How Outplayed's AI Signals Read the Market
A plain-English look at what our models actually measure — volume, spread, float distribution and momentum — and what they deliberately ignore.
By Outplayed Research
"AI signal" is doing a lot of work in a lot of places these days, so it's worth being precise about what ours actually does — and, just as importantly, what it doesn't.
What the models measure
Our signals blend a handful of observable, auditable inputs:
- Volume trend — is real trading activity rising or fading?
- Bid-ask spread — tightening spreads mean conviction; widening spreads mean uncertainty.
- Float distribution — where the liquid supply sits within a wear bracket.
- Momentum and mean-reversion — short-term price behaviour relative to the item's own history.
None of these are exotic. The value isn't a secret indicator — it's measuring the boring things consistently, across thousands of items, without getting bored or emotional.
What they deliberately ignore
A good signal is as much about what it refuses to react to as what it reacts to.
We don't trade on hype keywords, streamer mentions, or social sentiment spikes. Those are real forces, but they're noisy, reflexive, and impossible to act on without front-running yourself. The models stay on observable market structure.
How to use a signal (and how not to)
A signal is an input, not an instruction. Treat it like a second opinion:
- Use it to surface items worth a closer look.
- Confirm with your own read on float, pattern and liquidity.
- Size the position by your conviction, not the signal's score.
The fastest way to lose money with any signal is to outsource judgement to it. Use it to widen your funnel, then do the work.
Where to see them
Signals live on the dashboard with the underlying metrics exposed, because a score you can't interrogate is a score you shouldn't trust.