analyzing churn of a strategy I was working on.

on this model the cost of trading was about 5% of equity a year, which is quite high. the turnover is also a bit high considering the nature of the strategy.
35% of trades entered and exited in a single day and 59% of trades where completely gone by the 3rd day. that would be fine if the strategy was medium frequency but that's not the case.
32% of total costs come from these round-trips.
this strategy obviously has a churn problem generated by weak churn controls at the boundary of available slots in the portfolio. this is not so much a strategy problem, as the strategy has decent edge, but more of a design flaw.
for example lets say that you allow 50 positions to be opened at any time. what happens to those positions say that are in the bottom 10% , and that slightly drop off that top list, without any significant improvement in edge? how much are you getting paid really for that adjustment? and the answer is in most cases the change is so small that it certainly doesn't pay for the full cost of round-tripping an entire position.
this becomes even more problematic when you're trading on venues and markets that don't have as much liquidity and on top of the fees you're also paying a lot more to cross the book.
it's important to keep strict thresholds for rebalancing. either based on the expected value of that decision outweighing the costs or setting a standard threshold % that it must break before churning the existing positions.
(for the tests an hypothetical $10k starting size was used)