commodity price forecasting

Why price forecasting papers should benchmark against market expectations

Most price forecasting papers compare against ARIMA or a random walk and call it a day. That misses the point. Markets already have forecasts built into options-implied volatility and prediction markets where people put real money on outcomes. Beating a random walk is a low bar. For assets like Bitcoin, interval coverage matters more than RMSE. An 80% confidence interval that actually holds 80% of the time tells you more. Our piece on time-series AI models stumbling against simple baselines explores this gap further.