I cringe when I see so much goes to building the IT infrastructure and so little to modeling and validating algorithms. Someone points out that most quant funds

I cringe when I see so much goes to building the IT infrastructure and so little to modeling and validating algorithms. Someone points out that most quant funds fail. Key reason is their ML models are badly overfitted. By fitting enough models, you'd always find one with a good looking Sharpe Ratio. This is the classic multiple hypothesis testing problem. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3274354 https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3274354