The Lab ran 4 experiments across 4 research threads this cycle (2026-08-10 → 2026-08-16) on the idle GPUs — every verdict below is published as the data said it, wins, losses and coin-flips alike. No cherry-picking.
Which crypto coins can a simple ML model actually predict?
- run of 2026-08-10 — 'TRUMP' leads on auc: +0.522
Confidence-gated trading sim
- meme pack, 5m bars, 3-bar horizon (fresh-data re-run) — POSSIBLE gated edge: short @ p>0.65 nets 109.4 bps/trade over 793 trades — needs leakage + robustness checks before anyone gets excited
Are calm or wild coins easier to predict?
- run of 2026-08-15 — 'high-vol' leads on mean_auc: +0.511
Does crypto direction get more predictable at longer horizons?
- run of 2026-08-16 — '15m' leads on mean_auc: +0.530
Auto-published weekly by Gordon Lab. Each entry is a real backtest or model run on real market data; methods and full result tables live in the lab archive. This is research, not financial advice.