XIRR for capital deployed

[nareshtedla]

Hi,

I have been doing back testing for a strategy for a period of 3 years of data. My returns are as below based on the stats:
Return [%] 114.659196
Buy & Hold Return [%] 209.190101
Return (Ann.) [%] 27.910586
Trades 379
Avg. Trade Duration 1 days 04:49:00

Return (Ann.) [%] assumes that my money is sitting idle when there is no trade. However, in real world situation, money is deployed into other trades. Is there a easy way to calculate the XIRR only for the period of time money is deployed instead of the whole duration?

Also, XIRR for Buy & Hold duration?

Any help in this regard is greatly appreciated. Thanks.

[kernc]

Return (Ann.) [%] assumes that my money is sitting idle when there is no trade. However, in real world situation, money is deployed into other trades.

You could look into and extrapolate stats['Exposure Time [%]']. If exposure time is, say, 10%, you could assume the real returns to be 10-times greater if all the money were allocated all the time in strategies of similar performance.

[nareshtedla]

This actually helps to an extent. Thanks for quick turnaround.

[kernc]

But, now reading what I wrote, the linear 10x doesn't take into account compounding, making it a pessimistic (i.e. underestimating) measure. 😅

[nareshtedla]

Agreed. Compounding can happen either way :-)