Breakouts led the week while sweet spot setups stayed selective
Backtest patterns, sweet spot performance, and missed-trade analysis for July 20 to July 24.
THE EDGE THIS WEEK
Breakout volume stayed active, but selectivity did not convert
This week showed a clear split between scanner activity and actual breakout efficiency. Setup volume was healthy, but the 18.9% breakout rate and the weak 27.5% result in the usual pressure-and-timing sweet spot suggest that participation was broad while follow-through was narrow.
The stronger edge appeared in symbol-specific persistence rather than broad market-level confirmation. A small cluster of names delivered extremely clean outcomes, while many otherwise tradable-looking setups failed to expand, which is a pattern traders should treat as an environment that rewards precision over coverage.
BY THE NUMBERS
Weekly backtest and scanner profile
The broad backtest remained constructive at the aggregate level, with 64.9% average win rate across 2,353 trades. But live scanner behavior told a more selective story, where a large setup count did not translate into a high conversion rate.
That gap matters. When setup volume is high but breakout rate stays under 20%, traders often need to shift from abundance thinking to filtration thinking. In practical terms, that means fewer automatic takes, more emphasis on context, and tighter attention to symbols already proving they can trend. For current candidates, see live setups in the scanner.
SWEET SPOT REPORT
The usual low-pressure timing window underperformed sharply
The sweet spot setup, defined by pressure below 60 and 3 to 5 bars of structure, delivered only 27.5% this week versus a 59.4% historical win rate across 229 trades. That is not a small drift. It is a regime-level deviation in behavior, even if the broader volatility regime could not be classified.
Research-wise, this suggests the market did not reward the normal balance of compression and early timing. In weeks like this, lower-pressure setups can appear technically clean but fail because they never attract enough urgency after trigger. The pattern usually points to one of two conditions: either breakouts are happening from stronger institutional sponsorship than the sweet spot captures, or price is requiring more confirmation before expansion.
When a historically reliable timing window cuts its win rate by more than half, the lesson is not to abandon the setup. It is to reduce default confidence and demand stronger confirmation from symbol quality, sector participation, or post-trigger momentum.
SYMBOL SPOTLIGHT
Three names that reveal the week’s structure
LLY finished in the 100% club with 10 wins in 10 trades and 0.9999 average R. The key research point is not just perfection. It is sample size plus consistency. A 100% result with double-digit trade count suggests repeated alignment between setup quality and follow-through, which is exactly the type of symbol behavior traders should prioritize when the broader scanner is noisy.
GDX posted a 94.44% win rate over 18 trades with 0.8863 average R. This is one of the more important entries in the report because it combines high frequency with high efficiency. When a symbol generates this many opportunities and still maintains that level of output, it often reflects a cleaner trend structure than the general market is offering.
COST was the opposite case, with just 18.18% win rate across 11 trades and -0.6366 average R. Weak symbols like this are useful because they reveal what failed setups looked like this week: respectable activity, repeated triggers, and very poor reward conversion. That profile often shows up when a name is technically active but structurally resistant to sustained expansion.
Other strong names such as NOW, URI, CVX, ZM, and V reinforce the same conclusion. The edge was concentrated in a relatively small group of symbols that kept proving themselves, not evenly distributed across the opportunity set.
WHAT THE BOTS MISSED
Filter discipline prevented noise, but it also blocked meaningful upside
The bots missed 14 trades worth +50.0R, including 7 runners that reached TP3. The dominant blocker was rvol_threshold, responsible for 11 of the 14 misses, far outweighing spy_alignment and not_on_watchlist.
This is a valuable research result because it shows the cost of strict volume gating in an environment where broad breakout conversion was weak but isolated names still expanded well. Relative volume filters are excellent at reducing low-quality participation, but they can lag when valid moves emerge from quieter starts and only accelerate after trigger.
The implication is not that rvol should be removed. It is that the filter may need adaptive treatment. When scanner-wide breakout rates are low, a rigid rvol requirement can protect capital. But when missed-trade R accumulates this quickly, it may be worth testing a secondary pathway for elite symbols or sectors already showing repeat success. Traders can watch the bots in the Edge Lab to see how these filters behave in real time.
The most expensive mistake this week was not overtrading. It was allowing a useful filter to become too blunt in a market where leadership was narrow but real.
SECTOR HEAT MAP
Breakouts clustered in growth and leadership pockets
Technology led the breakout count with 23, followed by Financials and Healthcare at 10 each, ETFs at 8, Consumer at 7, Chinese ADRs at 5, and Industrials at 4. There was also a separate technology bucket with 4 breakouts, which likely reflects a classification split rather than a distinct group. Combined, that would push technology-related participation even higher.
The pattern is important. Breakouts were not evenly dispersed. They concentrated in areas where institutional sponsorship can create repeated trend continuation, especially in large-cap growth and liquid thematic exposure. That aligns with the strong symbol-level outcomes in names such as NOW and the high scanner activity in semis and infrastructure-adjacent names like $ASML, $AVGO, $AMD, and $VRT.
At the same time, weaker performance in names like XLF and XLB shows that sector breakout counts alone were not enough. Traders needed both sector participation and symbol-level quality. Concentration helped, but only when paired with instruments that could actually convert triggers into sustained movement.
RESEARCH NOTE
This was a week for narrowing the funnel, not widening it
The deeper takeaway is that high activity can be misleading when follow-through is uneven. With 465 setups, only 88 breakouts, and a sweet spot win rate far below its historical norm, the market rewarded traders who concentrated capital in proven names rather than distributing attention across every technically valid trigger.
That creates a useful framework for next-week testing. Instead of asking whether the scanner is busy, ask whether leadership is repeating. If the same symbols, sectors, or structural profiles keep producing, then concentration is justified. If not, setup count becomes noise. The research edge is often found not in finding more signals, but in recognizing when the market only wants to pay a small subset of them.