⚠️ Personal research and trading journal — not investment advice. The author does not provide licensed advisory services.
When I built the IBD transcript corpus — 4,800+ videos, a decade of daily commentary — I wasn't just looking for stock-level signals. I wanted to know whether IBD's mentions predicted anything at the sector level.
They do. But not in the way I expected.
What I Found
I analyzed IBD mentions by sector across the full 2016-2026 corpus, then cross-referenced with 30-day forward sector returns. The pattern:
Top performers (sectors where IBD buy-mentions led to strongest 30d forward returns): 1. Electronic Technology (semiconductors, hardware, software infrastructure) 2. Technology Services (cloud, cybersecurity, data platforms) 3. Medical/Health Technology
Bottom performers (sectors where buy-mentions led to weakest or negative 30d forward returns): 1. Consumer Services (retail services, travel, food) 2. Finance (traditional banks, insurance) 3. Utilities / Energy (mentioned mostly for defensive positioning)
The spread between top and bottom sectors was material — approximately 3-5 percentage points on 30-day median forward return when comparing ACTIONABLE_BUY mentions in Electronic Tech vs. Consumer Services.
The Mechanism
This isn't hard to explain. IBD commentary reflects its own methodology: it covers stocks that are already showing relative strength, volume confirmation, and earnings growth. Those conditions cluster in specific sectors depending on the macro cycle.
Electronic Technology tends to dominate when earnings growth is concentrated in tech (most of the period I studied), when institutional money is rotating toward growth, and when the broad market is in confirmed uptrend. Consumer Services has structurally different characteristics — more mean-reverting behavior, lower institutional ownership concentrations, and earnings that correlate less with market momentum.
IBD mentions don't create this pattern. They reflect it. When a stock gets an ACTIONABLE_BUY or PATTERN_BREAKOUT_FRESH mention in IBD commentary, the sector context tells you something about the tail risk of that setup.
What Didn't Work
The full IBD high-conviction filter I tested earlier — combining mention frequency, buy language, and sector — lagged a simple RS≥80 baseline as a standalone direction-signal. The problem: high-conviction IBD mention language is an AFTER-the-move signal. By the time a stock is featured in daily video with multiple buy-language anchors, it's often extended.
This was falsified explicitly: #WON1 IBD high-conviction filter UPDATED 2026-05-14 — lags RS≥80 baseline by a material margin. A candidate-seed for further study, never a direction signal.
The sector pattern survived that falsification. Individual stock timing from IBD language is noisy; sector allocation from the sector composition of what IBD features is cleaner.
How I Use This Now
I don't filter setups out based on sector — that kills too many valid candidates with exceptions (GUNKUL, AMATA, and other Thai examples don't fit neat US sector boxes). But I use sector context as a tiebreaker:
When two setups look equally well-formed, I'll prioritize the one in a sector where the macro wind is at its back. If the overall IBD commentary in a given week is heavily skewed toward Electronic Tech mentions, that's relevant context for a semiconductor setup I'm evaluating — it tells me I'm not fighting the rotation.
For Thai stocks, the sector map is different (Materials and Industrials carry more weight than in the US; Consumer Services behaves differently given the tourism and export exposure), so I don't apply the US sector ranking mechanically. The principle transfers; the specific rankings don't.
The Broader Lesson
When you build a signal corpus, the aggregate information at the sector level is often more stable than the individual stock level. Individual IBD mentions are event-driven, time-sensitive, and pick-specific. The sector composition of what gets mentioned in confirmed uptrends — that's a slower-moving signal with more persistence.
If you're going to mine any analyst or commentator corpus for trading signals, start by asking: does what they cover cluster by sector? Does that clustering have predictive content? The answer is often yes, and it's cleaner than the individual call.
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: 4,800+ IBD transcript corpus 2016-2026, phrase-first extraction methodology. #WON1 finding: high-conviction IBD filter lags RS≥80 (see feedback_ibd_high_conviction_filter.md). Sector patterns directional across full corpus; precise spread statistics in backtest data. Thai sector rankings not derived from this corpus — US corpus only. Finding: SECTOR PATTERN holds; individual timing signal does not.