Research notes
Short write-ups of what was tested, what failed and why. Negative results are published too: knowing what doesn't work is most of the job.
2026
- CR-003
The error the lab caught itself
How a valuation bug flattered early results, how it was found, and what changed afterwards.
Forthcoming
- CR-002
For a UK investor, the currency is the risk
Why a dollar trend portfolio has to be judged in pounds, and what hedging does to the drawdowns.
Forthcoming
- CR-001
Ten thousand trials: what failed, and why
What was tested, why most indicator-based ideas behave like coin flips, and what it takes for an idea to survive.
Forthcoming
Study logA selection, newest first
| Study | Question | Result |
|---|---|---|
| QL-0126 | When the Nasdaq moves first, do lagging stocks catch up in a tradeable way? | Rejected |
| QL-0125 | Are stop-losses hunted more often than chance would predict? | Rejected |
| QL-0124 | Does order-flow data predict the session beyond what price already says? | Rejected |
| QL-0118 | Does adding gold protect a portfolio in crashes, or does it only add gold's own return? | Rejected |
| QL-0084 | Does daily trend-following in futures work at a £10,000 account size? | Rejected |
| QL-0031 | Can a trend filter on index funds avoid the deep drawdowns of simply holding them? | Kept, live from November 2026 |
164 studies and more than 11,000 counted variants since October 2026. Almost all were rejected.