⚠️ Personal research and trading journal — not investment advice. The author does not provide licensed advisory services.
Every scanner runs on hard thresholds. A stock either qualifies or it doesn't. That binary logic is what makes scanners fast and consistent — and it's also what creates systematic blind spots.
After building and running my breakout scanner for months, I identified three categories of setups the scanner is structurally incapable of finding. Not because the setups are bad — in some cases they're the best setups in the list — but because they don't cross the numerical thresholds I chose.
I built a secondary tool, the "near-miss tracker," specifically to surface these. Here's what it finds.
Blind Spot 1: RS 74-79
My primary scanner requires RS Rating ≥80. The RS filter is one of the most validated elements in my entire methodology — it passed bootstrap walk-forward confidence intervals that excluded zero, something most filters I've tested don't do.
But what about stocks with RS 74-79?
These stocks are in the top quartile of the market by relative strength. They're outperforming 74-79% of all other stocks over the trailing 12 months. In most markets, that's strong. My scanner doesn't see them.
The near-miss tracker pulls candidates in the RS 74-79 range that otherwise would qualify on every other criterion: Stage 2 uptrend, contracting base, volume dry-up. It surfaces them separately, not as primary candidates, but as context.
Why it matters: sometimes a stock is building RS momentum and passes through 74 → 77 → 81 over several weeks. If I'm only looking at ≥80, I don't see the trajectory — I just suddenly see a new name appear on the scanner with no prior context. The near-miss list shows me the stock was already on my radar two weeks earlier, with RS 76, contracting nicely.
The near-miss tracker doesn't override the RS≥80 hard gate. It just shows what's warming up nearby.
Blind Spot 2: Bases 8-14% Below the Prior High
A classic breakout scanner looks for stocks near their pivot — within 5% of a prior high, typically. This catches stocks ready to break.
But what about stocks that are contracting beautifully, showing higher lows, with volume drying up — but still 10% below the prior high?
The base is not complete yet. The stock isn't "near the pivot" because the pivot hasn't been established. It's building the base that will create the pivot.
My scanner can't find these. Its logic is: stock near high + volume pop = breakout candidate. A stock still 10% below its high doesn't show up as "near a breakout." But it may be setting up the best risk/reward of its current base, because:
1. The stop (prior low) is clearly defined 2. The base is still contracting (higher lows are being printed) 3. Volume is quieter than usual, signaling lack of distribution
A GUNKUL setup from 2026 illustrated this. The stock was building a 5-week cup — valid base by the definition I use — but was 9% below the prior high when the scanner would have first noticed it. The near-miss tracker had it for two weeks before it crossed into scanner range.
Blind Spot 3: Bases That Are Contracting But Not Yet Near a Recovery High
Related to blind spot 2, but subtly different: these are stocks in early contraction.
The breakout methodology I use defines a valid base as contracting higher lows moving toward a recovery high (the second peak after a pullback). But the early-stage of a base looks like: first low, first bounce, second pullback — lower high. The stock hasn't shown a recovery high yet.
If I filter for "stocks near prior highs," I miss the early-base stocks by definition. If I filter for "higher lows," I might find them — but still can't know if the contraction will hold.
The near-miss tracker tags stocks showing the trajectory without requiring the resolution: base is narrowing (last 3 pullback lows are each higher), RS is rising, stock is above its 50-day moving average. These don't get flagged as primary candidates. They get flagged as "watch — base building, not yet complete."
Why Hard Thresholds Are Still Right
I built the near-miss tracker as a complement, not a replacement for the main scanner. The hard thresholds in the scanner exist for a reason: they eliminate ambiguity. If I pass everything near-miss candidates to the primary entry logic, I dilute the edge I've validated.
The scanner is conservative by design. Rigid gates are what make it consistent. The research (bootstrap walk-forward on 20+ years of SET data) was done on the hard-gated version. Weakening the gates means I'm running a different system than what was tested.
The near-miss tracker is for chart reading context — not for adding candidates to the trading queue. When I see a near-miss stock, I add it to my watchlist as something to monitor. I don't trade it until it crosses the hard threshold. The tracker answers "what are the interesting things developing nearby?" not "what should I buy now?"
The Setup That Taught Me This
GUNKUL was in the near-miss bucket for two weeks before the main scanner picked it up. By the time the scanner triggered at RS≥80 within 5% of the pivot, the optimal entry — based on the base structure — had been clear for several days.
That's not a failure of the scanner. That's what scanners do: they confirm, they don't anticipate. The chart reader anticipates. The scanner confirms when the setup is hard to miss.
The distinction is important. If I'd taken the near-miss trigger as an entry signal for GUNKUL, I'd have entered 2 weeks early, with less defined risk, and I'd be running a different strategy than the one I've tested. The value of the near-miss list is awareness, not action.
Track. Study. Wait. Strike.
Personal research and trading journal — not investment advice. The author does not provide licensed advisory services. — MOEasymmetry
Draft 2026-06-12. Source: MOEasymmetry near-miss tracker (scan_near_miss.py → near_miss.html). The 3 blind spots are structural artifacts of the scanner thresholds — RS≥80 gate, within-5%-of-pivot filter, and recovery-high requirement. GUNKUL example from 2026 (approximate). Near-miss tracker built as watchlist supplement, not as a primary signal tool.