SPY falls 1.08% as 26 failed breakouts outpace 20 winners; AMD leads picks
Market recap, bot performance, and scanner analysis for Thursday, July 23.
The UnxEdge breakout scanner picked up 71 wedge pattern setups today, and the ugly part is that the better trades were mostly on the short side as sellers finally got paid. This was a wedge breakout session where weak breadth mattered more than stock picking, and the scanner saw it before the indexes finished sliding.
$SPY lost 1.08%, $QQQ dropped 1.36%, and small caps held up better, but the move was not random. Today’s tape leaned risk-off as speculative biotech headlines stayed messy, from funding stress around SMMT-linked sentiment damage in the group to more low-quality deal flow and listing churn, while bank earnings were fine but not strong enough to carry financials into leadership. The practical takeaway was simple: strength had to be earned, and most clean continuation moves showed up in inverse tech products, index shorts, and isolated relative-strength names.
Sector rotation stayed narrow. Gold rotated in week over week while housing rotated out, which fits a tape that is getting more defensive under the surface even before any major macro event forces the issue. Semis were still strong on a 5-day relative basis despite this week’s underperformance trend, which means traders need to separate swing leadership from daily execution. That distinction matters when names like AMD can still screen well statistically even as the broader $QQQ complex bleeds lower. For context, revisit yesterday's debrief.
Market Snapshot
Index damage was real, but the scanner stayed selective
The scanner skewed bearish from the start, with 45 bear setups against 26 bulls. Only 20 names actually broke out while 26 failed, which tells you this was not clean trend expansion across the board. It was selective downside pressure with a few real runners and a lot of noise. Setup grades remained respectable, including 26 A-range setups, and traders can see live setups in the scanner to track how that quality is evolving into tomorrow.
Yesterday's Pick
FUBO bear got stopped, no excuses
This one failed cleanly. The setup never delivered downside follow-through and hit the stop before testing targets. Accountability matters more than narrative here: yesterday’s call did not work.
Arxe Pick of the Day
AMD got the nod on stats, not vibes
Verdict: trade_it
AMD had the best combined backtest profile of the approved list: 77.78% win rate, 3.5 profit factor, and the largest sample size at 18 trades. That matters more than marginally better pressure elsewhere because sample size is what keeps a setup honest.
The setup was fresh at three bars and still actionable, which is exactly what you want in a tape where late entries get punished. Even with semis losing relative ground this week, AMD still offered the strongest statistical edge among the approved candidates. Traders looking at grade quality and trigger structure can see live setups in the scanner.
Today’s actual Arxe swing pick was XBI BEAR with HIGH conviction, but neither bot converted it. The reason was straightforward: the setup failed bot-side qualification before execution, likely on contract or pre-filter constraints rather than discretionary rejection. That is frustrating, but it is better than forcing low-integrity fills.
AMD scanner chart – July 23, 2026 | UnxEdge
Bot Performance
Nothing traded, and that is the honest state of play
Wex and Xcel are in training mode while strategy is being refined. The scanner and Arxe intelligence remain fully active. When trading resumes, every trade will be published here. You can watch Wex and Xcel trade live in the Edge Lab once execution is back on.
Missed Trades
The tape gave runners, and the system sat out all of them
Zero breakouts traded. Thirteen missed. Total opportunity cost: +32.09R. That is not a small miss, but it also was not one type of miss. Some of these were exactly the kind of outsized moves filters are designed to avoid chasing. Others were real blind spots.
The standout was TMO at +12.5R, which is the kind of move that distorts the whole day’s missed-trade stats. The index complex also handed traders textbook downside vehicles through DIA, QQQ, SQQQ, and TQQQ. When the same directional message repeats across cash index, inverse ETF, and levered inverse ETF, the filter stack needs to recognize that alignment faster.
Breakout Scorecard
More failures than breakouts, which is exactly why selectivity mattered
That 43.5% breakout rate is mediocre at best. It confirms the tape was not forgiving. If you were buying random wedge breakouts, you likely got chopped. If you stayed aligned with index pressure or found isolated names with real momentum, you had a chance.
Notable triggers included ADBE BEAR B+, ANET BULL B+, BMY BULL B, C BEAR B+, CEG BULL B+, and DIA BEAR B.
Watch Tomorrow
The A-range names still coiling into Friday
This is the section that matters. These are the unresolved names with enough quality to stay on the board, but they still need confirmation. Levels below are framework levels built from the active coil. Traders should refine against the live range before execution and see live setups in the scanner for real-time updates.
Live breakout scanner · Autonomous AI traders · Fully transparent track record.