System record · research & decision intelligence
Quant is where we pressure-test our own decisions before they cost us. Two systems live today: R22, the trading system we built and run, and the Blindspot Engine - which reads everything we've researched and tried, then finds what we're missing.
01 / Two systems, one method
Demand makes the market see you. Quant makes sure the thing they see is actually right. Both start from the same habit: every claim we make has to survive a check.
Trader: the product
A gold trend-following system we built, validated and run. The point isn't the trade: it's that every decision it makes sits on a documented chain of research, backtest and verdict.
The Blindspot Engine: the concept
It reads everything we've ever researched, tried and concluded, then looks for what we're not seeing: and the connections we never made. Trading was the first place we ran it. The method isn't stuck to trading.
02 / The concept
Think of it as a detective who reads your whole diary before saying anything. It can't guarantee the right answer: it guarantees the questions you'd never think to ask get asked, with sources attached.
Decisions, research notes, session logs and live systems are linked so the engine can see the chain behind any claim: "this rule passed because that signal was strong: and that signal is a broker-quote proxy, not true order flow." Before, that warning and that pass never touched. Now the engine can connect them.
Instead of "is this ready?", it asks the question the other way: "this already failed: how?" It hunts the dramatic failure and the boring one, and it refuses to re-raise anything we've already tried or ruled out. What remains is ranked, so the strongest questions come first.
Each finding is kept with its source and its date - 65 findings across 7 runs so far. What we ruled out is kept too, so we never waste time re-arguing a closed question. The output is a decision you can make and stand behind.
Honest about what this is: Quant is our own research operation, shown openly. It is not a promise of trading returns and it is not client-revenue proof. We won't dress a trading system up as business advice: the offer is the method, and it earns its name only after it's proven on a non-trading problem too.
03 / One method, every decision
The engine doesn't care what the subject is. Trading was the first and hardest test. The same process: read the history, ask the question backwards, keep the sources: works anywhere a wrong assumption is expensive.
Applied today
Runs on our trading research: hidden dependencies, decayed edges, factors whose meaning was never questioned, alternatives we never tested.
Proven · liveNext: the natural fit
Before the outbound engine spends a quarter on a client's offer, run it backwards: "this campaign already failed: how?" The findings feed the qualification spec we write on day one.
First non-trading pilotInternal
A blindspot pass across a live cold-email system: domain health, deliverability assumptions, claims that drifted from evidence: before they cost a reputation.
Internal use caseThe Outbound Machine link
The content service already blocks claims that drift from facts. A blindspot pass goes the other way: finding the strongest proof you're not publishing yet.
Future capability04 / Start
Tell us what you're deciding on. If we take it on, the first run is the proof: findings with sources attached, before you pay for a second.
Book a 30-minute Google Meet - bring the decision you're stuck on.
Questions Where this runsInside the Lush Associates stack: the same machine room that books meetings on the Demand Engine side of this site.