My firm (an asset manager) experimented with ensemble techniques a good five years ago or more. We spent about 200-250 hours (in aggregate, across many people) — most of it programmers, but also quants and portfolio analytics. There was a significant cost in time and money. We do 15-20 such projects every year, let’s call it research and development. We don’t publicize the results outside our firm.

We learn what works and what doesn’t. Some of it can be used for our business, a lot of it just teaches us what not to do. We don’t share this information with our competitors for free.

N.B. CFA institute charges its members an annual fee — despite being registered as a charity — it isn’t giving away information for free either. But it tends to publish information only after it becomes mainstream. It doesn’t publish cutting edge research