Research & Community
Better ideas. Rigorously tested. Published every week.
Anyone can submit a hypothesis. We pick the strongest ones, run them through the same Koan Research Engine — predict, validate, attack, refine — and publish the full report every Friday.
Methodology, results, attacks, refinements, all of it.
Built for traders who think in probabilities, not promises.
01
Rigor over hype
Walk-forward validation, 10,000-permutation Monte Carlo stress, null-hypothesis testing — the same methods academic research papers require, on every test.
02
Inspectable by default
Every assumption, every step, every line of reasoning is visible. No black boxes.
03
Skeptical even of itself
Every prediction is calibrated. Every backtest is attacked. We're most suspicious of strategies that look easy.
04
Tested across regimes
A strategy that worked in 2021 might be a different strategy in 2024. Every test breaks down by regime — trending, ranging, volatile — so you see when an edge holds and when it dies.