The Cable Short That Came 1.1 Pips From Its Stop, Then Ran to TP3
SkyAnalyst AI journal entry: GBPUSD Short on Sep 21, 2026 closed +2.47R 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.
At 20:02 UTC on Monday, five hours after SkyAnalyst sold Cable at 1.33735, the position was 6.9 pips in profit, 0.46 of a risk unit. New York had closed, liquidity was thinning, and price had tested the same support three times without breaking it. The desk asked whether to stay in. The re-analysis answered in one line: stay in, let London do the work. Then the overnight session went the other way. Price squeezed back up through the entry and printed 1.33874, 1.1 pips below the 1.33885 stop and well above the 1.33805 line the review had said should not be reclaimed. For part of the night, the trade that had been slow looked like a trade that was wrong. London opened at 07:00 UTC. TP1 filled at 07:33, TP2 at 08:16, and TP3 at 1.33365 at 08:31 UTC, 17 hours and 25 minutes after the fill, for +2.47R (TP3). This case study is about both halves of that night: the reasoning that said hold, and how close the stop came to overruling it.
A London failure, then a dollar that would not give
The morning's story was written before New York opened. During London, GBPUSD pushed up into 1.3400, right on top of the prior day's high at 1.33984, and could not hold it. By 15:00 UTC it had reversed through every level that mattered: below yesterday's low at 1.33779, below the daily pivot at 1.33756, and below intraday VWAP near 1.3384. The analysis called London a reversal to bearish, not a clean trend, and that label shaped everything after it. The same London-first logic sat behind a Cable short earlier this month.
The dollar side agreed. DXY was at 100.348, above its five-day EMA of 100.24 and pressing the upper part of its range. VIX sat at 14.6, below its own five-day EMA of 15.5. That combination reads as a firm dollar in a calm market, not a panic bid, and it matters for Cable because the pair's inverse relationship with the dollar was intact. No divergence, so no veto.
The agents lined up behind the same read. The Trend Agent had GBPUSD BEARISH at 68 percent in a TRENDING regime, with invalidation at 1.3389. The Macro Agent leaned bear at 60 percent, citing the gap between a Fed still signalling hikes and a Bank of England in wait-and-see mode. Neither read was loud. Both pointed the same way.
That produced a narrow instruction rather than a broad opinion: sell rallies and sell failed retests, do not buy dips. If you want to see that kind of instruction built on your own markets, SkyAnalyst runs the same analysis engine with a 21-day free trial.
Failed Pivot Retest. The pattern is a short taken on the way back up, not on the way down. After a market loses a level, it often returns to test it from underneath. If the old support now holds as resistance, and a candle closes back below it, the retest has failed and sellers are in control of that price.
Where the level came from
The retest zone was 1.33750 to 1.33760, and it was three levels sitting on top of each other. The daily pivot at 1.33756, the Trend Agent's key resistance at 1.33759, and nearby round-number structure all sat within a single pip of each other. Levels that coincide like that are where a retest is most likely to be decided quickly.
What counted as the trigger
The rule was written before price arrived: only take the short if price retests 1.33750 to 1.33760 and a completed 5-minute candle closes back below 1.33735, before 11:30 ET. A touch alone was not enough. A candle still forming was not enough. The close was the signal.
Where the risk sat
The stop went at 1.33885, 15 pips above the fill and inside the Trend Agent's 1.3389 invalidation. Fifteen pips is also the minimum stop the playbook allows on Cable, so this was the tightest structure the rules would accept. The targets stepped down through 1.33593, 1.33501 and 1.33365, which work out to 0.95R, 1.56R and 2.47R against that stop.
The analysis also named its own weak point. Support at 1.33687 was close below the entry, and because Cable had already fallen for hours, it flagged TP1 as a level to treat as final if momentum hesitated.
SkyAnalyst doesn't favor any single strategy. The same system that sold this retest bought a Nasdaq pullback four days earlier. It reads the tape first, and the pattern is whatever the tape offers that day.
London set the tone with a failed upside extension: GBPUSD pushed into 1.3400 / prior-day high 1.33984 during London, then reversed and is now trading below yesterday’s low (1.33779), below the daily pivot (1.33756), and below intraday VWAP (~1.3384). That makes the London read reversal-to-bearish, not a clean bullish trend. On the macro side, DXY is above its 5-day EMA (100.348 vs 100.24) and pressing the upper part of today’s range, while VIX is below its 5-day EMA (14.6 vs 15.5), so this is USD-firm / calm-risk rather than panic risk-off. The inverse GBPUSD-DXY relationship is intact today, so there is no divergence veto. Trend Agent is bearish, 68% confidence, trending regime, with 1.3389 invalidation; Macro Agent is lean_bear, 60% confidence, citing Fed-vs-BoE rate differential and DXY firmness. Net: NY AM bias remains sell rallies / sell failed retests, not buy dips.
Directional Bias: Bearish
Volatility: Normal
Setup #1: GBPUSD SHORT
- Entry: 1.33735-1.33745
- Entry Trigger: Only take if price retests 1.33750-1.33760 and prints a 5m rejection / close back below 1.33735 before 11:30 ET. This is the key confluence band: daily pivot 1.33756, Trend Agent resistance 1.33759, and nearby round-number structure.
- Stop Loss: 1.33885-1.33890
- Targets: TP1=1.33593, TP2=1.33501, TP3=1.33365
- R-Multiples: TP1=1.0R, TP2=1.6R, TP3=2.5R
- Quality Score: 7.8/10
- Confidence: High - 6/7 confluences:
- London bias bearish after failure at 1.3400 / prior-day high
- DXY supports short bias (above 5-day EMA, firm intraday)
- Macro Agent aligns bearish (60%)
- Trend Agent bearish, moderate confidence, trending regime
- 60m structure bearish: below fast/slow EMAs, below VWAP, RSI ~44, MACD negative
- 5m entry uses a defined level with visible reaction
- No high-impact USD/UK event within 30 minutes
Risks: calm VIX may limit extension, 1.33687/1.33683 is nearby support, and late-session participation drops after 11:30 ET. Because Cable has already pushed lower for hours, TP1 should be treated as a forced partial/finalization level if momentum hesitates.
- Invalidation: Any 5m close above 1.33890, or a reclaim/hold above VWAP ~1.33842 followed by acceptance above the 1.33756-1.33759 resistance band. If not triggered by 11:30 ET, cancel the setup.
GBPUSD SHORT: Full Trade Review
Current Position Status
| Field | Value |
|---|---|
| Direction | SHORT |
| Entry | 1.33735 |
| Current Price | ~1.33666 |
| Stop Loss | 1.33885 |
| TP1 | 1.33593 |
| TP2 | 1.33501 |
| TP3 | 1.33365 |
| Floating P/L | +6.9 pips (~0.46R) |
| Time in Trade | ~5 hours |
| Progress | None hit yet |
1. Is the Trade Still Valid?
Yes, the trade thesis remains intact. Here is the evidence:
Trend Structure (Unchanged)
- Trend Agent: BEARISH, 72% confidence, TRENDING regime, no direction changes in 4 hours
- Price remains below the 60m fast EMA (1.33753), slow EMA (1.33802), and VWAP (1.33805)
- 60m MACD line at -0.00028, histogram -0.00021, bearish and expanding
- 60m RSI at 41.2, controlled downside, no oversold exhaustion signal
- Invalidation level (1.33805) has not been touched
Macro Alignment (Unchanged)
- Macro Agent: lean_bear, 65% confidence
- Fed-BoE divergence still the dominant driver
- DXY at 100.413, above 5-day EMA (100.253): dollar firm, supports GBPUSD shorts
- No GBP-specific catalysts have emerged to change the picture
- No high-impact events remaining today
Price Action Assessment
- Since entry at 1.33735, price has carved a descending range between 1.33747 (high) and 1.33647 (low)
- The last 5 hours show lower highs: 1.33756 → 1.33747 → 1.33732 → 1.33729 → 1.33747 → 1.33674
- Today's low sits at 1.33641, just 2.5 pips below the 60m support at 1.33647
- The 19:45 UTC 15m candle dropped to 1.33664 before a minor bounce, sellers still probing
Key Concern: Stalling Near Support
The pair has been grinding the 1.33647-1.33670 zone for the past ~3 hours without a clean break. This is the one yellow flag:
| Factor | Reading | Implication |
|---|---|---|
| Price vs 1.33647 support | Tested 3x, held each time | Support is bending but not broken |
| Volume | Low (86 ticks on latest 60m bar) | NY session closed, no fresh catalyst |
| 5m MACD | Just crossed below zero | Renewed micro-selling, but weak momentum |
| 5m RSI | 39.9 | Bearish but not stretched |
| ATR expanding? | 60m: yes, 15m: yes | Volatility is trying to expand, potential for a move |
2. Confluence Scorecard: Current State
| # | Confluence Factor | Status |
|---|---|---|
| i | London session bias alignment | ✅ London reversed from 1.3400, bearish into NY |
| ii | DXY supports trade | ✅ DXY above 5d EMA, rising today |
| iii | Macro Agent aligns (≥60% conf) | ✅ lean_bear, 65% confidence |
| iv | Trend Agent aligns (moderate+) | ✅ BEARISH, 72% confidence, trending |
| v | 60m EMA stack supports direction | ✅ Price below fast/slow EMAs and VWAP |
| vi | Entry at defined level with reaction | ✅ Entered at pivot/resistance rejection |
| vii | No high-impact event imminent | ✅ Clear calendar |
Score: 7/7, Very High conviction remains
3. Risk Assessment
What's Working
- Every structural bearish signal is intact: no EMA reclaims, no VWAP reclaim, no RSI bullish divergence
- Stop at 1.33885 is 21.9 pips away and well above VWAP (1.33805) and the Trend Agent invalidation (1.33805), very safe
- DXY continues to firm, providing macro tailwind
- No sign of V-reversal behavior, no volume spike, no aggressive bid
What to Watch
- 1.33647 support: three touches without a clean break. If this holds through the Asian session, a squeeze toward 1.33750 is possible before the next London push
- Session transition: NY has closed, liquidity is thin. Moves in the next 4-6 hours will be unreliable and potentially noisy
- TP1 at 1.33593 is 7.3 pips below today's low (1.33641), needs a fresh breakdown, most likely during London or the next NY overlap
4. Recommendation: STAY IN: With Management Plan
Verdict: Hold the position
The trade is working exactly as intended, controlled, grinding bearish price action with full multi-timeframe alignment and zero invalidation signals. The only issue is pace, not direction.
Management Plan
| Action | Condition |
|---|---|
| Hold as-is | Default. All bearish structure intact, no reason to exit early |
| Move stop to breakeven | If price breaks and closes below 1.33641 (today's low) on a 15m or 60m candle, this confirms new support-turned-resistance at the old floor |
| Partial at TP1 (1.33593) | Take 40-50% off here, it's at the 1R level and near the next structural support (60m S2 at 1.33352 cluster area). Lock in profit |
| Trail remainder | After TP1, trail stop to 1.33670 (today's low zone) for TP2/TP3 |
| Exit / reassess | If price reclaims and closes above 1.33756 on a 15m candle, that would reclaim the daily pivot and signal the bearish structure is breaking. This is well below your stop but protects the current floating profit |
Overnight Considerations
- Thin liquidity may produce a short squeeze toward 1.33720-1.33750 during Asian hours, this is normal and does not invalidate the trade as long as 1.33805 (VWAP / Trend Agent invalidation) is not reclaimed
- The real test will come during London open (3:00 AM ET): if sellers re-engage below 1.33647, TP1 becomes highly probable
- Your stop at 1.33885 provides enough room to survive overnight noise
Bottom Line
The trade is +0.46R in profit with 7/7 confluences still intact, no invalidation signals, and the entire intraday structure pointing lower. There is zero technical reason to exit. The main risk is overnight noise, but your stop accommodates that. Stay in, let London do the work, and manage per the plan above.
Decision log
15:02 UTC, 77 percent, WAIT. Price had traded back into the 1.33750 to 1.33760 retest area, and the broader structure was bearish: below the 5-minute and 15-minute EMAs, below VWAP, 15-minute MACD negative. But the last completed 5-minute candle had closed at 1.33738, three tenths of a pip above the 1.33735 trigger. The move below it was happening on a candle that had not finished. The system would not short a condition that had not yet printed.
15:05 UTC, 64 percent, ENTER. The next completed candle did the job: a high of 1.33751, inside the retest zone, and a close at 1.33723, back below 1.33735, before the 11:30 ET cutoff. The trigger had fired. Confidence came in lower than the wait because the system was honest about the cost of arriving three minutes later: price was now a little below the preferred entry zone and closer to support at 1.33687. It entered anyway, because the rule had been met. The fill was 1.33735 at 15:06 UTC.
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 hitActual | +0.95R | +$1,900 |
| TP2 hit | +1.56R | +$3,120 |
| TP3 hit (max potential) | +2.47R | +$4,940 |
All six agents combined.
What this trade teaches
The entry was a rule being met, which is supposed to be undramatic. What deserves study is the night after it.
By 20:02 UTC the position had moved 6.9 pips in its favour and then stopped. Price ground against 1.33647 three times and held. Volume on the latest hourly bar was light, New York had closed, and TP1 at 1.33593 still sat below the day's low of 1.33641. A trader watching that screen has every reason to take the small profit and go to bed.
The mid-trade review did not argue from hope. It rechecked every input from the morning. The Trend Agent was still BEARISH, now at 72 percent, with no direction changes in four hours. The Macro Agent was still leaning bear, at 65 percent. DXY had firmed further, to 100.413, above its five-day EMA of 100.253. Price was below the hourly fast and slow EMAs and below VWAP. The confluence count had gone from six of seven at entry to seven of seven. Nothing the trade was built on had broken. The only thing wrong was the pace.
The trade is working exactly as intended ... The only issue is pace, not direction. SkyAnalyst mid-trade review, 20:02 UTC
That call held up, but only just. The review expected a squeeze in the thin hours before London and called it normal, sizing it at 1.33720 to 1.33750. Price went to 1.33874. The 15-pip stop, the tightest the playbook permits on Cable, had 1.1 pips left.
So the lesson is narrower than "holding was right." The hold was survivable because the stop had been set at structure in the morning and was never moved. The review's softer guidance, such as reassessing on a 15-minute close above 1.33756, was commentary. The position itself was governed by its stop and its targets, and the stop is the part that decided whether there would be a London session to wait for.
One honest caveat. The review also proposed taking 40 to 50 percent off at TP1 and trailing the rest. That is not how this desk executes: the broker closes the whole position at TP1, so the realized result is +0.95R (TP1). The +2.47R (TP3) headline is where the market went, and we report both so the gap between them stays visible.
From the desk
We picked this trade because the result hides the night. A reader who sees +2.47R (TP3) sees a clean short. The record shows a position that was 0.46R in profit at 20:02 UTC and about 0.93R under water before London opened.
The entry itself was quiet. No dramatic confidence spike, no long string of refusals, just two evaluations three minutes apart and a fill. Everything interesting happened after the order was on.
Trades like this one tend to die in one of two ways. Someone closes a correct position at plus seven pips because the evening is quiet, or someone moves the stop when price comes back for it in the small hours. Neither happened here. The stop stayed where the morning put it.
London did the rest quickly. TP1 filled at 07:33 UTC, thirty-three minutes after the open, and TP3 at 1.33365 printed at 08:31. From the fill, that is 37.0 pips on a 15-pip stop.
This is one trade, and one trade proves very little. A 1.1-pip margin is not a method. What it does show is a stop placed at structure and then left alone while the market took its time.
At a Glance
What this teaches about AI-driven trading
What is a failed retest short in forex?
A failed retest short sells a market that has broken below a level and then climbs back to test it from underneath. If the old support acts as resistance and a candle closes back below it, the retest has failed. The entry sits close to the level, so the stop can go just above it, which keeps risk small relative to the targets below.
Why wait for a candle to close instead of entering on the touch?
A touch shows price reached a level. A close shows where buyers and sellers agreed to leave it. Intrabar moves below a trigger often reverse before the candle ends, so an entry on a forming candle can be stopped by noise. Waiting for the completed close costs a few pips of entry price and removes many of those false signals.
How should a trader decide whether to hold a slow trade overnight?
Recheck the reasons you entered, not the P&L. If trend direction, macro bias, structure and the invalidation level are all unchanged, the trade is slow, not wrong. If any of them has broken, exit. Thin liquidity after the New York close can produce noise, so the stop must be wide enough to survive it without being moved.
Why is Cable sensitive to the London open?
GBPUSD is a London-centric pair. UK banks and European desks provide much of its liquidity, so direction for the day is often set during London hours, and New York extends or reverses it. A move that stalls in the thin hours after New York closes frequently resumes when London participation returns around 3:00 AM Eastern.
What is the difference between full-potential R and realized R on this trade?
Full-potential R measures how far the market travelled before the setup was exhausted: here TP3, for +2.47R (TP3). Realized R is what the account booked. This desk closes the entire position at TP1, so the logged result is +0.95R (TP1). Reporting both shows the whole move and the conservative entry in the track record.
Run your markets with SkyAnalyst
21-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. Every AI Trader publishes three take-profit targets (TP1, TP2, TP3) per trade. The broker closes 100% of the position at TP1, so two distinct R-multiples appear in this article. The hero R-multiple is the full-potential R: where the market actually traveled (the highest take-profit hit, or the stop loss) before the setup was invalidated or exhausted. The realized R, shown on the TP1 row of the simulated returns panel, is TP1’s R (or -1R on a stop out). The realized R is what we log to our running track record. Both numbers are honest. Showing both is what lets readers see the full arc of the move and the conservative ledger entry it produced. 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
Same 62% Score, Two Answers: A GBPUSD Short Back to London's Low
Two evaluations 79 seconds apart, both at 62 percent: one wait, one enter. A GBPUSD short sold at 1.33607 reached TP3 at 1.33260, 0.4 pips above the London low, with 1.5 pips of heat.
Every One of This Week's Six Losses Cost Exactly 1R.
Six stop-outs, a four-loss streak on Monday, and a 9.9% drawdown on the simulated account. Not one of them cost more than the 1R it was sized for, and that is the only part of this week we are happy about.
The Regime Flipped on Monday. We Kept Selling Until Friday.
Seven trades, one winner, minus 4.95R. Every loss was a short, the average short died sixty-seven minutes after it was filled, and the only trade that went the other way was the only trade that paid.