Normally, trade signals are determined by the crossing of some indicators, the comparison of price levels, or the comparison of some indicator with a threshold value. All those are binary functions that do not consider the 'steepness' of the crossing or the distance to a threshold. My idea is to use fuzzy logic functions instead that return a value between 0..1. For filtering, the binary && could be replaced by a fuzzy 'and' function that returns a value like sqrt(a*b) or min(a,b) instead of a && b.
Theoretically this could remove some randomness from the trade signals and make strategies more robust. We plan to test this in the next time and, if successful, add fuzzy logic in one of the next Zorro versions.
Fuzzy logic for trade signals
-
- Trader
- Posts: 82
- Joined: Wed Oct 31, 2012 8:04 am
- Location: Frankfurt / Germany
- DING
- Trader
- Posts: 103
- Joined: Wed Nov 16, 2011 11:53 am
Re: Fuzzy logic for trade signals
fuzzy logic :!: , maybe plus neural fuzzy be more cool! I am expecting it
Re: Fuzzy logic for trade signals
I'll be interested to see what you find. The concept makes sense, but I don't know how well it would work out in practice... Too many of my ideas that "make sense" end up in the bin.