I’ve been experimenting with deep‑learning models to find leading indicators for the Nasdaq‑100 (NQ). Over the past month the approach delivered a 32 % portfolio gain, which I’m treating as a lucky outlier until the data says otherwise. I selected the following crypto/Future/ETF/Stock (46 tickers) to train the model: ADA‑USD, BNB‑USD, BOIL, BTC‑USD, CL=F, CNY=X, DOGE‑USD, DRIP, ETH‑USD, EUR=X, EWT, FAS, GBTC, GC=F, GLD, GOLD, HG=F, HKD=X, IJR, IWF, MSTR, NG=F, NQ=F, PAXG‑USD, QQQ, SI=F, SLV, SOL‑USD, SOXL, SPY, TLT, TWD=X, UB=F, UCO, UDOW, USO, XRP‑USD, YINN, YM=F, ZN=F, ^FVX, ^SOX, ^TNX, ^TWII, ^TYX, ^VIX.

I collected data started from 2017/11/10 for building feature matrix. I’ve shared the real-time results in this Google Sheet: https://ai2x.co/ai

• Columns R–V show the various indicators.

• Row 2 contains each indicator’s correlation with NQ on a one‑hour look‑ahead basis.

Feedback, alternative metrics, or data sources are very welcome!