Scanner Retrospective — May 29 – June 5, 2026
70 setups across 7 scans. 19 resolved unique signals (8 wins / 11 losses / 0 breakevens), ~30+ pending. Portfolio (optimal combo) period return +0.43%. Best pick AVGO (+12.48%, 05/27 — TP2 hit); worst HON (-5.0%, 06/03 — same-day stop). Period dominated by regime-label lag: score dropped to 41 while label stayed RISK-ON — 8 consecutive stop losses from Jun 1-3 scans. Profit factor 0.90 with zero breakevens — trailing stops not triggered before SL on most new entries.
Quick Dashboard
Resolved Trades — 19 Unique Signals
Period: May 29 to June 5, 2026 (7 scans). Only resolved signals shown — ~30+ pending trades excluded from this table.
| Scan | Ticker | Strategy | PnL | Status | Notes |
|---|---|---|---|---|---|
| 27/05 | AVGO | Momentum | +12.48% | TP2 ✔ | RISK-ON 85 · AI capex thesis · Full TP2 hit |
| 13/05 | AAPL | Breakout | +5.91% | Expired ✔ | RISK-ON · Mega-cap quality through regime oscillation |
| 22/05 | ARM | Breakout | +4.22% | Trail ✔ | RISK-ON 38 · AI breakout, quick +4% locked |
| 19/05 | MS | Momentum | +3.22% | Trail ✔ | RISK-ON 45 · Financials momentum confirmed |
| 13/05 | SLB | Momentum | +1.73% | Expired ✔ | RISK-ON · Energy momentum on long horizon |
| 27/05 | GE | Breakout | +1.71% | Trail ✔ | RISK-ON 85 · Industrials breakout captured |
| 27/05 | AMZN | Momentum | +1.06% | Trail ✔ | RISK-ON 85 · Mega-cap momentum, marginal |
| 22/05 | AAPL | Breakout | +0.14% | Trail ✔ | RISK-ON 38 · Second AAPL entry, breakeven-like |
| 03/06 | HON | Momentum | -5.00% | Stop ✘ | RISK-ON 41 (de facto NEUTRAL) · Same-day exit |
| 03/06 | HUBS | Momentum | -5.00% | Stop ✘ | RISK-ON 41 (de facto NEUTRAL) · Same-day exit |
| 02/06 | PLTR | Momentum | -4.83% | Stop ✘ | RISK-ON 47 (de facto NEUTRAL) · High-beta stopped in 2d |
| 03/06 | NVDA | Momentum | -4.00% | Stop ✘ | RISK-ON 41 · Semis Momentum in fading regime |
| 01/06 | AMZN | Momentum | -3.85% | Stop ✘ | RISK-ON 58 · Score still RISK-ON but fading |
| 01/06 | LLY | Momentum | -3.15% | Stop ✘ | RISK-ON 58 · Healthcare Momentum into headwind |
| 26/05 | NVDA | Momentum | -2.76% | Stop ✘ | RISK-ON 85 · Pre-regime-fade stop |
| 02/06 | SPOT | Breakout | -2.15% | Stop ✘ | RISK-ON 47 · Breakout failed in fading regime |
| 01/06 | GOOGL | Breakout | -2.04% | Stop ✘ | RISK-ON 58 · Breakout killed by fading score |
| 22/05 | DXCM | Breakout | -1.11% | Stop ✘ | RISK-ON 38 · Healthcare breakout stalled |
| 26/05 | SPOT | Momentum | -0.12% | Stop ✘ | RISK-ON 85 · Near-breakeven stop |
Strategy Analysis
Strategy observation — period highlight
The period reveals a stark regime-dependent performance split. Momentum's aggregate PnL is deeply negative (-4.49%) despite a nominal 36% WR — the 5 wins all came from older scans (May 13-27) while the 9 losses all came from Jun 1-3. Breakout's 60% WR and +6.54% aggregate PnL demonstrate resilience, but the sample is small (5 signals). The critical takeaway: when regime score is declining (even if the label hasn't changed), Momentum strategy should be paused in favor of Breakout or no-signal. A regime-score-based strategy gate — not label-based — would have avoided 8 of 11 losses.
Regime Analysis
Regime lesson of the period — Label lag is the #1 systemic risk
The regime score trajectory tells the full story: 56 → 58 → 47 → 41 → 87 → 36. The label stayed RISK-ON until Jun 5 when it finally flipped to NEUTRAL — by which point the Jun 1-3 scans had already caused 8 consecutive stop losses. Three critical takeaways: (1) Regime score is more actionable than regime label — a score below 50 should trigger defensive posture regardless of label. (2) Momentum strategy at score < 50 is catastrophic — 0/8 WR this period. (3) The label change mechanism needs a faster response — a 3-day lag turns a manageable drawdown into a cascade. Proposed fix: gate Momentum behind score > 50, regardless of label.
Top 3 Setups
AVGO — Broadcom (+12.48% TP2)
AVGO hit the full TP2 target — the best setup of the entire period by a wide margin. The AI capex thesis played out perfectly: Broadcom's custom AI accelerator revenue, combined with VMware integration synergies, drove the stock through both target levels. Strong volume confirmation + dark pool accumulation signals provided high-conviction entry. This is the archetype of what Momentum strategy delivers when regime conditions are genuinely supportive (score 85) — a setup that starts in strong conditions and rides a secular trend through to completion. The +12.48% return on a single setup is more than all 11 losses combined would take to offset at average sizing.
AAPL — Apple (+5.91% expired)
AAPL's breakout from May 13 held through the entire regime oscillation of the period — from RISK-ON 85 down to RISK-ON 41 and back up to RISK-ON 87 — without ever hitting its stop. This is the mega-cap quality shield at work: Apple's massive market cap, diversified revenue streams, and institutional holding base create a natural floor that absorbs regime volatility. The breakout strategy with a wide enough stop (not too tight) allowed the position to breathe through the turbulence. Expired at +5.91% on horizon — could have been more with a wider horizon, but capital preservation was the priority in this volatile period.
ARM — Arm Holdings (+4.22% trailing)
ARM appears in the Top 3 for the third consecutive retrospective — confirming that the AI semiconductor breakout thesis has persistent structural momentum. Entered during a volatile RISK-ON 38 regime on May 22, the trailing stop quickly captured +4.22% as the stock moved through its breakout zone. The trailing mechanism correctly locked profit before the Jun 1-3 regime degradation could erode gains. ARM's IP royalties + AI accelerator revenue model creates a fundamental floor that technical stops can't easily break. The consistency of ARM's appearances across retrospectives makes it a candidate for extended horizon tracking in future scans.
Flop 3 Setups
HON — Honeywell (-5.0% stopped, same-day exit)
HON is the poster child of the regime-label lag problem. Entered as a Momentum play on Jun 3 when the regime label was still RISK-ON but the score was 41 — deep in NEUTRAL/EARLY RISK-OFF territory. The stock hit its maximum stop loss on the same day of entry. A Momentum signal on an industrial cyclical at regime score 41 is structurally unsound: cyclicals need genuine RISK-ON conviction (score > 60) to sustain momentum. The 5% same-day loss represents maximum position damage. Actionable rule: block all Momentum signals when regime score < 50, regardless of label. This single rule would have prevented this trade entirely.
HUBS — HubSpot (-5.0% stopped, same-day exit)
Same pattern as HON — Momentum entry at regime score 41, same-day stop at max loss. HUBS as a high-growth SaaS name is particularly sensitive to risk sentiment: when institutional buyers step back (which they do at regime score < 50), high-beta growth stocks lose support immediately. The scanner identified HUBS's technical setup (above EMA20, volume expansion) but failed to weight the macro headwind. Two same-day exits from the same scan is a red flag for stop calibration: either the stops were too tight for the volatility level, or — more likely — the entries should never have been taken. The regime-score gate would have prevented both.
PLTR — Palantir (-4.83% stopped)
PLTR combines two risk factors: high-beta tech (volatility amplifier) and Momentum strategy in a fading regime (score 47). PLTR's beta of ~2.0x means any market weakness is amplified, making it especially vulnerable when regime confidence is declining. The stop at -4.83% was hit in just 2 days — rapid deterioration consistent with high-beta behavior in adverse conditions. This reinforces the same lesson as HON and HUBS: the regime score, not the label, should gate Momentum entries. PLTR would have been filtered by a score > 50 requirement. Additionally, high-beta names (beta > 1.5) should require an even higher score threshold — perhaps > 60 — to compensate for their amplified downside sensitivity.
Lessons & Improvements
What the scanner got right
- AVGO (+12.48%) hit full TP2 — the best single trade across all recent retrospectives
- Breakout strategy: 60% WR, +6.54% aggregate PnL — resilient even through regime fade
- Mega-cap quality picks (AAPL, ARM, MS) survived the regime oscillation
- Trailing stops locked profits on older positions before the Jun 1-3 cascade
- Zero breakevens — a marked improvement from last retro's 40% BE rate
- ARM in Top 3 for the 3rd consecutive retro — structural thesis validated
What the scanner missed
- Regime-label lag: 3-day delay between score decline and label change caused 8 consecutive SLs
- Jun 1-3 scans: 0% WR on 8 signals — all Momentum entries into a fading market
- Same-day exits (HON, HUBS on Jun 3) — stops may be too tight for high-vol days
- Profit factor 0.90 — below breakeven, first sub-1.0 PF in 5 retrospectives
- Momentum at score < 50: 0/8 WR — strategy should have been paused
- NVDA appeared twice (May 26 + Jun 3) — anti-doublon failed or position closed between entries
Lesson 1 — Regime score gates must override regime label for Momentum
The #1 lesson of this retrospective. The regime label stayed RISK-ON for 3 days after the score dropped below 50. All 8 Momentum signals from Jun 1-3 hit stop loss. Action: implement a hard gate: Momentum strategy requires regime score > 50, regardless of the current label. This is now severity=blocking. For high-beta names (beta > 1.5), the threshold should be score > 60.
Lesson 2 — Same-day exits indicate stop tightness relative to volatility
HON and HUBS both hit max stop on the day of entry (Jun 3). When 2 signals from the same scan are stopped intraday, it signals that ATR-based stops are too tight for the prevailing volatility regime. Action: when regime score is declining (2+ days of consecutive score decrease), widen stops by 0.5x ATR or skip the scan entirely. Advisory severity — to be validated next retro.
Lesson 3 — Breakout outperforms Momentum in regime transitions
Breakout: 60% WR, +6.54%. Momentum: 36% WR, -4.49%. In regime transition periods (score moving significantly between scans), Breakout is structurally more resilient because it relies on price clearing a defined level rather than sustained directional pressure. Action: when regime score volatility is high (standard deviation of last 5 scores > 15), increase Breakout allocation to 60-70% of scan signals. Advisory severity.
Lesson 4 — Older scans carry winners, newer scans carry risk
All 8 winners came from scans dated May 13-27. All 11 losses included 8 from Jun 1-3 + 3 from May 22-26. The pattern: signals entered during high-regime-score periods (85, 56, 45) had time to develop and trailing stops locked profit. Signals entered during declining-score periods (58, 47, 41) were stopped quickly. Action: the retrospective grade should more aggressively weight recent scan performance. Consider a 2:1 recency weight for the final 3 scans vs older scans.
Open Questions from Prior Retros — Resolution
Automation — data/scanner-lessons.json
Key updates to the machine-readable lesson file consumed by /scanner Phase 0.8:
- New blocking: Momentum strategy requires regime score > 50 (not just label = RISK-ON)
- New blocking: High-beta (beta > 1.5) Momentum requires regime score > 60
- New advisory: When regime score declining 2+ consecutive days, widen stops by 0.5x ATR
- New advisory: When regime score stdev > 15 over last 5 scans, allocate 60-70% to Breakout
- Closed: Q1 (RSI ceiling), Q3 (VWAP calibration), Q4 (BE persistence) — all resolved, not actionable
- New open questions: (Q5) Does the regime-score gate at 50 persist as optimal threshold? (Q6) Same-day exit rate — is ATR stop width correlated?
New patterns flagged for monitoring
- Regime-score-label lag: Score dropped to 41 while label stayed RISK-ON — caused 8 SL cascade (Jun 1-3). Now blocking rule.
- Fortress same-day exits: 2 same-day exits on Jun 3 (HUBS, HON) — stops may be too tight for the volatility at score < 50.
- Momentum 0% WR at score < 50: All 8 Momentum entries from scans with regime score < 50 hit stop loss. Zero exceptions.
Methodology note — Rolling system
The C+* grade reflects the 19 resolved trades of the period with a recency adjustment. The ~30+ pending trades will be reclassified in the next retro. The provisional marker (*) acknowledges that only ~38% of signals are resolved. The portfolio return (+0.43%) captures real performance agnostic of individual trade resolution timing. The recency-weight adjustment (C+ instead of B) reflects the 0% WR on the most recent 8 resolved signals — a forward-looking indicator of scanner health.
Pillar 2 — Portfolio Simulation
Mode used: Optimal combo (best risk-adjusted returns). Period return extracted from portfolio tracking data.
Grade History
| Period | Grade | Hit Rate | Portfolio Return | Key Lesson |
|---|---|---|---|---|
| Feb 20, 2026 | C+ | — | — | Early calibration period |
| Feb 28, 2026 | B+ | — | — | First strong performance week |
| Mar 6, 2026 | B- | — | — | Moderate performance |
| Mar 13, 2026 | B | — | — | Consistent mid-range |
| Mar 20, 2026 | C | — | — | Stops too tight in volatile regime |
| Mar 27, 2026 | B+ | — | — | Recovery + momentum alignment |
| Apr 3, 2026 | C | — | — | Regime oscillation hurts |
| Apr 10, 2026 | C* | — | — | Provisional, many pending |
| Apr 17, 2026 | A* | — | — | Best week — regime + setup alignment |
| Apr 25, 2026 | B* | 54% | +3.2% | Momentum > Breakout in sustained RISK-ON |
| May 2, 2026 | B+ | 52% | +2.9% | Pre-Squeeze underutilized in EARLY RISK-OFF |
| May 15, 2026 | C+ | 20% | N/A | Stops too tight, breakout penalized by rotation |
| May 22, 2026 | C | 16.7% | +1.32% | 4 rotations in 5 sessions |
| May 29, 2026 | B* | 40.0% | +1.22% | Energy + EARLY R-OFF toxic · 40% BE from trailing |
| Jun 6, 2026 | C+* | 42.1% | +0.43% | Regime-label lag · 8 SL cascade Jun 1-3 · PF 0.90 · Score gate needed |
Sources & Disclaimer
Market Data
- Signals:
scanner/YYYYMMDD/signals.json - Trades:
data/backtest-trades.json - Portfolio:
data/backtest-results.json - Live prices: Yahoo Finance via allorigins.win
- Regimes: DailyTickers MCP Gateway
Methodology
- Hit Rate: wins (PnL > 0) / resolved positions
- Portfolio: Optimal combo (best risk-adjusted)
- Grade: round((P1_score + P2_score) / 2) + recency adj.
- Rolling 10-day system — inflight reclassified next retro
- Next retrospective: Friday June 12, 2026