Paper — every trading figure here (positions, orders, P&L, equity, returns) is simulated on an Alpaca paper account. No real capital is at risk, and no order is ever placed with your money. Running costs shown under ops, such as LLM spend, are real.

Going deeper

Public

How the Research Engine Vets Opportunities

Before anything trades, an idea is passed through a research funnel. The goal is simple: by the time a strategy raises a trade, the system is as confident as it can be that the idea is sound — and well timed.

The one question that matters

When something has fallen a long way, there are two very different reasons it could be cheap — and telling them apart is the whole job:

  • ·A value dislocation — a sound sector knocked down by a passing shock. The business is fine; the price over-reacted. This is the kind of cheap worth buying, because it tends to recover.
  • ·A value trap— something cheap because it deserves to be. The decline reflects a real, lasting problem. It looks like a bargain, but there's no spring in it. This is the kind of cheap to avoid.

Why this is the keystone

A purely technical "it's oversold" reading can't tell these two apart — both look identical on a price chart. The research funnel adds the missing lens: it asks whether the thing is cheap-and-sound or cheap-and-broken before it ever lets a timing signal matter.

The funnel, lens by lens

An idea has to earn its way through each lens in turn. A weak answer at any stage filters it out — the system would rather pass than force a low-quality trade.

  • ·1. The macro & cycle lens— first, what's the weather? The system reads where the broad economy sits in its cycle and whether the backdrop favours a given sector. Some sectors do well early in a recovery, others late — buying into a tailwind beats fighting a headwind.
  • ·2. The soundness lens (cheap, or broken?)— next, the value-trap check. Using the sector's underlying health, the system asks whether a sell-off looks like a sound asset on sale or a genuine deterioration. Only the cheap-and-sound ideas move on; the value traps are set aside.
  • ·3. The dislocation lens (how oversold?) — for what survives, how stretched is the price? A bigger sell-off in something fundamentally sound means a bigger potential snap-back. This measures the size of the opportunity — but only after soundness is established, never before.
  • ·4. The timing lens— finally, is now the moment? Being cheap and sound isn't enough; the price has to actually start turning up. The system waits for that confirmation rather than trying to catch the exact bottom of a falling price.

The doctrine

Macro tailwind, then cheap-and-sound, then how-oversold, then timing — in that order. By the time an idea reaches a strategy as a tradeable candidate, every one of those questions has already been answered yes. That is what "maximally confident, well timed" means in practice.

What this does and doesn't promise

A more careful research funnel raises conviction — it does not guarantee an outcome. It filters forsound, well-timed ideas; it can't manufacture an edge where none exists. That's why every approach is still held to out-of-sample proof before it's trusted with money, and re-checked over time.

See How Strategies Are Proven for how an approach earns its place, and Reading the Signals for why a vetted idea still might not become a trade.

Where you see this in the app

The Positioning page is the radar where these ideas surface — beaten-down sectors with a real reason they could recover, tracked until the timing confirms. The Insights page is where you can ask the system why it sees an opportunity the way it does.