Quant Strategies Implemented by the Quantopian Community(blog.quantopian.com)
blog.quantopian.com
Quant Strategies Implemented by the Quantopian Community
http://blog.quantopian.com/5-basic-quant-strategies-implemented-by-the-quantopian-community/
7 comments
Quite a few of these strategies appear to be underperforming their benchmark index. Some others are known to not work anymore like the 'Turtle' system or only work under special circumstances like the gold/gold miner arbitrage (requires sufficiently high crude oil prices). I haven't checked this in detail, but I would also be very wary of curve fitted backtest results that won't hold up for future market movements.
Total returns versus the benchmark is not the only important measure. For instance, the minimum variance strategy visibly underperforms its benchmark but it has way lower volatility and fewer drawdowns. These are more important in the real world where you are likely to be heavily levered.
I love the one that used AAPL in the backtest and then advertises another site.
See what I'd like to see is a strategy that computes the consensus view of all those strategies.
Build a random forest with all these quant algorithms as the decision trees?
I wonder how good that'd be.
I wonder how good that'd be.
My guess: over-fitted. It will make money until it explodes.
Isn't that called "the market"?