⚠️ Personal research and trading journal — not investment advice. The author does not provide licensed advisory services.
IBD's buy mentions do predict a run-up. That part is real. Across 77 ACTIONABLE_BUY + PATTERN_BREAKOUT_FRESH signals from 2019-2026, the mean maximum run-up was +23.2% (median +14.6%). 91% of signals had a positive maximum run-up at some point.
But here's the problem: mean 120-day returns turn negative.
UNH: -49% from peak. SMCI: -45%. The average signal that looked brilliant at +20% was giving it all back and more if you held to 120 days.
This is the hold-and-pray failure mode. And it's the reason partial-TP is mandatory in my system.
The Full Picture
I analyzed the complete distribution of outcomes for signals that hit +20% or more on maximum run-up:
- 36% of signals hit ≥+20% maximum run-up at some point (vs 22.5% baseline = +13.5pp edge)
- Mean t+120d return: negative across the full signal set
- Tail risk asymmetric: right tail (NVDA-type runs) exists but left tail (extended reversals) dominates the average
The signals are real. The timing is the problem. IBD mentions a stock that has already begun breaking out. The window for capturing the move is limited. Miss it, and you're now holding a stock that may be entering distribution phase.
Why This Happens
When a stock gets featured in IBD video with ACTIONABLE_BUY language, it has typically already broken out, is at or near new highs, and has attracted the audience that generates additional buying pressure. That buying creates the run-up. The 91% positive max run-up number reflects that mechanical tailwind.
But stocks that attract that level of spotlight attention tend to also attract aggressive selling as they extend. Institutional investors who bought earlier take profits. The stock digests the move. Often, the digestion is violent.
The specific stocks with the worst outcomes (UNH, SMCI) weren't random: they combined IBD featured status with fundamental deterioration that wasn't visible at the time of the mention. By 120 days, the story had changed. Long-term holders absorbed the entire reversal.
What the Fix Is
Partial-TP at first target (typically 20-25%) and a trailing exit on EMA50 or MA21 captures the bulk of the run-up while avoiding the reversal.
The analysis validates this mathematically: the 36% of signals hitting ≥+20% at peak is the capture opportunity. The fix isn't finding better signals — it's not giving the run-up back.
In my current system, partial-TP is structured as: - Sell half at 2× initial risk (approximately 10-14% from a typical entry) - Trail the remainder on MA21 weekly - Hard exit if stock drops back below breakout level on volume
The alternative I'm testing (IBD canonical sell rule) is: 20-25% partial + EMA50 trail (vs the current 2R + MA21). These are in A/B testing for the post-freeze comparison. The principle in both cases is the same: capture, don't hold and pray.
The Broader Lesson
This is a testable version of a common trading maxim: "take your profits." But the data makes it concrete. You're not leaving money on the table by taking the partial at +20%. You're protecting yourself from a population of outcomes where stocks that hit +20% have a measurable probability of ending at -30% by 120 days.
The IBD buy signals work. The exits matter more than the entries.
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: n=77 ACTIONABLE_BUY + PATTERN_BREAKOUT_FRESH signals, IBD corpus 2019-2026. Mean max run-up +23.2%, median +14.6%. 36% hit ≥20% vs 22.5% baseline = +13.5pp. T+120d mean: negative. Named stocks (UNH -49%, SMCI -45%) represent tail outcomes, not median. Validated in feedback_ibd_buy_runup_analysis_2026-05-15.md. Partial-TP at 2R+MA21 is locked config; 20-25%+EMA50 is A/B candidate per feedback_ibd_sell_rule_20_25_ema50.md.