unchanged. This drawdown report is not the recap. The companion weekly recap covers all seven trades and lands at +1.21R net, with four TP3 winners carrying the week. See February's corrected monthly recap for closing-month context.
Act 1: Tuesday's US30 mean-revert short stops between two TP3 winners
A US30 mean-revert short at 15:33 UTC stopped at minus 1R as the index held resistance. The stop sat between two NAS100 TP3 winners (15:01 and 16:41), so Tuesday closed at plus 0.29R cumulative.
Act 2: Thursday's US30 buy-the-dip stops, the week trough sets
A US30 buy-the-dip long at 16:11 UTC stopped at minus 1R as support failed inside the next hour. Equity dropped to $98,580.44, the week's trough, leaving the week at minus 0.71R cumulative heading into Friday.
Act 3: Friday's US500 pullback long stops six minutes after the equity peak
Friday's US30 primary fade short ran to TP3 at 16:33 UTC, peaking equity at $104,417.08. Six minutes later a US500 pullback long at 16:39 UTC stopped at minus 1R as the regime softened intraday. Equity settled at $102,417.08. The Friday stop was the third and final loss of the week, isolated, with no follow-on.
Every trade the system took.
| Date | Time | Instrument | Dir | Model | Setup | Grade | R | $ Sim | Result | Details |
|---|---|---|---|---|---|---|---|---|---|---|
| Feb 24 | 15:33 UTC | US30 | Short | unknown | US30 SHORT (mean-revert at resistance) | C+ | -1.0R | -$2,000 | Stop hit | - |
| Feb 26 | 16:11 UTC | US30 | Long | unknown | US30 LONG (Buy-the-dip) | C+ | -1.0R | -$2,000 | Stop hit | - |
| Feb 27 | 16:39 UTC | US500 | Long | unknown | US500 LONG (pullback buy) | C+ | -1.0R | -$2,000 | 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 three losses did not share a setup. Two instruments, three distinct setups, all stopped at minus 1R: Tuesday's US30 mean-revert short, Thursday's US30 buy-the-dip long, Friday's US500 pullback long.
What the pattern is
Every confluence score landed in the actionable range, with macro, structure, and cross-asset reads positive at trigger. What was not present in any was the second-bar absorption signal that confirms a level is defending, not just being touched.
Why these failed
A confluence trade at the floor of the actionable range carries roughly a 40 percent failure rate by construction. Three independent floor-of-the-range trades resolved at the failure end across three sessions. Wrong tail, not off distribution.
What we keep doing
The same setups, at the same scores, will be taken again. Tightening the floor would have skipped Tuesday's NAS100 longs and Friday's US30 primary fade under any reasonable tuning.
Decision highlights
The Risk Agent did not engage a circuit breaker because the system does not have one. A pause after Tuesday's US30 stop would have skipped the Tuesday 16:41 NAS100 long (+0.69R credited) and the Friday US30 primary fade (+1.73R credited). A pause after Thursday would have skipped the entire Friday session.
The Trend Agent's confluence threshold stayed at the published floor through the window. The three losses and four TP3 winners were all evaluated under the same scoring rules. A system that tightens the floor under stress is a discretionary trader pretending to be a system.
The Macro Agent held risk-tolerant through the week. The regime read was right; what the read does not guarantee is per-trade confirmation. The Friday US500 stopped because the regime softened inside the trade lifecycle, not because the trigger was wrong.
Claude vs GPT: who led the week.
SkyAnalyst runs multiple foundation models in parallel across its four-agent system. When two models trade the same instrument in the same week, the results are directly comparable. This is that comparison.
Same signals, same risk framework, different foundation model.
Six instruments, six stories.
US30 took two losses for minus 2R net. Tuesday's mean-revert short and Thursday's buy-the-dip long both stopped at minus 1R inside the next hour. Distinct setups, different sessions, no streak.
All US30 this week →NAS100 took zero losses. Both NAS100 entries (Tuesday 15:01 and 16:41 UTC) ran to TP3.
All NAS100 this week →US500 took one loss for minus 1R net. Friday's pullback long at 16:39 UTC, six minutes after the US30 primary fade peaked the equity curve.
All US500 this week →Loss of the week: US500 Long · -1R
Losses worth learning from
Loss 1: US30 Short Feb 24 15:33 UTC
What the system saw: mean-revert short at a tested resistance shelf. Macro gated risk-tolerant on a lean-bullish DXY read. Cross-Asset confirmed with NAS100 holding pace. Confluence at the published floor.
What went wrong: within the first hour the index defended the shelf as a higher low and the thesis inverted. Volume on the test was at the 60-period average; no rejection candle printed.
Lesson: every macro input was inside tolerance. The inversion happened inside the trade lifecycle. We would take the trade again under the same scoring.
Loss 2: US30 Long Feb 26 16:11 UTC
What the system saw: buy-the-dip long at a tested support shelf inside a risk-tolerant regime, index trending higher on the rolling sample. Cross-Asset confirmed with NAS100 bid. Confluence at the floor.
What went wrong: the support shelf failed inside the first hour as a broader risk-off ripple ran through the equity indices. Volume-scoring of resting orders below the shelf was thin. The stop set the week trough at $98,580.44.
Lesson: the bullish read was inside the regime envelope. Variance is the cost of trading the floor. Removing the floor would lower expected value across calendar windows.
Loss 3: US500 Long Feb 27 16:39 UTC
What the system saw: pullback long six minutes after the US30 primary fade peaked equity at $104,417.08. Macro risk-tolerant. Cross-Asset cleared the index pair as decoupled. Confluence at the floor.
What went wrong: the regime softened intraday and the pullback failed to hold. Setup clean at trigger; post-entry tape did not pay.
Lesson: a discretionary trader holding the +1.73R gain would have skipped this entry. The system does not consult the recent trade record. The cost this week was one isolated minus 1R stop after the peak.
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 | -3R | −$6,000 |
All six agents combined.
From the desk
Through Mar 2, 2026, the cumulative ledger reads +7.43R YTD across 24 trades from Jan 12 inception. The same $100,000 account at 2 percent risk per trade sits at $114,860 on the static line and $115,393 on the compounded line — the spread is the cost (or benefit) of compounding through a positive-expectancy edge as winners cluster around losses.
The honest reading is that the system traded its full sizing, took every setup that cleared threshold, and gave back minus 3R on three independent trades, while the same five sessions produced four TP3 winners and closed the recap at +1.21R net. The asymmetry is the system's central edge.
The republish: the original drawdown report counted four losses and a streak-of-two on Feb 26. The fourth row was a paused-mid-trade NAS100 entry the dashboard never had. The cancelled-trade exclusion fix drops it. The companion weekly recap and February monthly recap regenerated against the same correction. Adjacent: prior drawdown report and the offsetting US30 primary fade case study.
What we're tuning
There is no entry-side tuning signal in the three losses. All three cleared the published threshold; minus 3R sits inside the rolling-record interquartile range; the streak of one is below the long-run median.
The tuning happened upstream. The cancelled-trade exclusion in the data layer is now patched, and a regression test blocks paused-mid-trade rows from reaching the drawdown aggregator. The Thursday stop surfaces the same volume-scoring concern we track on resting-order treatment, but a single instance is not a basis for change.
What the numbers actually mean
A 35-40 percent long-run win rate paired with the asymmetric tail the four TP3 winners on the recap side illustrate is the rate-and-reward profile this system was designed around. One TP3 winner at +1.73R credited covers 1.73 losers; the four TP3 winners this week (+0.60R, +0.69R, +1.19R, +1.73R credited) covered the three losers and then some at the broker fill. The corrected weekly recap closes at +1.21R net inside the same five sessions that produced this minus 3R drawdown ledger. That is what asymmetry looks like at week resolution; usually it is only visible at the rolling-100-trade horizon.
A three-loss week with a longest streak of one is well inside expected variance. Van Tharp's R-multiple framework, Schwager's analysis of trend-following systems, and standard binomial treatment of independent outcomes all conclude that a 35-40 percent system has expected longest losing streaks of 5-8 trades inside any rolling 100-trade window. This week's streak of one is far below that median. A minus 3R intraweek drawdown represents 5.59 percent of equity at the trough relative to the Friday peak; drawdowns of 5-10 percent are inside the first standard deviation of expected variance for this volatility profile.
The concept worth holding onto: judge a system on its 100-trade rolling window, not its weekly one. The shorter the window, the more variance dominates the signal. A drawdown report exists to make the variance visible. The math, extended to the right horizon, is what makes the variance pay. The corrected ledger sits cleanly inside that long-run profile.
At a Glance
Drawdown questions
Why is this article being republished, and what changed in the numbers?
A cancelled-trade exclusion fix dropped one paused-mid-trade NAS100 row from Feb 26 the admin dashboard never counted. Corrected loss-side ledger: three losses, minus 3R, streak of one. Original: four losses, minus 4R, streak-of-two.
How does this week compare to the long-run drawdown distribution?
Well inside the first standard deviation. Binomial treatment of a 35-40 percent system predicts longest streaks of 5-8 inside any rolling 100-trade window. This week's streak of one is far below that median.
How does the recap's net R relate to the drawdown report's net R?
The corrected <a href="/blog/weekly-recap-2026-02-23">weekly recap</a> projects winners at TP1 baseline and counts losses the same way, landing at +1.21R net across the seven trades. This drawdown counts only losses, reporting minus 3R. Same methodology, different slices.
Why did the system not pause after the Thursday US30 stop set the week trough?
The system has no streak-aware circuit breaker. A pause after Thursday would have skipped the Friday US500 fade and US30 primary fade to TP3, the two largest credited winners of the week. The threshold is the threshold.
Should a 5.59 percent intraweek drawdown change how I think about this system?
No. Drawdowns of 5-10 percent are routine for this risk profile, and the corrected ledger closed the five sessions at +1.21R net. The honest signal is the data-layer fix, not the drawdown itself.
Trade with the system that publishes its drawdowns.
Subscribers receive every signal — winners and losers — 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: +7.43R YTD across 24 trades, see stats strip.
From the SkyAnalyst Journal
The Year SkyAnalyst Learned to Trust Claude: 99 Trades, +16.57R YTD
Ninety-nine trades since launch on Jan 12, 2026. Plus 16.57R net at a 58.6 percent win rate. The headline isn't the number — it's how a desk that opened with three trades in January became a system holding expectancy across four months.
US500 Short on March 19: A Failed Bounce Inside the Deepest Drawdown Week
A SHORT at 6596.9 into VWAP and prior-day-low resistance, four waits and one enter at 74 percent confidence, a 3h 55m hold to TP1 for +1.18R inside the worst week of the published record.
US500 Long on March 16: A B-Grade Buy-the-Dip Inside the Deepest Drawdown Week
A LONG pullback at 6706 with two waits, one enter at 72 percent confidence into a lean-bear FOMC backdrop, a 59-minute ride to TP1 for +1.5R inside the worst weekly stretch of the published record.