"To be sure, many would argue that Quantumrock's impressive performance is a fluke and unlikely to be repeated consistently in the future" There you go!
They're $550M in size ? That's a bacteria compered to the big players. For that size - 34% isn't impressive.
Those figures are meaningless without knowing drawdown. “If you have a situation where the strategy is performing well just because it utilizes noise and not real patterns, we can detect those situations in advance.” Is that a long way to say the algo got backtested?
The old 2000 brokers tried to do this, 1 got lucky and there setting matched market condtions so they made a 1 off nice profit and ignore the other 1999 routine. Unless Machine Learning and AI suddenly become real things not just hype and are actually magic and can predict the future ofcourse. Nothing can predict the future better than the mk1 human eyeball.
I think it’s a waste of time to keep on predicting the future. Rely on market analysis and take decisions
Those who poo pah the general idea are incorrect. Billions have been made by funds going down this path. Medallion fund, one of the best, from Renaissance Technologies, has been automated for decades and they have been heavy into machine learning for many years.
Watching interviews of people involved with renaissance can give few good keywords about designing algo trading system...