I Made My Lorentzian Model Right More Often. It Got Poorer.

wiseapple.com

Build an indicator no one else has: wiseapple.com/app

I tried my own idea on my own version of the Lorentzian Classification model: feature decomposition, a setting that squeezes many features into a few strong signals. The win rate went up. The profit went down. Each result is the same 81 markets, tested one at a time.
The whole experiment: “I Tested a 1950s Math Trap on Lorentzian Classification”
All trades shown are simulated. Not financial advice.