Six trades, three winners, three losers, -1.01R net at the TP1 baseline. Tuesday opened with a 3-for-3 morning printing +1.99R inside thirty minutes; Wednesday
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.
Six trades, three winners, three losers, -1.01R net on the TP1 baseline. That is the scorecard for May 25-31, 2026 across the canonical-instrument set, a thin week with Memorial Day taking Monday and no triggers Friday. Cumulative equity opened at $100,000, climbed to $103,970.45 by 14:32 UTC Tuesday after three back-to-back winners ran TP1 or better, then gave back through two simultaneous Wednesday stops and a Thursday US30 short to close the window at $97,970.45. Through May 31, 2026, the system has banked +20.00R YTD across 121 trades since the Jan 12 inception. A $100,000 simulated account at 2% risk per trade sits at $139,996 (static) or $145,328 (compounded) — a modest give-back week against a year still net positive in the high teens of R-multiples.
Tue May 26 carried three trades, all triggered between 14:05 and 14:32 UTC, all cleared TP1 or better. At 14:05 UTC a Claude Opus 4.7 NAS100 long on a pullback-to-5m-dynamic-support pattern ran TP1 for +0.78R at C+ grade. Two minutes later, at 14:07 UTC, a GPT-5.5 GBPUSD short ran TP3 for +0.63R at C+ grade. At 14:32 UTC a GPT-5.5 USDJPY long ran TP3 for +0.57R, also at C+ grade. The three winners summed to +1.99R inside thirty minutes, taking the cumulative equity from $100,000 at the open to $103,970.45 by the end of the morning. Tuesday closed at +1.99R net on a three-winner, zero-loser session.
Wed May 27 produced two trades at the same 14:42 UTC trigger window, both losses. A Claude Opus 4.7 US30 long on a pullback continuation stopped for -1R at C+ grade. A GPT-5.5 EURUSD long on a pullback-to-VWAP setup stopped for -1R at B grade. Two simultaneous full stops took cumulative equity from $103,970.45 back to $99,970.45, erasing Tuesday's +1.99R print and crossing flat by 14:42 UTC. Wednesday closed at -0.01R net, the architecture two basis points below the open after a four-trade swing.
Thu May 28 carried a single trade. At 14:36 UTC a GPT-5.5 US30 short on a pullback failure stopped for -1R at C+ grade. Cumulative equity moved from $99,970.45 to $97,970.45, closing the window at -1.01R net. Friday May 29 produced no triggers across either model family on any of the six canonical instruments. The week's full arc ran inside three trading sessions on three consecutive days, six trades, three winners, three losers, and a -1.01R close.
| Date | Time | Instrument | Dir | Model | Setup | Grade | R | $ Sim | Result | Details |
|---|---|---|---|---|---|---|---|---|---|---|
| May 26 | 14:05 UTC | NAS100 | Long | Claude Opus 4.7 | Breakout Retest Long | C+ | +0.78R(TP1) | +$1,563(TP1) | TP1 hit · ★ Trade of the week | Read case → |
| May 26 | 14:07 UTC | GBPUSD | Short | GPT-5.5 | GBPUSD Post-Data Pullback / Failed Retest Short | C+ | +0.63R(TP1) | +$1,268(TP1) | TP3 hit | Read case → |
| May 26 | 14:32 UTC | USDJPY | Long | GPT-5.5 | USDJPY Conditional Pullback Long | C+ | +0.57R(TP1) | +$1,140(TP1) | TP3 hit | Read case → |
| May 27 | 14:42 UTC | US30 | Long | Claude Opus 4.7 | US30 Long Pullback to Broken Resistance | C+ | -1.0R(SL) | -$2,000(SL) | Stop hit | - |
| May 27 | 14:42 UTC | EURUSD | Long | GPT-5.5 | EURUSD Long Pullback Continuation | B | -1.0R(SL) | -$2,000(SL) | Stop hit | - |
| May 28 | 14:36 UTC | US30 | Short | GPT-5.5 | Short Failed Reclaim of OR High / VWAP | 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.
The week's recurring pattern was a clean Tuesday-morning continuation that printed three for three in thirty minutes, then two-against-two chop on US30 across Wednesday-Thursday that gave back twice what Tuesday banked on a single instrument. Tuesday's three winners triggered on retest continuations into NY AM session opens with the Macro Agent quiet on regime risk. Wednesday-Thursday's three losses ran on US30 twice and EURUSD once, with the US30 column carrying the full -2R drag of the week.
Inside the six canonical trades, NAS100, GBPUSD, and USDJPY each banked one winner for a combined +1.99R, US30 took two losses for -2.0R as the heaviest instrument drag, EURUSD took one loss for -1.0R, and US500 sat the window out. Three winners contributed +1.99R combined at the TP1 baseline; three losers contributed -3.0R combined, each printing as a clean full stop with no partial structural exits. The 50 percent win rate paired against an average winner of +0.66R and an average loser of -1.0R produces a slightly negative R-multiple by expectancy math, and that is what landed.
Three winners summed to +1.99R: +0.78, +0.63, and +0.57. Three losers summed to -3.0R: three full stops at -1R apiece. The net at -1.01R is the arithmetic of a 50 percent win rate with an average winner of +0.66R and an average loser of -1R. When average winners do not clear 1R, a 50/50 split closes slightly red rather than scratch. The expectancy math is well documented; this is the published distribution running on its left half on a thin week.
The Tuesday morning decision to fire three back-to-back entries across two model families and three different instruments inside thirty minutes is the clearest discipline read of the week. Between 14:05 and 14:32 UTC the system sized into a Claude NAS100 long, a GPT GBPUSD short, and a GPT USDJPY long, each at the same 2 percent risk on equity at trigger. The Risk Agent did not down-weight any of the three after the first two cleared TP, and the third entry ran TP3 to complete a +1.99R morning. Each entry triggered on its own confluence card.
The Wednesday decision to fire the Claude US30 long and the GPT EURUSD long simultaneously at 14:42 UTC is the per-trade isolation read of the window. Two different model families opened two different instruments at the same trigger, each sized at 2 percent risk on its respective equity-at-trigger, and the architecture did not collapse them into a correlated position. The two stops printed independently and accounted for -2R of the week's -3R loss column. The decision logic at trigger time treated them as independent confluence cards, which is the design.
The Thursday decision to take the lone US30 short at 14:36 UTC after two US30-instrument-family losses already printed in the column is the discipline read of the week's close. The entry triggered on a pullback-failure pattern at C+ grade with the Macro Agent not flagging regime risk. The Risk Agent ran the standard 2 percent sizing on equity-at-trigger and did not adjust on the basis of the prior US30 stop. The trade stopped at -1R, closing the window at -1.01R net. The fixed-R policy and the confluence floor ran unchanged through the sequence.
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.
EURUSD took one trade for a 0 percent win rate and -1.0R net. The May 27 GPT long on a pullback-to-VWAP setup at B grade stopped for -1R at the same 14:42 UTC trigger as the Claude US30 long stop. The pair did not produce a triggered green entry inside the window.
All EURUSD this week →GBPUSD took one trade for a 100 percent win rate and +0.63R net. The May 26 GPT short ran TP3 at C+ grade as part of the Tuesday morning three-print cluster.
All GBPUSD this week →US30 took two trades for a 0 percent win rate and -2.0R net, the heaviest single-instrument drag of the window. The May 27 Claude long on a pullback continuation stopped for -1R at C+ grade. The May 28 GPT short on a pullback failure stopped for -1R at C+ grade. Both trades cleared the confluence threshold at trigger; the tape did not produce the runner.
All US30 this week →NAS100 took one trade for a 100 percent win rate and +0.78R net. The May 26 Claude long on a pullback-to-5m-dynamic-support pattern ran TP1 at C+ grade, the largest single contribution of the window.
All NAS100 this week →USDJPY took one trade for a 100 percent win rate and +0.57R net. The May 26 GPT long ran TP3 at C+ grade, completing the Tuesday morning three-print cluster.
All USDJPY this week →US500: no trades this week — the index sat outside the architecture's setup criteria across the available trigger windows on Tuesday through Thursday.
All US500 this week →Win of the week: NAS100 Long · +0.78R
Each of the three losers cleared the published confluence threshold at trigger. The May 27 Claude US30 long triggered on a pullback continuation pattern with confluence signals that scored above threshold; the May 27 GPT EURUSD long entered on a pullback-to-VWAP setup at B grade, the highest grade on any trade of the week; the May 28 GPT US30 short ran a pullback failure pattern with the Macro Agent quiet on regime risk. Every grade described an entry card that scored above the published floor, the Macro Agent did not veto regime on any of the three, and the Risk Agent ran fixed-R sizing unchanged through each entry.
The US30 column carried the heaviest weight, two trades on opposite directions across consecutive days for -2.0R net, and the pattern inside that column matters more than either single stop. Wednesday's long and Thursday's short both triggered on pullback continuations or failures that did not produce the second leg inside the runner window. The architecture read the entry card cleanly on each, the tape simply did not extend. The EURUSD long stop is the cleaner loss to read; the entry took a session pullback into VWAP and the tape continued through the level rather than reverting. The recap surfaces the pattern — US30 chop across two sessions, EURUSD a single rejection — not a trade-by-trade autopsy of three stops that each cleared the published floor at trigger.
Each trade risks +$2,000 (1R). The system's actual scale-out behavior may differ, see disclaimer.
| Scenario | R-multiple | Profit on $100k |
|---|---|---|
| Window netActual | -1.01R | −$2,020 |
The honest reading is that this was a thin week that printed exactly what a 50/50 win rate produces when the average winner does not clear 1R. Six trades, three winners, three losers, -1.01R net. Tuesday's clean three-print morning made the case that the architecture's NY AM continuation entries still find the tape; Wednesday-Thursday's two US30 stops and one EURUSD stop made the case that the confluence floor does not guarantee the runner on any single entry.
Through May 31, 2026, the system has banked +20.00R YTD across 121 trades since the Jan 12 inception. A $100,000 simulated account at 2% risk per trade sits at $139,996 (static) or $145,328 (compounded). The difference between the two — roughly $5,300 — is the compounding effect of disciplined fixed-fraction sizing across 121 trades: every winner reinvests at a slightly larger base, every loser reduces it. The static figure is the simpler read across periods; the compounded figure is what an operator actually banks if they leave the account size to drift with the curve.
The architecture point on this window is that the sizing did not flinch on either side. Tuesday's three sequential entries were sized at the same 2 percent risk on equity at trigger that Wednesday's two simultaneous opens used and that Thursday's lone US30 short ran. A discretionary trader sitting on a +1.99R Tuesday morning might have stepped aside Wednesday after the first stop printed; the system did not. The fixed-R policy and the published confluence threshold ran unchanged through the six-trade sequence across both model families and across the index-to-FX instrument boundary. From the SkyAnalyst Team.
The US30 column is the one signal worth flagging out of this window. Two trades on opposite directions stopping for -2R across consecutive days is the kind of single-instrument print that warrants a closer review of the confluence floor on the index, not a threshold change on the basis of one week's six-trade sample. We are reviewing whether the Wednesday-Thursday continuation and pullback-failure entries on US30 were over-weighting confluence factors that decayed inside the runner window across this specific session's tape. The fixed-R policy and the 2 percent sizing rule are unchanged across both Claude and GPT through the review.
The Friday no-trade outcome is the second item on the bench, in a positive sense. The architecture did not fire on any instrument on May 29 despite the window's six-trade sample running half against. The setup-grade discipline held the line at threshold rather than reaching for a recovery print. We are not opening a new instrumentation track from this week; a -1.01R close on six trades with the Tuesday three-print morning intact is the published expectancy running on a thin sample, not a regime drift signal.
Three winners summed to +1.99R at the TP1 baseline: +0.78R, +0.63R, and +0.57R. Three losers cost -3.0R: three full stops at -1R apiece. The net arithmetic settled at -1.01R across six trades. A 50 percent win rate with an average winner of +0.66R and an average loser of -1R produces a slightly negative result inside the published expectancy band. Average winners that do not clear 1R require a higher-than-50-percent hit rate to print flat.
Claude ran two trades for -0.22R net at 50 percent win rate; GPT ran four trades for -0.80R net at 50 percent win rate. The win rate on both sides tied at 50 percent. Claude's single winner was the May 26 NAS100 long at +0.78R; GPT's two winners were the May 26 GBPUSD short at +0.63R and the USDJPY long at +0.57R. Six trades across two model families is too small a sample to read anything other than tape variance through a single window.
The Wednesday Claude long on a pullback continuation and the Thursday GPT short on a pullback failure both triggered on confluence cards that scored above the published threshold at entry. Neither trade's second leg formed inside the runner window. The architecture reads each entry independently on the confluence math at trigger and does not down-weight the second US30 entry because of an adjacent loss on the same instrument family. The fixed-R policy and the 2 percent sizing rule ran unchanged across both stops.
The recap projects every winner using a TP1 exit on the simulated $100,000 account. This is the simplest baseline for comparing across periods. The May 26 GBPUSD short ran TP3 and the May 26 USDJPY long also ran TP3, so subscribers running their own scale-out or trail strategy banked closer to the headline higher-TP figures on those entries than the TP1-baseline +0.63R and +0.57R figures here. The full -1R stops print identically across baselines.
No. A six-trade week with three winners and three losers landing at -1.01R is well inside the published expectancy band. The May 11-17 window banked +5.96R on a stronger seven-three split; the May 18-24 window printed -2.82R on a tougher seven-eleven sample; this week landed in the middle on half the trade count. Year-to-date the system sits at +20.00R across 121 trades, win rate 57.9 percent. The next window's entries will trigger on the same confluence floor, the same fixed-R policy, and the same 2 percent sizing rule that ran this week.
Subscribers receive the same pre-trade AI analysis three minutes before entry.
We project the recap totals using a TP1 exit on every winning trade. This is the simplest baseline for comparing across periods. Traders running their own scale-out, trail, or TP2/TP3 hold strategies will see different totals. Dollar figures are simulated on a $100,000 account at 2% risk per trade. Actual subscriber P&L varies with account size and execution. Past performance is not a guarantee of future results.
121 trades since Jan 12 inception, 70 winners, 51 losers, +20.00R net at TP1 baseline. A $100,000 simulated account at 2% risk sits at $139,996 static, $145,328 compounded. A system that launched slow, ramped staged, consolidated by May.
Forty-seven trades, twenty-seven winners, twenty losers, +8.34R net at the TP1 baseline. Claude ran +6.89R on 26 trades at 61.5 percent; GPT printed +1.45R on 21 trades at 52.4 percent. The two-model lineup stabilized this month.
Three stops, three R given back, a peak-to-trough drawdown of 5.77 percent, and a longest losing streak of three across two trading sessions. What each stop taught us, and what the curve says about a week the runner never showed up.