I Backtested 58 Intel Insider Buys Since 2019. The Results Were Interesting
So I've spent a stupid amount of time digging through SEC Edgar filings and backtesting insider transactions against actual price outcomes.
Let me share what the data says specifically about Intel ($INTC), because the numbers are genuinely interesting, not in a Reddit hype way, but in a "this is statistically weird and probably worth understanding" way.
**The Setup**
Between 2019 and early 2026, Intel executives and directors made 58 open-market purchases of their own stock, meaning they used personal cash, not options, not grants. Just bought shares like anyone else would.
When that happens, what does the stock do?
* Day 1 median return: +0.63%
* Day 3 median return: +1.53%
* Day 5 median return: +2.59%
* Day 20 median return: +2.16%
That last number is what I find most interesting. The drift doesn't continue forever, it actually fades slightly after Day 5. So, whatever signal insider buying creates, it plays out fast. You're looking at a roughly 5-day edge, then the market absorbs it and moves on.
The 5-day win rate across all 58 cases: **69%**. For a large-cap stock, that's a statistically notable number. It's not 90%. It's not a magic formula. But 69% is meaningfully above coin-flip territory.
**What Happens When Insiders Buy Into a Falling Stock**
This is where it gets more interesting.
Of those 58 buys, 27 happened when Intel had already dropped more than 2% in the 5 days before the filing. The executive bought *into* weakness, not strength.
In those specific cases, the 5-day median return expanded to **+3.27%.** an extra 0.68% over the baseline.
The interpretation isn't complicated: when someone with deep knowledge of the business buys the dip with their own money, it's a stronger signal than buying into a flat or rising tape. The "floor" behavior is more pronounced.
**Real Examples From the Data (Both Wins and Losses)**
**January 26, 2026:** Intel's CFO David Zinsner bought \~$250k worth of shares at $42.50. The stock returned +12.1% over the next 5 days. Huge. Clean signal.
**August 5, 2024:** CEO Pat Gelsinger made two separate purchases on the same day ($152k and $99k). Cluster buys from a CEO tend to carry more conviction than a single order. 5-day return: +4.89%.
**May 1, 2024 and April 29, 2024:** Also Gelsinger buys. Stock drifted slightly negative (\~-0.8%). These are the failures that make the 69% realistic. Any backtest that doesn't include the losses is selling you something.
**Why Most People Ignore This Data**
The main reason is that it's annoying to collect. SEC EDGAR filings aren't cleanly structured, the dates don't always align cleanly with trading days, and you have to manually filter out option exercises, gift transfers, and 10b5-1 plan purchases from genuine open-market buys. The noise is high.
When you do filter it properly and align it with actual OHLC data, though, you get a surprisingly coherent picture, at least for a stock like Intel, where insider transactions are frequent enough to build a meaningful sample.
This is one ticker, and 58 data points isn't a massive sample. I'm not saying this works universally or that it's some edge case Wall Street is sleeping on. Smart money already watches this, the question is whether retail traders have the same quality of signal or are always reacting a few days late.
Curious if anyone else has done similar backtesting on insider filings for other large-caps, or if there's a sector where this signal tends to be stronger or weaker. Energy and semiconductors seem like obvious candidates to me, but I haven't gone deep on them yet.
And honestly, if a system existed that tracked this automatically, surfaced the weak vs. strong setups, and delivered the historical context upfront instead of requiring you to dig for it yourself, would that be something you'd actually want to use? Seems like a gap that's always existed but has never been filled well for regular investors.
[](https://www.reddit.com/submit/?source_id=t3_1thg49u&composer_entry=crosspost_prompt)