EURUSD Short on March 18: Eleven Evaluations, Ten Waits, +1.81R to TP3
SkyAnalyst AI journal entry: EURUSD Short on Mar 18, 2026 closed +1.81R on TP3. Full workspace view, decision log, and AI reasoning, unedited.
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. That’s what makes the system auditable — and it’s what this case study will show, step by step, on a specific setup the trend agent almost passed on.
The macro that set the table, again
The macro print into NY-AM had not changed from the morning's first short. The Macro Agent was carrying regime strong_bear at 72 percent confidence into shared state, with US 10-year yields above their 5-day EMA and rising from yesterday's close. The DXY confirmation read mixed: the dollar index sat at 99.796 against a 5-day EMA of 99.895, technically below the average but pressing higher than yesterday's print. That ragged-but-bearish DXY shape was the same one the system had shorted into earlier in the session.
The VIX read elevated but below its 5-day EMA on the longer lookback, signaling risk aversion present but not strengthening. FOMC sat in the afternoon calendar, which compressed the size the Risk Agent was willing to allocate but did not block the trade. None of these readings on their own would have unlocked a fresh entry. Together, they kept the regime gate open for a second pass at the same playbook.
EURUSD itself had bounced from the morning's lows back toward VWAP at 1.1530, retracing into the 1.15230 to 1.15305 band that overlapped the 5-minute fib resistance and the broken intraday support-turned-resistance. The 60-minute fast EMA was still above slow EMA, so higher-timeframe trend had not fully flipped, but price was below fast EMA, RSI on 15-minute had faded to the mid-40s, and MACD histogram had been negative for several candles. A bearish pullback regime, not a bullish continuation. What the tape did not yet have was a confirmed rejection candle inside the band.
The setup at 15:48 UTC was a bounce-fade short into broken intraday VWAP. The same playbook the system had run two hours earlier on the same instrument. Walking through the structural requirement explains why the system sat through ten waits before firing.
What the pattern is
Price has been rolling over on the 60-minute chart and rejected at session VWAP. Somewhere inside the decline, it bounces back into a short-term resistance zone, typically the broken VWAP band or a fib-EMA confluence. A professional does not short the touch. They wait for the confirmation: a 5-minute bearish rejection candle inside the zone, RSI rolling back below 50, and a close back below the breached level on volume above the 60-period average.
How pros actually use it
Tested resistance levels reject roughly 40 to 50 percent of the time on a first touch, closer to 70 percent when the touch prints a rejection candle on meaningful volume. The tell is the closed bar. Without it the pattern is noise. With it, it is signal that the resting offers have absorbed the bid. The same statistical edge applies to the second instance of the pattern on the same instrument the same day, because the offers above VWAP regenerate as the rally retraces.
Why this graded C+ rather than B
Three things kept the grade modest. The DXY confirmation was technically below its 5-day EMA but pressing the 5-day from below, which the system reads as mixed rather than clean bearish-EURUSD. The 60-minute EMA stack was still in transition rather than fully inverted. And the FOMC calendar sat ahead, which compressed the Risk Agent's size envelope. C+ means tradeable, every floor clears, conviction not high enough for B.
Why the second instance is its own evaluation, not a re-entry
The system does not size a second position because the first one worked. Each setup is scored against the live tape on its own merits. The morning's first short had already closed at TP3 by the time this second evaluation cycle began. The Trend Agent reads the new bounce as a fresh opportunity if and only if the structural premise re-prints: bounce into broken VWAP, rejection on a closed bar, confirmation volume. Scoring the second instance from scratch is what stops a single winning trade from becoming a martingale.
How the system reads this, dynamically not dogmatically
The bounce-fade short into broken VWAP is one playbook of many. The same afternoon the Trend Agent was watching a fade-to-VWAP setup on US30 that the macro alignment did not support, a long on XAUUSD that did not clear cross-asset confirmation, and a continuation breakdown on EURUSD below 1.15040 that the system gated as lower quality than the rally-fade.
SkyAnalyst doesn't favor any single strategy. The confluence math picks the playbook each evaluation cycle. The fact that the same playbook fired twice on the same instrument on the same day is a property of the tape, not a preference of the system. On a different morning the same setup would have scored below threshold and the system would have skipped it. The math reads the tape first, then fits the pattern to what is there.
NY AM EURUSD Summary
Macro backdrop is bearish EURUSD, but not at the strongest possible level. The Macro Agent is strong bear, 72% confidence, and the Trend Agent is also bearish, 66% confidence, so agent alignment is supportive for shorts. However, the DXY filter does not fully align with the hard-rule bear condition: DXY is below its 5-day EMA (99.796 vs 99.895) even though it is up vs yesterday’s close, so it is not a clean “above EMA and rising” bearish confirmation. Secondary macro confirms are modestly bearish for EURUSD: US 10Y yields are above 5-day EMA and up vs yesterday (yield support for USD), while VIX is elevated but below its 5-day EMA, so risk aversion is present but not strengthening on the 5-day filter.
Technically, the intraday structure remains sell-the-rally / transition bearish. Trend Agent keeps price below VWAP ~1.1530 and below the 15m EMA stack, with downside levels at 1.15043 and 1.1491, and invalidation at 1.15385. On 60m, the fast EMA remains above slow EMA, so higher-timeframe trend has not fully flipped bearish, but price is below fast EMA, RSI has faded into the mid-40s, and MACD histogram has been negative for several candles, which supports a bearish pullback regime rather than bullish continuation.
Calendar gate: the major 8:30am USD data is already past and settled. There is no high-impact event within the next 30 minutes from the snapshot time, but FOMC later today means volatility risk remains elevated and argues for reduced size.
Directional Bias: Bearish
Volatility: High
Setup #1: EURUSD SHORT
- Entry: 1.15230-1.15305
- Entry Trigger: 5m rejection candle or failed reclaim at/just below VWAP, ideally after tagging 5m fib resistance / intraday VWAP band and closing back below 1.15230
- Stop Loss: 1.15410
- Targets: TP1=1.15098, TP2=1.15043, TP3=1.14910
- R-Multiples: TP1=1.0R, TP2=1.4R, TP3=2.2R
- Quality Score: 7.0/10
- Confidence: Medium-High - Macro and Trend agents align bearish, 15m trend structure is still down, and the preferred trade is a fade into VWAP/resistance rather than chasing mid-range price.
Confluences
- Macro Agent bearish, confidence >= 60
- Trend Agent bearish, confidence >= 60
- 10Y yield trend supports short bias
- 60m structure supports bearish pullback: price below fast EMA, MACD histogram negative
- 5m entry zone aligns with VWAP / fib / intraday resistance
- 15m RSI is below 50 / near 48-49, not extreme
- No high-impact event within 30 minutes
Failed / weaker confluences
- DXY 5-day trend is not a clean bearish confirmation because it is still below its 5-day EMA
- 60m EMA stack is not fully bearish; fast EMA still above slow EMA
Risks
- DXY is not fully aligned with the short thesis
- 5m momentum has stabilized and is no longer impulsively bearish
- Midday liquidity can produce chop and false breaks
- FOMC later today can suppress clean follow-through
Setup #2: EURUSD SHORT
- Entry: Breakdown below 1.15040 on a 5m close, or retest-fail of 1.15040 after breakdown
- Stop Loss: 1.15155
- Targets: TP1=1.14910, TP2=1.14857, TP3=1.14770
- R-Multiples: TP1=1.1R, TP2=1.6R, TP3=2.3R
- Quality Score: 6.0/10
- Confidence: Medium - Valid only if price re-accepts below NY support; this is a continuation setup with less favorable location than the rally-fade.
Confluences
- Macro Agent bearish, confidence >= 60
- Trend Agent bearish, confidence >= 60
- 10Y yield trend supports short bias
- 60m MACD histogram remains negative
- Entry is at a key session support / Trend Agent support 1.15043
- No high-impact event within 30 minutes
Failed / weaker confluences
- DXY 5-day trend not cleanly bearish
- 15m RSI is not strongly bearish, only slightly below neutral
- Breakdown entry is vulnerable to support sweep / reversal if 1.1504 is not accepted
Risks
- Selling support rather than resistance reduces edge
- NY low / session extreme may attract mean reversion buyers
- If price snaps back above 1.1504 after breakdown, failure odds rise quickly
Best Use
The higher-quality NY AM setup is Setup #1 only on a rally into 1.15230-1.15305. That matches the Trend Agent’s sell-the-rip under 1.1535 guidance and gives cleaner structure than chasing downside at current levels. If EURUSD does not retest resistance or does not break and hold below 1.15043, then No trade.
Decision log
First evaluation, 15:33 UTC, confidence 63 percent. Price has tagged the lower edge of the 1.15230 to 1.15305 retracement band. The structural premise is intact: Macro Agent strong_bear at 72 percent, Trend Agent bearish at 66 percent, EMA stacks rolling, RSI below 50. But the touch is too quiet and the 5-minute bar in progress has not closed with a rejection body. I want a closed-bar rejection inside the zone before the trigger clears. Declining.
Second evaluation, 15:35 UTC, confidence 62 percent. The bar that printed was an indecision body with wicks both sides, neither rejection nor reclaim. Volume sat below the 60-period average. The premise has not weakened, it simply has not strengthened either. Declining.
Third evaluation, 15:36 UTC, confidence 52 percent. Confidence has slipped below the 60 percent floor as the bar in progress wicked up into the upper half of the band. Below threshold means the gate is shut, full stop. Declining.
Fourth evaluation, 15:37 UTC, confidence 63 percent. Confidence has recovered as the next bar opened lower and traded back toward the band midpoint, but the closed-bar rejection candle has still not printed. A barely-passing read on a setup that has not yet produced its trigger is the configuration the system was designed to wait through. Declining.
Fifth evaluation, 15:38 UTC, confidence 68 percent. The current bar is forming as a small bearish body inside the zone, but it is not yet closed. Cross-Asset returned neutral, DXY firmed marginally, the 15-minute MACD histogram extended negative. Premise tightening, trigger not in. Declining.
Sixth evaluation, 15:40 UTC, confidence 74 percent. Highest score of the cycle. The most recent 5-minute closed bearish, but volume came in at the 60-period average rather than meaningfully above it. The system requires confirmation volume on the trigger bar, not just adequate volume. Declining.
Seventh evaluation, 15:41 UTC, confidence 66 percent. Follow-through bar opened back above the prior bar's midpoint, partially filling the rejection. The score retreated as the structure of the setup softened. The closed-bar rejection from the prior cycle did not get confirmation from the next print. Declining.
Eighth evaluation, 15:43 UTC, confidence 58 percent. Confidence dropped below the floor as price probed back toward 1.15280. Below threshold, no entry, no exception. Declining.
Ninth evaluation, 15:45 UTC, confidence 63 percent. Score recovered to the floor as the bounce stalled and price rotated back below 1.15240. The bar in progress is forming as a bearish reversal candle, but it has not closed. The rule is to act on confirmed signals, not probable ones. Declining.
Tenth evaluation, 15:46 UTC, confidence 58 percent. The bar that closed printed as a small bearish body but volume sat just under the 60-period average. The threshold cleared on the score but the confirmation volume did not. Declining.
Eleventh evaluation, 15:48 UTC, confidence 62 percent. The most recent 5-minute closed as a bearish rejection candle inside the band, the next bar opened below 1.15240 and held, the 15-minute MACD histogram ticked further negative, and volume came in above the 60-period average. This is the trigger I have been watching for across the prior ten evaluations. Entering short at 1.15232, stop 1.1541, TP1 1.15098, TP2 1.15043, TP3 1.1491.
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 |
|---|---|---|
| Stop hit (invalidated) | -1R | −$2,000 |
| TP1 hit | +0.75R | +$1,500 |
| TP2 hit | +1.06R | +$2,120 |
| TP3 hit (max potential)Actual | +1.81R | +$3,620 |
All six agents combined.
What this trade teaches
The sixth evaluation is the one to focus on. Confidence printed at 74 percent, the highest of the cycle, but the system declined because the trigger bar's volume came in at the 60-period average rather than meaningfully above it. A discretionary trader watching the score peak at 74 percent would have shorted on that bar. They would have been short at a worse price than the system entered eight minutes later, then sat through the 15:41 to 15:46 sequence as the bounce probed back toward 1.15280 and likely been picked off on the wick at the eighth evaluation when score dropped to 58 percent.
The trade closed at +1.81R (TP3) over two hours and fifty-five minutes, with zero drawdown from entry to exit. That outcome is not the system identifying a hidden edge in the C+ grade. The grade describes the setup card at entry; it says nothing about what the tape will do over the next three hours. Above the threshold floor, the variance of the tape determines the result.
The system fired on the eleventh evaluation at 62 percent confidence, three points below the score that had peaked at 74 on the sixth. The number was never the trigger. The closed bar with confirmation volume was. - From the post-trade review
The shape pairs cleanly with the morning's first short, documented in the March 18 EURUSD short at +1.07R. Same instrument, same bounce-fade playbook, same closed-bar trigger requirement, same TP3 outcome. The MTD book stood at 19 trades and +3.58R net after this trade closed, with the 31.6 percent win rate carrying outsized average winner contribution from the two paired EURUSD shorts.
From the desk
This trade did not look special on the setup card. A C+ grade. A 62 percent confluence score on the entry evaluation, having traveled through a 74 percent peak and a 52 percent trough across the prior ten cycles. None of those numbers, on their own, would have any reader marking this as the larger of two paired TP3 winners on the same downward tape.
What separated it from the bounces that stopped earlier in the quarter was the tape, and the tape is not something the system claims to predict. We do not say "this will run 32 pips." We say "this clears every floor, the bias is intact across timeframes, the macro is supportive, the closed-bar rejection has finally printed with confirmation volume." The system places the stop above structural invalidation, sets targets at the next three references, and lets the position run.
On March 18 at 15:33 UTC the Macro Agent had written regime strong_bear at 72 percent into the shared state, the same value it had carried through the morning's first short. The Trend Agent on each of its eleven evaluations read that value verbatim. If the Macro Agent had been chatting in prose about mixed signals, the Trend Agent would have had to interpret tone. It does not, so it did not. The fact that two paired entries on the same instrument the same day both ran to TP3 is the coordination at work, not a doubled bet on a single read.
From the SkyAnalyst Team.
At a Glance
What this teaches about AI-driven trading
How does the system decide to take a second position on the same instrument the same day?
Each setup is scored against the live tape on its own merits. The morning's first EURUSD short had already closed at TP3 by the time this second evaluation cycle began at 15:33 UTC. The Trend Agent does not increase size on a fresh entry because a prior trade worked, and the system does not skip a setup because a prior one already ran. The second entry is gated by the same closed-bar trigger and confirmation volume as the first.
Why did the system wait through ten evaluations when confidence cleared the floor on most of them?
Confidence above threshold is a permission to act once the trigger prints, not an instruction to act before it does. Across the ten waits, the score floated above 60 percent on seven of them and peaked at 74 percent on the sixth, but the closed-bar rejection candle with confirmation volume did not appear until the eleventh. The system requires both the score floor and the trigger conditions on the same evaluation cycle.
What does +1.81R translate into dollar terms for a typical account?
On a hypothetical $100,000 account at 2 percent risk per trade, 1R equals $2,000, so +1.81R (TP3) translates to roughly +$3,620 of potential return. That figure assumes the position is held to the highest take-profit reached. In live execution the broker scales out at TP1 for risk management, so the recorded broker P&L is smaller than the full-arc R-multiple shown.
How does the system handle the FOMC sitting ahead in the calendar at trade entry?
The calendar gate blocks new entries inside a window before scheduled high-impact events. On March 18 at 15:48 UTC, the 19:00 UTC FOMC was outside the no-entry window, so the gate cleared. The Risk Agent did compress the size envelope because volatility risk remained elevated heading into the print, but the trade was permitted to size and run. The position closed at 18:43 UTC, before FOMC, and the targets were not affected by the event.
Run your markets with SkyAnalyst
Seven-day free trial. No credit card. Full access to the Trend Agent, Macro Agent, and six-factor confluence scoring.
Trading involves substantial risk of loss. Past performance is not indicative of future results. The analysis shown was produced by an AI model operating on SkyAnalyst’s live trading infrastructure; it is shared for educational and research purposes only and is not financial advice. About reported results. Each model outputs three take-profit targets (TP1, TP2, TP3) per trade. In live execution, models typically scale out at TP1 for risk management — the broker position records this as a TP1 exit. The R-multiples and dollar returns shown in this article reflect the full potential of the trade: where the market actually traveled to (the highest take-profit hit, or stop loss) before the setup was invalidated or exhausted. This lets readers see the complete arc of each setup, not just where the position was closed. Simulated returns in this article are calculated against a hypothetical $100,000 account at 2% risk per trade (1R = $2,000). These are educational reference figures and do not reflect any specific account or broker execution. Your actual result depends on your position size, your risk parameters, and live market conditions.
From the SkyAnalyst Journal
March 2026 Monthly Recap: 42 Trades, Variance Both Directions, Break-Even Net
Forty-two trades. Twenty-two winners, twenty losers, 52.4 percent win rate. Net minus 0.13R, essentially flat on a TP1 baseline. The month produced both the deepest published drawdown and the bumper week of the record.
USDJPY Short on March 31 - Closing the Month on a TP3 Winner
A pullback short on USDJPY entered at 159.23 ran to TP3 at 158.75 in 2h 32m, closing at +3.20R. The closing-day winner of a March that finished -0.13R / 22W-20L on the TP1-baseline tally.
EURUSD Long on March 31 - Second of Two Same-Day TP3 Winners
A Bullish Pullback long on EURUSD entered at 1.1520 ran to TP3 at 1.1558 over four hours and seventeen minutes, closing at +1.58R. The second of two TP3 winners on the closing day of a near-flat March.