Signals are overlays, not bets
Most of the signals people get excited about — insiders buying, a buyback announced, short interest spiking — are real but short-lived. They nudge returns for a few weeks, then fade. That makes them genuinely useful, but useful in a specific way: they're good at telling you when to act and how hard, and bad at being the reason to own something for a year. We treat them accordingly.
What it is
We split the things we know about a company into two kinds. Durable strands — business quality, valuation, capital allocation, management credibility — are the reasons to own it at all; they're what the multi-month hold actually rests on. Conviction and timing signals — insider clusters, a buyback kicking in, a sentiment extreme — are overlays: they sharpen the entry and inform the size, but they don't carry the position by themselves.
Why it matters
The evidence keeps saying the same thing. Take our own insider-cluster study: measured properly, clusters of insiders buying earn a real ~1-month abnormal return and a modest 3-month one — then nothing. As a standalone "buy and hold for a year" rule it does little. As an overlay on a name we already wanted to own for durable reasons, it's exactly what it should be: a well-timed, higher-conviction entry. Treating a 1-month signal as a 12-month thesis is how people lose money with a statistic that was actually true.
How we use it
- A signal may only be load-bearing if its own evidence supports the horizon we're holding for. Short-horizon signals shape entry and size; they don't justify the hold.
- The hold is justified by the durable strands. The overlays then stack on top — the more independent ones that agree, the more conviction (see confluence).
- Where a signal's edge lives in places we can't trade well (tiny, illiquid names, for instance), we note it and down-weight it for our universe rather than pretending it transfers.
The honest caveat
"Overlay" is not a licence to use weak signals freely. An overlay still has to have earned its place with evidence — it's just being asked to do a smaller, honest job. A pile of fading signals is not a thesis; it's a reason to look more carefully at the durable strands underneath.
A plain-English explainer of how we think — part of our evidence-driven framework. Not investment advice.