AI Resonance · 阅读最新日报 · AI 入门推荐

2026-11-14 · America/Los_Angeles · 社区动态 · #6

Should short-horizon price forecasting papers benchmark against market-implied forecasts, not just naive baselines? [D]

been reading a bunch of papers on short term crypto and commodity price forecasting (transformers, LSTMs, the usual). almost all of them compare against ARIMA, a random walk, maybe a simple LSTM, and report a better RMSE. what i rarely see is a comparison against what the market already expects. options implied vol for the range, or a crowd forecast where people put money on outcomes. those already use alot of public info, so beating a random walk doesnt say much if you cant beat the market's own forecast. also most papers only score the point forecast. for something like btc the interval…

Should short-horizon price forecasting papers benchmark against market-implied forecasts, not just naive baselines? [D]

热度 45.6 / 100;排名与评分保留该期记录。