Every experiment on Factor Lab passed a strict statistical gate — deflated Sharpe, FDR adjustment, permutation tests, cost realism. Save the ones you like. Fork them into your Build workspace. It all syncs to the iOS app the moment you log in.
Every card is a finished, validated experiment with a hypothesis, parameters, and evidence behind it. Save what catches your eye — fork what you want to iterate on.
Each experiment ships with its full payload — the hypothesis it tested, the parameters and factor IDs, the symbol and timeframe, the metrics, and a complete evidence panel. Fork it and every byte lands in your Build workspace on iOS.
fork preserves all of it.After realized vol regime expansion, ETH 1h carries persistent directional momentum net of funding. Null: post-expansion forward returns are zero in expectation.
factor_ids: ["ret_1h", "realized_vol_24h", "funding_8h"] method_id: "vol_scaled_momentum_v2" params: { lookback: 24, vol_target: 0.35, decay: 0.92 } symbol: "ETHUSDT" market: "um_futures" interval: "1h" exit: { stop_loss_pct: 0.04, max_hold_bars: 48 } cost_model: { taker_bps: 5, slip_base: 1, slip_k: 25 }
Every action you take here is account-bound. Sign in once — the same saves and forks show up in your iOS app the moment you open it.
Me · Saved on iOS within seconds. Organize with collections, sort by your own rating.
Build · Drafts with the lineage intact.
JSON is the full payload (re-importable). CSV is the trade ledger plus per-bar PnL. Report is a shareable HTML summary with credibility headers.
Most platforms show you a backtest curve and a Sharpe. We show those last. Every experiment fronts the statistical caveats — not because the result is bad, but because the gate is real.
FDR-adjusted p-values and Deflated Sharpe Ratio handle the multiple-testing tax. PBO bounds overfit risk. Nothing passes by luck of the draw.
IC and rank-IC across forecast horizons, signal decay profile, regime-conditional consistency. We want the edge to make sense before we trust the curve.
Taker fees, slippage that scales with order size, funding costs on futures, capacity estimate. If your fork uses different costs, the gate re-evaluates.
You browse the archive. This is the autonomous research program that fills it — its live status, what has passed the gate, and the most recently archived survivors. Read-only. No knobs.
The research program runs server-side on a fixed schedule. The website is read-only by design — you browse and fork the survivors, you never tune the search. Every survivor carries a reproducible manifest.
Factor Lab is the source of validated raw material. Build is where you bend it — constraints, your costs, your sizing. Both sides talk.
Hypotheses generated, optimized, gated, archived. New cohorts published weekly. Every experiment is reproducible from its manifest.
Forks arrive in Build with the full payload. Change parameters, re-run with your costs, publish back as your own variant — or push to Monitor as a watch-only signal.
Browse without signing in. When you find one you like, sign in and it follows you to your phone.