← Back to Blog

general

Which crypto coins can a simple ML model actually predict?

Auto-generated by Gordon Lab from a real backtest — published as a draft for review, not live.

Which crypto coins can a simple ML model actually predict?

The question: We trained XGBoost on 25 coins to predict 15-minute direction out-of-sample on the P100. How many actually beat a coin-flip?

What the data said

'ARB' leads on auc: +0.532

Full results

group metric
ARB 0.5321
BONK 0.5225
CRV 0.519
BTC 0.5148
HYPE 0.514
ETH 0.5137
LINK 0.51
XRP 0.5094
TRUMP 0.5087
SHIB 0.5071
UNI 0.5066
GRT 0.5064
PEPE 0.5057
AVAX 0.5051
DOGE 0.5047

Generated by Gordon Lab from a real backtest. Method + caveats live in the harness; this is research, not advice — and the losing runs are shown too.