Having worked for a 700M$ quant investment fund, I'd say, uhem no. It's pretty simple: algo investing doesn't work well or at all. algo trading does. nobody publishes very good algorithms. there are perhaps only 20 very good systems in the world. usually they are statistical arbitrage or HF. these kinds of systems are measured in sharpe ratio and turnover at least once a week. although I've spoken to a 25 year old who managed to make a killing auto-trading an emerging market, but these things don't happen at scale.<p>also, Bloomberg is in completely different business.<p>if somebody really wants to work on replacing Bloomberg let me know. this is a huge market, which nobody is working on AFAIK (=~Angellist).
Revenue per user might be high, but it's targeting a smaller market. Bloomberg is useful to anyone who buys or sells stocks in any way. Quantopian seems to be about people who want to find a technical system to beat the market. This is by its nature self-defeating. If everyone is using Quantopian, everyone cannot be beating the market. The article makes this very point on page two.
I've been lurking on Quantopian for a few months. The algotrading community is secretive. Currently, you code in python which is then saved and run on their servers. I fear Quantopian copying the best trading strategies for their own use. Even if you black box the code, all Quantopian needs is I/O with your program. What's to stop this?
I have a fair bit of insight into this world as I work in it and this quote seems abit off.<p>> Now, 5% of asset managers are algorithmic traders. In a decade, 50% to 70% of asset managers will use trading algorithms.”<p>It may be true that if you include all mutual and pension funds that only 5% have their own algos, and even that seems very low, its important to note that these funds trade through other brokers like Goldman and Knight capital.<p>I'd guess that vast majority( > 90%) of trade that they do are done algorithmically whether on a simple VWAP or a more complex "work it" style algos.<p>Maybe the author meant alpha seeking algos rather than basic order allocation algos?
In this space there is also <a href="http://www.quantconnect.com" rel="nofollow">http://www.quantconnect.com</a>.<p>As a concept it can take off for sure, but not sure the Bloomberg comparison is fully accurate.