You Can’t Always Get What you Quant
FT Tech Tonic
00:00
The Problem With Genetic Algorithms
Genealogicalogrithms are an approach where basically you get a whole bunch of computer entities. They all have parimeters only, i dont kno, may be a holding period. Thete might be a hundred thousand of these things, all designed to buy or sell base on certain set parameters. And their job is to sort of eat market data and produce money. These are busy trading algaritms that learn by themselves and propagate just like a living organism. If there's something new that hasn't happened before, ot hasn't happened in that data set, such as a global pandemic, then a i systems can get a little bit such.
Transcript
Play full episode
