Here's one of my contest entries. Perhaps of interest to folks. Nothing to write home about, but not butt-ugly either.
Here's one of my contest entries. Perhaps of interest to folks. Nothing to write home about, but not butt-ugly either.
Apologies for the naive request, but can you provide some insight into these graphs?
Hello Moataz -
I don't know precisely how the turnover is calculated to obtain the Daily turnover, Average daily turnover, by month, and Average daily turnover, net plots. Presumably, for example, if my portfolio consists of 3 stocks, with relative weights (0.5,0.5,0) and I go to in one day (0,0,1.0), it would correspond to 1.0 turnover for that day. But what if I go from (0.5,0.5,0) to (0.4,0.2,0.4)?
Regarding the spike in daily trading volume starting at the end of 2008, it is related to the spike in profit as the market melted down. The algo has more capital and since it is all deployed, the number of shares traded goes up. I suspect that the algo is very profitable over this time period because it looks for price discrepancies over a max time frame of 5 trading days. I'd expect that it would pick up on inefficiencies driven by short-term uncertainty.
Presumably, the Distribution of Daily Turnover Rates is simply a histogram of the Daily turnover. There is a line and a bar, so I'm not sure how to interpret the plot. For a histogram, I'd expect only bars (bins).
There should be some documentation out there on what the tear sheet is calculating, in detail. We have:
http://quantopian.github.io/pyfolio/
https://github.com/quantopian/pyfolio
I'm not sure which one is authoritative and corresponds to the current revision available on Quantopian.