US30 Short Hits TP2 Overnight — GPT-5.5's First Win of Season 2 at +1.65R
GPT-5.5 entered the Dow short at 49,525 on bearish breadth and held through the European session. The broker filled TP2 at 49,265 — the largest realized R on either side of the benchmark this season.
The AI Trading Benchmark runs Claude and GPT side by side on the same live market data, the same instruments, the same risk per trade, and asks one question: which model trades better. Season 2 opened with two losses for GPT — a NAS100 short stopped out on April 30 and another stopped on May 6 — and the model carried a -$2,000 hole into mid-May. Today changes that. GPT-5.5 shorted US30 into the New York morning breakdown at 49,525, held overnight while price ground lower through the European session, and was filled at TP2 (49,265) on the broker side. The realized R, per the benchmark's TP1-full-close policy, is +1.65R — the highest single-trade R of the season so far on either side of the experiment. This article walks through the setup that produced it, the macro context around the entry, and what the trade reveals about GPT-5.5's decision-making under transitional regimes.
About reported results. Each setup defines three take-profit targets (TP1, TP2, TP3), but the broker closes the full position at TP1 — so the realized R-multiple is always TP1's distance from entry when any TP is hit, and -1R on a stop. The dollar P&L shown in this article is the actual broker close at TP1 (or stop). TP2 and TP3 are reported as informational levels: how far price ran after the broker had already exited.
Result
R-Multiple
AI Confidence
Win Rate
Season Record
Market Environment — May 19, 2026
The Dow opened May 19 with a clean directional bias and held it. Pre-market sentiment was already cautious — FOMC minutes were due the next afternoon, and traders were positioning short into the risk window rather than fighting it.
The macro tape backed the bearish read. The US Dollar Index (DXY) was bid at 99.42 ahead of the NY open, trading above its 5-day EMA at 99.01 and pressing yesterday's high — a headwind for Dow multinationals that earn outside the US. The 10-year Treasury yield (US10Y) printed 4.679, above its 5-day EMA of 4.578 and above yesterday's high, applying rate pressure to financials and rate-sensitive sectors. The VIX held at 18.00 versus a 5-day EMA of 17.91 — not panicked, but elevated enough to favor short setups with wider stops. None of these readings were extreme on their own. Together they painted a transitional-to-risk-off regime, with breadth indicators (NYAD at -1,019 vs a 5-day EMA of -463) deteriorating into the session.
Equity breadth was the cleanest tell. The NYSE advance-decline line printed a new intraday 5-day low (-1,647 at the bottom) before recovering to -1,019 — still well below its EMA and yesterday's low. For an index like the Dow, where 30 large-cap names carry the tape, breadth that broken keeps the default bias short. The opening range formed between 49,265 and 49,601; the upper bound failed on a VWAP retest before 11:00 ET, and price began to compress under the 49,531 shelf.
Pending Home Sales had already cleared at 10:00 ET, removing the morning's only US data risk. With FOMC minutes still 30+ hours away, GPT-5.5 had a clean window to act on the breakdown setup that the multi-timeframe data was lining up against the 49,520–49,530 zone.
US30 SHORT
Setup: US30 Short Continuation
Analysis by SkyAnalyst AI
Strategy Analysis
GPT-5.5 took the US30 short at 49,525 — right at the upper edge of the 49,520–49,530 entry zone the pre-trade analysis had flagged as the high-quality breakdown/retest area. The model's confidence at entry registered at 62%, classified as moderate. That number matters. Many setups in this experiment have entered at 75–85% confidence and lost; this one entered with measured conviction and produced the season's largest realized R. Confidence is a model's self-reported uncertainty, not its skill.
The setup was a continuation short — not a reversal. Breadth had already broken (NYAD -1,019 versus a 5-day EMA of -463, with an intraday low at -1,647), the Trend Authority Agent classified US30 as bearish with 66% confidence and a transitioning regime, and the 60-minute structure had price below both VWAP (49,640) and the fast/slow EMA cluster (49,591–49,594). The Macro Analysis Agent leaned bearish on the group at 64% but rated US30 itself as neutral at 57%. Mixed but workable. The 5/7 confluence score put the setup in Medium-High range — exactly the band the benchmark's rules require for a tradeable entry.
GPT-5.5 placed the stop at 49,625, just below the Trend Agent's 49,630 invalidation line. That gave the trade approximately 100 points of risk against the 60-minute ATR of 102 — meeting the minimum 1x ATR rule without bloating the stop. TP1 was structured at 49,360 (the prior support shelf, ~1.6R from entry), TP2 at 49,265 (the intraday low retest, ~2.6R), and TP3 was held conditional at 49,084 — only valid if breadth deteriorated further. The model did not chase below 49,510, did not flip directions during the bounce off VWAP, and did not size up on the conviction. It executed the planned trade.
Price moved through TP1 at 49,360 during the European overlap and continued into the Asian session, hitting TP2 at 49,265 at 02:11 UTC the following morning — an 11-hour, 26-minute hold. Per the TP1-full-close policy that governs this benchmark, the broker had already closed the full position at TP1; the realized R is +1.65R (TP2) and the dollar P&L is +$1,650.00 (TP2) on the standard $1,000 risk unit. TP2 is informational — how far price ran after the broker had exited.
This trade is GPT-5.5's first win of Season 2. After two stopped-out shorts (NAS100 on April 30, NAS100 again on May 6) put the account at -$2,000, the US30 short alone covers most of the season's drawdown and lifts the equity walk back toward break-even territory. More important than the dollar number: the entry mechanics held. The model entered inside the planned zone, sized the stop against ATR not against hope, and held the position through the New York afternoon and the European session rather than scratching on the first retest. That is the discipline the benchmark is built to measure.
It is worth saying what the trade did not prove. One Medium-High setup landing TP2 does not validate the model under all regimes. The macro tape lined up against US30 specifically — DXY bid, yields up, breadth broken — and the entry was at the structural level the pre-trade analysis specified, not a chase. Under cleaner conditions, the model has converted at a lower rate. Whether this becomes a pattern or a one-off depends on what happens around FOMC minutes the next day and whether the breadth break holds into the back half of the week.
NEUTRAL 62%Regime:
CHOPPYMacro:
HEADWINDlean_bear 72%The US dollar maintains a modest bid with DXY at 99.29, supported by hawkish Fed minutes signaling rate hikes remain on the table if inflation persists, while the ECB, BoE, and BoC are positioned closer to easing. Growth divergence is pronounced—Eurozone and UK PMIs collapsed today while US manufacturing remains robust—creating a structural USD-positive environment across major pairs, though the move lacks explosive conviction.
17.27 (normal)99.29 bullish4.58 stable104.6 bearishWhat the AI Actually Saw
Full 7.7k-character analysis produced by the AI before entering this trade. Unedited.
- Multi-timeframe technical analysis
- Key support and resistance levels
- Pattern recognition and trade setup identification
- Risk-reward assessment with defined invalidation
- Macro context and cross-asset correlation check
Powered by SkyAnalyst AI
SkyAnalyst AIUS30 NY AM Environment — Bias: Bearish / Sell-the-bounce
Current read: US30 remains breadth-driven and breadth is negative enough to keep the default NY AM bias short. One setup qualifies at Medium-High only if entry is taken near the 49,520–49,530 breakdown/retest zone; chasing lower or shorting too high into VWAP creates risk-rule problems.
1) Breadth + Volatility Driver
| Driver | Current Read | Implication |
|---|---|---|
| NYAD / ADD | -1,019 vs 5-day EMA -463 | Breadth is below EMA and supports shorts |
| Today ADD range | Low -1,647, high -910 | Breadth printed a new intraday 5-day low, but is off the extreme now |
| Vs yesterday | Below yesterday’s low | Clear breadth deterioration |
| VIX | 18.00 vs 5-day EMA 17.91 | Slightly elevated; supports shorts / wider stops |
| VIX extreme? | Not above 5-day high / yesterday high | No “only shorts” compound veto, but longs are low quality |
Breadth bias: Bearish.
Breadth-extreme veto: No longs should be considered while price is weak and NYAD remains deeply negative. Shorts are allowed because NYAD is not at a 5-day high.
2) Macro Regime + Cross-Asset Confirmation
Macro Agent:
- Group bias: lean_bear, 64% confidence
- US30 bias: neutral, 57% confidence
- Tradeability: moderate, 65/100
- Key factors: price below 5-day EMA, breadth deterioration, FOMC minutes tomorrow, Pending Home Sales already released.
Cross-asset confirmation:
- US10Y: 4.679, above 5-day EMA 4.578 and above yesterday’s high → rate-pressure headwind for Dow/financials.
- DXY: 99.422, above 5-day EMA 99.013, near yesterday’s high → headwind for Dow multinationals.
- VIX: modestly above EMA → not panic, but not risk-on.
Regime classification: Transitional-to-risk-off.
Breadth and VIX both lean bearish, but VIX is only mildly elevated and Macro Agent’s US30-specific confidence is below 60%, so conviction is not maximum.
3) Trend Structure + Key Levels
Trend Agent:
- Direction: Bearish
- Confidence: 66%
- Regime: Transitioning
- Recommendation: Reduce size
- Key resistance / invalidation: 49,630
- Key support: 49,360
- VWAP: 49,646 area
60-minute structure:
- Price is below 60m VWAP near 49,640–49,646.
- Price is also below fast/slow EMAs near 49,591–49,594.
- RSI: 46.97, neutral-bearish.
- MACD: above zero but below signal, histogram negative → momentum still deteriorating.
- 60m ATR: ~102 points, so stops must be at least about 100 points and structural.
Important levels:
- Current reaction zone: 49,524–49,531
- Intraday/VWAP resistance: 49,579–49,601
- Trend invalidation: 49,630
- Support: 49,360
- Intraday low / OR low: 49,265 / 49,252
- Prior-day low magnet: 49,084–49,099
4) Lower-Timeframe Entry Read
15-minute:
- Price below EMA fast/slow.
- RSI near 47, not oversold anymore.
- MACD below zero, histogram negative but improving → bounce is corrective, not yet a bullish reversal.
- Bias: still bearish, but avoid chasing into support.
5-minute:
- Price below VWAP near 49,579.
- EMA fast below slow; price has bounced but remains structurally below VWAP.
- MACD histogram positive from oversold bounce → this is a countertrend rebound, not a long setup.
- Opening range: roughly 49,265–49,601. The high at 49,601 failed; price remains inside the OR.
Post-data note: Pending Home Sales was already released at 10:00 ET. The market has already produced the “second-chance” type behavior: bounce toward VWAP/OR resistance, then failure. Do not chase the first move; only act on a structured retest/breakdown.
Qualified Setup: US30 Short Continuation
Directional Bias
Short / bearish continuation, aligned with weak breadth, elevated VIX, bearish Trend Agent, and price below VWAP.
Entry Zone
Preferred entry zone: 49,520–49,530
This is the key breakdown/retest zone around the 5m/15m shelf at 49,524–49,531.
Do not chase below 49,510.
If price breaks straight down without a retest, risk/reward deteriorates because the structural stop must remain near the Trend Agent invalidation area.
Entry Trigger
Enter short only if one of the following occurs:
- 5-minute close below 49,524, followed by a failed retest of 49,524–49,531, or
- Price rejects VWAP/OR resistance at 49,579–49,601, then breaks back below 49,531 with a 5m lower high.
The cleanest trigger is:
Short 49,520–49,530 after a failed reclaim of 49,531, with 5m price holding below VWAP and NYAD still negative.
Stop Loss Zone
Stop zone: 49,625–49,630 maximum, including execution buffer
Reasoning:
- Trend Agent invalidation is 49,630.
- 60m ATR is about 102 points.
- From a 49,525 entry, a stop near 49,627 gives roughly 102 points of risk, satisfying the minimum 1x 60m ATR rule.
- Stop is above the failed OR/VWAP resistance cluster and below/at the Trend Agent invalidation ceiling.
Important: If execution slippage would require a protective stop above 49,630, skip the trade. The stop would exceed the Trend Agent invalidation.
Take-Profit Levels
| Target | Level | Logic | Approx. R from 49,525 / 49,627 risk |
|---|---|---|---|
| TP1 | 49,360 | Trend Agent support / structural support | ~1.6R |
| TP2 | 49,265–49,252 | NY low / intraday low retest | ~2.6R |
| TP3 conditional | 49,099–49,084 | Prior-day low magnet | ~4R+ |
TP3 is conditional only. Use it only if NYAD expands back toward the session low, VIX holds above EMA, and price accepts below 49,265. Without renewed breadth deterioration, TP2 is the more realistic morning-session objective.
Confluence Score: 5/7 — Medium-High, 7.0/10
| Confluence | Pass/Fail | Notes |
|---|---|---|
| NYAD matches short direction | ✅ | ADD -1,019, below 5-day EMA and below yesterday’s low |
| VIX supports short | ✅ | VIX 18.00, slightly above 5-day EMA |
| Macro Agent aligns ≥60 | ❌/Partial | Group lean_bear 64%, but US30-specific bias is neutral at 57% |
| Trend Agent aligns ≥60 | ✅ | Bearish, 66% confidence |
| 60m EMA stack supports | ❌/Partial | Price below EMAs/VWAP, but not a clean bearish EMA stack |
| Price at key level with reaction | ✅ | 49,524–49,531 shelf, OR/VWAP rejection above |
| No high-impact USD event within 30 min | ✅ | 10:00 ET Pending Home Sales was medium impact and already released |
Score: 5/7 = Medium-High.
Main Risks
- Transitioning regime: Trend Agent says bearish but transitioning, not a full trend day.
- 5m momentum bounce: MACD histogram has improved, so shorts should not be chased into lows.
- Macro is not strongly bearish for US30 specifically: US30 Macro Agent is neutral at 57%.
- Stop discipline matters: The setup only works if entered high enough near 49,520–49,530 to preserve 1.5R+ to TP1 while keeping the stop at/under 49,630.
Invalidation Condition
The short setup is invalid if:
- US30 reclaims and holds above 49,630, or
- 5m closes above 49,601–49,630 with NYAD improving materially, or
- price reclaims VWAP and holds above it for multiple 5m candles.
Above 49,630, the Trend Agent bearish thesis is invalidated and shorts should be avoided.
No Long Setup
No long setup qualifies. Longs fail the breadth, VIX, Trend Agent, and macro/trend confluence filters. With NYAD deeply negative and price below VWAP/EMAs, any long would be countertrend and below the required Medium-High threshold.
1 Evaluation — Quick Entry
The AI evaluated US30-Pepperstone 1 time before reaching a final decision. Each step shows confidence, reasoning, and whether the model was ready to enter.
Price is now inside the preferred short entry zone at 49,521 and has broken back below the 49,531 shelf after a sequence of lower 5-minute highs, while both 5m and 15m price remain below VWAP and the intraday trend agent is bearish. The exact textbook trigger of a completed 5-minute close below 49,524 followed by a failed retest is not fully confirmed from the provided candles, but the rejection below VWAP/OR resistance and current loss of 49,531 provide enough edge for a starter short near the planned level. Momentum is not strongly bearish yet, so confidence is moderate rather than high and the stop above 49,625 must be respected.
Account Performance
Profit taken at TP1 — the full position is closed at the first target to keep results measurable and comparable across models.
Key Takeaways
- Moderate confidence beats overconfidence. The 62% AI confidence at entry was not "high" — it was measured. This is the first win of Season 2 for GPT-5.5, and it came from a setup the model could justify on confluence, not from a setup it believed in unconditionally. Several earlier 75%+ confidence trades stopped out.
- Breadth is the directional anchor for index trades. US30 is a 30-stock index — NYAD breadth deeply negative against its 5-day EMA was the cleanest signal in the stack. Macro and technicals confirmed; breadth led.
- Stops scaled to ATR, not to hope. With 60-minute ATR around 102 points, a 100-point stop at 49,625 met the minimum risk rule without bloating to "give the trade room." Tighter stops on this setup would have been hit on the early-session noise.
- TP1-full-close caps the realized R. Realized +1.65R (TP2) is the broker's actual close at TP1 expressed in TP1-distance R. TP2 and TP3 are informational — how far the move extended after the broker had exited. The reported number is what hit the account, not what the chart showed.
- One trade does not define a season. This is the largest single-trade R of Season 2 on either side of the benchmark, but the Macro Agent's US30-specific read was only 57% neutral. The setup worked; the regime cooperated. The next few sessions will show whether the model can repeat the discipline when confluence is closer to the 4/7 floor.
Tomorrow brings the FOMC minutes — a known volatility event that compresses confluence scores by tightening the no-entry buffer around the release. Expect lower trade count and tighter execution discipline. If breadth holds negative and DXY stays bid, the bias remains short on indices; a reversal of either ends the thesis. — Eduardo, Senior Research Editor
Compare with Isaac’s analysis →Methodology
Both AI models receive identical market data, identical infrastructure, and identical risk parameters. No prompt engineering. No human intervention. Standard API temperature (0.0). Trades executed on demo accounts with institutional spread conditions via Pepperstone Markets. Each model operates with a $50,000 starting balance and 2% risk per trade. All positions are closed at TP1 — the first take-profit target — to keep results measurable and directly comparable across models.
Forex pairs and gold (XAUUSD) have standardized pricing across brokers — the prices in this article will closely match what you see on your own platform. US index CFDs (NAS100, US30, US500) are different: each broker constructs its own index price feed, so entry prices, stop distances, and P&L figures for index trades are specific to Pepperstone Markets. All trades in this experiment were analyzed, executed, and settled on Pepperstone demo accounts using Pepperstone's price feed.
Why This Cannot Be Replicated in ChatGPT or Claude Alone
Copying the analysis prompt into ChatGPT or Claude will not reproduce these results. Neither model has access to live market data — and the data is the foundation of everything.
Every analysis session, SkyAnalyst AI assembles a structured data packet of 50,000–100,000 tokens per instrument from live broker APIs. This is not a price quote. It contains 5 hours of multi-timeframe candle data across 60-minute, 15-minute, and 5-minute charts — each candle carrying full indicator overlays: EMA fast/slow, ATR, MACD with histogram, RSI, volume with SMA, VWAP with standard deviation bands, and others. On top of that: session structure levels (Tokyo, London, New York highs and lows), Fibonacci retracement and extension levels, a rolling 5-day macro window covering the 10Y yield, DXY, VIX, NYAD breadth, oil, and gold — along with additional proprietary data layers, all formatted as structured JSON specifically designed for LLM consumption.
The model never starts from raw data. Before Claude or GPT sees anything, two proprietary SkyAnalyst AI agents — among other internal systems — have already processed the environment: the Macro Analysis Agent produces directional bias with confidence scores and tradeability ratings across intraday and multi-day horizons, while the Trend Authority Agent evaluates technical structure — EMA alignment, momentum, regime classification — and outputs direction, confidence, key levels, and invalidation prices. The trading model synthesizes what these agents and preprocessing layers have already evaluated. This multi-agent pipeline is what produces the quality of analysis shown in this article — a single prompt to a single model, no matter how detailed, cannot replicate what multiple specialized systems produce in sequence.
The goal is to emulate what a professional trader actually does: read the macro environment, analyze multi-timeframe technicals, identify a setup with defined risk, wait for precise entry conditions, and execute with discipline. SkyAnalyst AI provides the infrastructure that gives the trading model everything it needs to do this — live data, preprocessed context, real-time monitoring, and broker execution. This is not a chatbot experiment. It is an institutional-grade trading pipeline where the AI model is the decision-maker, operating under the same conditions and constraints a professional desk would demand.
Trading involves substantial risk of loss. Past performance is not indicative of future results. These are AI model results shared for educational and research purposes only. Not financial advice.
Stay in the loop
Want the framework behind these trades?
Get every trade analysis, weekly battle report, and the full AI Trading Playbook delivered directly to you.
- Daily Trade Analysis
- Weekly Battle Reports
- AI Trading Insights
- The AI Trading Playbook (free)