SPY slips 0.27% as 33 failures outpace 25 breakouts; PYPL is the pick
Market recap, bot performance, and scanner analysis for Tuesday, May 12.
Hot CPI hit first, and tech never really recovered from it.
Tuesday was a clean example of why index-level damage can hide underneath a busy scanner. SPY only gave up 0.27%, but QQQ fell 1.03% and IWM lost 1.08%, which tells you exactly where the pressure lived: growth and small caps got sold while the broader tape tried to keep the damage contained. This was not a broad panic. It was a rotation-heavy, rate-sensitive sell day.
The macro driver mattered. April CPI landed this morning, and the market spent the session repricing what sticky inflation means for the next Fed step. With PPI due tomorrow, traders were not eager to add duration or chase high multiple names after the data. That pressure lined up with sector leadership too: semis still held up on a relative basis over the last five days, but today the Nasdaq was the weak link, while defensive and hard-asset pockets like gold and metals continued to outperform. Elsewhere, single-name headlines were mostly noise. AEP launching a $2.60B common stock offering reinforced the cost-of-capital backdrop, and the rest of the tape was dominated by earnings and corporate actions that did little to change the macro script. For context on how Monday set up today’s handoff, revisit yesterday's debrief.
The scanner still found plenty to work with, but the quality was mixed. There were 91 total setups and only 8 at A grade or better. That is enough to create opportunity, but not enough to justify force-feeding trades into a tape that was repricing around inflation data. In other words: motion was there, consistency was not. You can see live setups in the scanner when those higher-grade levels start to stack up.
Indexes bent, breadth stayed messy
The raw count looked active, but the resolution data was not great. Only 25 of 91 setups broke out, while 33 failed outright. That is not a momentum trader’s dream environment. It is a selective environment, which is exactly why grade discipline and backtest validation matter more than the setup count.
The right call was no call
Verdict: skip it.
There was no setup with grade A or better that also had a proven positive backtest edge across at least three historical trades. That rule exists for a reason, and today it kept the system out of weak evidence.
LRCX carried the top scanner grade at A++, but the backtest was bad: seven prior trades, 28.57% win rate, 0.4 profit factor. That is not edge, that is a trap wearing a nice grade. PYPL printed an A+ setup, but it only had one historical trade, which is nowhere near enough sample size to overrule the framework. The external backdrop did not help either. Recent analyst downgrades and softer sentiment around PayPal leaned bearish, so there was no reason to stretch for a long thesis just because the scanner printed a high score.
This is the kind of restraint that saves more capital than a forced trade ever makes. When grade and evidence do not line up, the only honest pick is a pass. As new levels build, see live setups in the scanner.
Light activity was the correct read
There was very little reason to trade aggressively into a CPI-driven tape with uneven follow-through. The systems stayed mostly restrained, which was the better outcome versus manufacturing action in a market that was not paying cleanly. You can watch Wex and Xcel trade live in the Edge Lab.
0 trades executed.
No issue here. A flat day is not inactivity for the sake of inactivity. It is a filter doing its job when the tape offers more chop than edge.
2 trades executed.
Manual close due to STC_FAILED orphan.
Translation: the trade did not become a real contributor, and the exit logic chose cleanup over hope. Flat is fine when execution friction shows up.
Exited via trailing floor after locking +12% from a +34% peak.
This one worked because the thesis matched the day. Financials were vulnerable in a hotter-inflation, policy-sensitive tape, and the exit engine did what it should do: let the position expand, then harvest what was left when momentum cooled.
The honest verdict: one meaningful winner, one cleanup trade, and zero overtrading. That is not flashy, but it is professional. Again, you can watch Wex and Xcel trade live in the Edge Lab.
Some money was left on the table, but not a fortune
The bots traded 2 breakouts and missed 9 that went on to produce a combined +4.86R. That number matters, but context matters more. There were no TP3 full runners, only one TP2, and just two TP1 hits. Most of the missed names were not clean trend monsters. They were partial payers in a noisy tape. That reduces the severity of the miss.
PDD B BEAR, +2.11R, TP2 hit
RVOL 0.35x, Day -2.93%
This is the biggest miss by R, but the low relative volume explains why it likely stayed filtered. Hard to call that a mistake when participation was thin.
XHB A- BEAR, +1.49R, TP1 hit
RVOL 2.36x, Day -2.28%
This is more interesting. Housing weakness matched the macro tape and volume confirmed it. This looks closer to a real blind spot than a healthy pass.
JPM A- BEAR, +1.26R, TP1 hit
RVOL 1.0x, Day +1.0%
The intraday breakout paid despite the stock finishing green on the day. That kind of path dependency is exactly why these are hard to capture consistently.
UPS A- BEAR, +0.92R, minor
RVOL 3.61x, Day -2.18%
Legitimate miss. Volume was there and the downside followed through enough to matter.
CRM A- BEAR, +0.77R, minor
RVOL 2.99x, Day -4.54%
Notable because it also showed up in the day’s breakout list. Clean enough to warrant attention, but still not a huge R capture after the trigger.
CVS A- BULL, +0.39R, minor
RVOL 0.35x, Day +3.18%
Low-volume bounce. Easy to justify skipping.
NET A- BEAR, +0.31R, minor
RVOL 0.48x, Day -3.21%
Another low-participation move. Filter probably did the right thing.
HAL A- BULL, +0.16R, minor
RVOL 1.0x, Day +4.41%
Too small to care about. This is noise, not missed edge.
Bottom line: XHB and UPS are the cleanest candidates for postmortem review. PDD paid the most, but the low RVOL makes that more acceptable as a pass than it looks at first glance.
Plenty of triggers, limited quality
The scoreboard was unimpressive. Failures outnumbered breakouts, which fits the tape. Inflation data created movement, but not broad trend reliability. Names that did work were mostly stock-specific or sector-aligned rather than part of a clean market-wide expansion.
Notable breakouts included AAPL bull A-, BLK bull A-, CAVA bear B+, CRM bear A-, CSCO bull A-, and CVNA bear A-. Mixed leadership across those names is the point. There was no single directional theme taking over the board. When that happens, selectivity beats activity. If you are tracking grade quality and breakout levels intraday, see live setups in the scanner.
The tape still trusts hard assets and selective tech
Gold
+10.98% vs SPY over 5d
Semis
+10.75% vs SPY over 5d
Metals
+3.51% vs SPY over 5d
Materials
+0.21% vs SPY over 5d
Biotech
-2.03% vs SPY over 5d
Housing
-2.78% vs SPY over 5d
Retail
-5.30% vs SPY over 5d
Energy
-6.71% vs SPY over 5d
Gold and semis remain the standout relative-strength groups over the last week, which is an odd but telling combination. One side is inflation and defensiveness, the other is concentrated leadership. Meanwhile housing, retail, and energy are lagging badly. Today’s action in XHB and XLF-backed bearish setups fits that regime.
Specific names, specific levels, no guesswork
These unresolved A and A- setups are still coiling into tomorrow. PPI at 8:30 ET means opening volatility can invalidate levels quickly, so treat triggers as conditional and wait for confirmation. For updated grading and live breakout levels, see live setups in the scanner.
Entry trigger: break below today’s bear trigger level once confirmed after the open.
Stop: above the opening reclaim high or the setup invalidation level, whichever is tighter.
TP1: first measured move at 1R.
Context: P score 68, distance 0.137%, RVOL 1.0x. Very close to trigger, which makes it actionable if software stays weak.
Entry trigger: break above today’s bull trigger level with confirmation.
Stop: below the opening hold level or setup invalidation.
TP1: first measured move at 1R.
Context: P score 68, distance 0.31%, RVOL 1.0x. Tight enough to matter, but only if