Weekly Losses: the Monday afternoon that cost us four R
Five losses, four of them in thirty-eight minutes on the same Monday session. The system gave back 4.5R against +23.11R YTD, then walked the curve back to a fre
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.
Through Jun 7, 2026, the system has banked +23.11R YTD across 133 trades. A $100,000 simulated account at 2 percent risk per trade sits at $146,231.82 static, or $154,125.13 compounded, against the inception equity from Jan 12. This week we gave 4.5R of that figure back. Five losses, four of them stacked into a thirty-eight-minute window on Monday afternoon. The curve recovered to a fresh peak by Friday close, but the Monday cluster is the only thing worth writing about, so that is what this article does.
Act 1: The Monday wipeout
The first loss landed at 15:06 UTC on Monday, a GBPUSD short stopping out for 1R flat. Nineteen minutes later, a USDJPY long was stopped at half-R as a partial position. One minute after that, the GPT trader's US30 short stopped for another full R. Eighteen minutes later, the GPT trader's EURUSD short stopped again. Four entries, four losses, thirty-eight minutes. The simulated curve walked from $100,000 to $93,000 in roughly the time it takes to drink a coffee.
Act 2: The Wednesday aftershock
The system stood down Tuesday entirely. Wednesday afternoon, Claude flagged a second-chance GBPUSD short at 14:33 UTC. It stopped at flat 1R, which is the second-worst kind of loss: the same instrument, the same direction, the same outcome as Monday's first trade. Drawdown deepened to 7.48 percent on the simulated equity walk. That was the trough.
Act 3: The walk-back
Wednesday's late afternoon and Thursday produced a string of unpublished winners that we have written about in this week's recap. By Friday at 14:36 UTC, the simulated equity printed a fresh peak. The week closed green on the recap side of the books. It still cost us 4.5R, a 7.48 percent drawdown, and the lesson written into the Pattern of the Week section below.
Every trade the system took.
| Date | Time | Instrument | Dir | Model | Setup | Grade | R | $ Sim | Result | Details |
|---|---|---|---|---|---|---|---|---|---|---|
| Jun 1 | 15:06 UTC | GBPUSD | Short | Claude Opus 4.7 | GBPUSD Post-ISM Pullback Short into VWAP/Fibonacci Resistance | C+ | -1.0R(SL) | -$2,000(SL) | Stop hit | - |
| Jun 1 | 15:25 UTC | USDJPY | Long | Claude Opus 4.7 | USDJPY Pullback Long to VWAP/Support Continuation | C+ | -0.50R(SL) | -$1,000(SL) | Stop hit | - |
| Jun 1 | 15:26 UTC | US30 | Short | GPT-5.5 | Pullback Short into Underside Resistance | B | -1.0R(SL) | -$2,000(SL) | Stop hit | - |
| Jun 1 | 15:44 UTC | EURUSD | Short | GPT-5.5 | Short the rebound into VWAP / Fib resistance | C+ | -1.0R(SL) | -$2,000(SL) | Stop hit | - |
| Jun 3 | 14:33 UTC | GBPUSD | Short | Claude Opus 4.7 | Post-Data Second-Chance Short | B | -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 this week is correlation, and the system did not see it in time.
Four of the five losses landed inside thirty-eight minutes on Monday afternoon. The setup agents flagged each pair independently: a GBPUSD pullback short into VWAP, a USDJPY continuation long, a US30 underside-resistance short, a EURUSD VWAP-fade short. Read individually, every setup graded at C+ or B. The Setup Agent does not currently weigh how many open positions already lean the same way before sizing a new one, and the four shorts all leaned the same way: short dollar bloc strength against the same intraday shelf that was about to hold.
When the dollar caught a bid in the 15:00 to 15:45 UTC window, every short stopped together. That is not five independent losses. That is one correlated event the system priced as five independent entries. The teardowns below name two of the worst examples; the fix lives in the What we tune next section.
Decision highlights
The Setup Agent sized the EURUSD short at 15:44 UTC Monday without checking that two correlated dollar-bloc shorts had already opened in the previous twenty minutes. The setup was a clean read on its own merits; the portfolio context was not part of the gate. This is the single most expensive judgment call of the week and the one that lands at the top of the tuning list.
The Risk Agent held flat 1R sizing through every entry of the Monday cluster, refusing to escalate after the first loss and refusing to reduce after the second. That discipline is what kept the give-back at 4.5R instead of the 8R or 10R that revenge sizing would have produced. The decision to do nothing is the decision the most experienced traders make and the one untested systems usually fail.
On Wednesday at 14:33 UTC, the Setup Agent re-entered GBPUSD short into the same direction that had stopped Monday. The read was independent on its own evaluation, but the portfolio memory of two days earlier was not weighted into the confluence score. We do not currently penalize a same-instrument, same-direction re-entry inside a 72-hour window. We will, after this article.
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.
- GBPUSD-2.0R · 2 trades
- USDJPY-0.5R · 1 trade
- EURUSD-1.0R · 1 trade
- US30-1.0R · 1 trade
Same signals, same risk framework, different foundation model.
Six instruments, six stories.
EURUSD: One loss this window. The GPT-traded short at 15:44 UTC Monday stopped for a flat 1R as the dollar bid that broke the entire afternoon's cluster carried EURUSD back through the VWAP fade level.
All EURUSD this week →GBPUSD: Two losses, both Claude. The Monday entry at 15:06 UTC stopped for 1R inside the wipeout window. The Wednesday second-chance entry at 14:33 UTC stopped again for 1R, taking the simulated curve to its 7.48 percent trough before the recovery began.
All GBPUSD this week →US30: One loss. GPT's underside-resistance short at 15:26 UTC Monday stopped for 1R when the index lifted with the broader risk-on rotation. Grade was B on the read; the loss came on the cluster, not the setup.
All US30 this week →NAS100: No losses this window. The Setup Agent flagged two NAS100 long candidates Thursday morning, both filled on the winning side and reported in the weekly recap.
All NAS100 this week →USDJPY: One half-R loss. Claude's pullback long at 15:25 UTC Monday took a partial stop as the dollar pop reversed the continuation read. This is the only scratch in the window and counts at half-R against the scoreboard.
All USDJPY this week →US500: No losses this window. The pair flagged once on a short setup that filled later in the week on the green side of the curve.
All US500 this week →Loss of the week: GBPUSD Short · -1R
Losses worth learning from
what was right
The grade B read on this US30 short at 15:26 UTC Monday. Underside resistance on a session that had pushed into a clean lower-high, VWAP rejected on the prior touch, volume on the rejection candle above the 60-period average. The Setup Agent had every confluence factor it needed to size in.
what was wrong
The entry came nineteen minutes after the GBPUSD short and one minute after the USDJPY partial stop. The dollar tape was already inflecting. The portfolio layer did not flag that three of the four open or just-closed positions all leaned short the dollar bloc. The US30 trade itself was a clean read; it was the timing inside the cluster that ate it.
what we'd do same
The setup read. The grade. The flat 1R sizing.
what we'd do different
A portfolio correlation gate that checks how many same-direction, same-asset-class trades are open or recently stopped before sizing the next one. This trade gets vetoed at the gate.
what was right
The Wednesday GBPUSD short at 14:33 UTC was a textbook second-chance entry on the same read that had stopped Monday. The Setup Agent does not currently penalize a re-entry on the same instrument, same direction, inside 72 hours, and the on-its-own-merits read was a grade B. The volume confirmation was there. The structure was there.
what was wrong
The portfolio memory was not. Monday's GBPUSD short had stopped on the same direction. The probability of the second-chance read paying off does drop measurably when the first read stopped two days earlier on the same instrument; the Setup Agent did not weight that signal because the gate does not currently check for it.
what we'd do same
The structural read. The grade. The flat sizing.
what we'd do different
A look-back window in the Setup Agent that downgrades confluence on a same-instrument, same-direction re-entry inside 72 hours. Probably by half a grade. The fix is in the next section.
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 | -4.5R | −$9,000 |
All six agents combined.
From the desk
A $100,000 simulated account at 2 percent risk per trade enters this week at $150,731.82 static, or $158,625.13 compounded, against an inception equity from Jan 12. It exits the week at $146,231.82 static or $154,125.13 compounded. This week's give-back is approximately $4,500 of the static figure, or roughly the same magnitude on the compounded side. The static path and the compounded path tell a slightly different story week-over-week, and the gap between the two ($7,893 in the system's favor on the compounded side) is the receipt for disciplined sizing across 133 trades.
We almost did not publish this article. The week closed green on the recap side of the books. By the time the Friday session ended, the curve was at a fresh peak. The natural editorial instinct is to write only that story and not this one. We are publishing this one because the Monday cluster is research data we owe the people who read this site, and because a system that only writes about its winning weeks is not a system, it is a marketing channel.
The four agents had a structurally bad afternoon. They did the parts of the job we ask of them, well: each Setup Agent flagged a clean independent read, the Risk Agent held the line on sizing, the Macro Agent had the dollar regime tagged correctly going into the session. The part of the job we have not asked them to do, yet, is the portfolio-level correlation check that would have stopped the EURUSD short from going on at the moment three other dollar-bloc shorts were already open. We are asking them to do it now.
What we're tuning
Two specific changes ship in this week's iteration cycle. First, a portfolio correlation gate in the Risk Agent that checks how many same-direction, same-asset-class positions are open or were stopped inside the prior thirty minutes before sizing a new entry. The Monday cluster would have been three losses instead of four; the fourth, EURUSD at 15:44 UTC, would have been vetoed at the gate.
Second, a 72-hour look-back in the Setup Agent that downgrades confluence by half a grade on a same-instrument, same-direction re-entry where the prior entry stopped. Wednesday's GBPUSD second-chance short would have re-evaluated at C+ instead of B and likely declined entry. Both changes are mechanical, both are testable on the historical record, and both will publish as their own analysis the week the back-test completes.
What the numbers actually mean
A win rate of 34.1 percent across the most recent 91 trades sounds low to anyone who has not read Van Tharp's `Trade Your Way to Financial Freedom`. Tharp's central point in that book is that a trading system is not graded on win rate; it is graded on expectancy, which is win rate times average win minus loss rate times average loss. A system that wins 34 percent at +3R average and loses 66 percent at -1R average has a positive expectancy of +0.36R per trade. That is the math the system is actually running on.
This week we drew a -4.5R cluster, with a longest losing streak of 4. Jack Schwager catalogs in `Market Wizards` the streak distributions that come out of systems with win rates in the 35 to 55 percent range; a four-loss streak is well inside the expected range for a system in our window, and a six-loss streak shows up roughly every three months on the historical math. The current 7.48 percent peak-to-trough drawdown is uncomfortable to look at, but it is not anomalous. We publish drawdown reports because every legitimate trading fund publishes them; the alternative is showing readers only the win-of-the-week and hoping they do not check the curve.
The Kelly criterion gives a conceptual answer to the question of how much to risk per trade. We do not publish a computed Kelly fraction because the inputs (true win rate, true average win, true average loss) drift on a system this young. Fixed 2 percent risk per trade is well below any reasonable Kelly half-fraction we could compute from current data, which means the system is structurally underbet, which means drawdowns of this size are absorbed and not amplified. The 91-trade sample size behind these numbers is still small. The next quarter of execution data will tighten the confidence interval; the conclusions will probably not change.
At a Glance
Drawdown questions
Is a 7.48 percent drawdown normal for a system with this kind of win rate?
It is well inside the expected range. A system that wins around 34 percent of its trades and runs at flat 2 percent risk per trade should expect drawdowns of 5 to 12 percent at multiple points across any given quarter. The distribution of streak lengths in Schwager's research suggests four-loss streaks are normal and six-loss streaks show up roughly every three months. The number to watch is not any single drawdown, it is the slope of the equity curve across hundreds of trades.
Why did the system not stop trading after the first or second loss on Monday?
Because the Risk Agent does not run on emotion. Each setup is evaluated on its own confluence math; an open or recently closed losing position does not currently invalidate a new independent setup. After this week, that will change for same-asset-class positions inside the same session. The change is in the Pattern of the Week section above.
How does the compounded path differ from the static path week to week?
Static dollars assume the account size never changes; every trade risks 2 percent of the original 100,000. Compounded dollars grow or shrink the per-trade dollar risk with the account, so a winning streak compounds faster and a losing streak loses faster. Across this YTD, compounded sits 7,893 above static, which is the gap that disciplined sizing buys you on the upside.
Are losing weeks like this one a sign the model is degrading?
Not on this sample size. Ninety-one trades is enough to compute the headline statistics but not enough to detect a slow drift in win rate or average R. The system runs continuous monitoring on both; the alert thresholds are set at multiples of historical standard deviation, and nothing in this week's data tripped them. We will publish a degradation analysis the quarter we see one.
Why publish the losses at all?
Because every legitimate trading fund publishes its drawdown reports. The alternative is showing readers only the win-of-the-week and treating the losses as marketing inconvenience. We are not running that playbook. The losses are research data. They are also a more accurate picture of what the system feels like to trade with than the wins.
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: +23.11R YTD across 133 trades, see stats strip.
From the SkyAnalyst Journal
Jun 1-7, 2026: a Monday wipeout, a Friday recovery
Twelve trades. Four losses inside thirty-eight minutes on Monday. One bounce-rejection short on NAS100 that paid the week. A 58.3% win rate that owes everything to discipline on Friday afternoon.
The NAS100 short the macro stack pre-cleared in one read
Yields at five-day highs, DXY firm, VIX rising, all six confluence factors clean on the first scan. A single evaluation took NAS100 short at 29,876.3, and the move ran 386 points.
US500 short: six of six confluences, four reads inside four minutes
NFP had cratered breadth and lit VIX. Six of six factors said short before the first evaluation ran. The interesting question wasn't whether. It was where.