about Portfolio OOS years computing

 I found two problems when I calculated portfolio for 10 bitcoin strategies:

1.The number of years for the portfolio strategy OOS appears to be miscalculated as the number of years for the entire sample.
2. As you can see in the attachment, the amount of single strategy OOS data I selected is one year, 10 months and 7 days. but total data years of SQX is 1 year. the calculation in the code is based on total data years, not total trading years.This leads to great errors in calculating the annualized rate of return and CAGR. 
3.I think CAGR and the annualized rate of return should be calculated accurately according to total trading month or day, and then converted into annualized. Just like the calculation of the Sharp ratio is also in this way. I think this calculation is more accurate and convincing.
   This phenomenon is easy to reproduce, so I will not upload the strategy. As we all know, building a portfolio with low correlation is the closest thing to a free lunch.It is strongly recommended that defects related to portfolio be fixed and related features as soon as possible.

Attachments
Portfolio-OOS years.jpg
(809.23 KiB)
  • Votes +4
  • Project StrategyQuant X
  • Type Bug
  • Status New
  • Priority Normal

History

b
#1

binhsir

13.11.2022 05:42

Task created

b
#2

binhsir

13.11.2022 05:42
Voted for this task.
AA
#3

Alex

13.11.2022 11:09
Voted for this task.
E
#4

Emmanuel

13.11.2022 19:03
Voted for this task.
b
#5

binhsir

14.11.2022 14:10

Attachment AnnualPctReturn.jpg added

AnnualPctReturn.jpg
(353.43 KiB)
b
#6

binhsir

19.11.2022 01:58

Description changed:

 I found two problems when I calculated portfolio for 10 bitcoin strategies:

1.The number of years for the portfolio strategy OOS appears to be miscalculated as the number of years for the entire sample.
2. As you can see in the attachment, the amount of single strategy OOS data I selected is one year, 10 months and 7 days. but total data years of SQX is 1 year. the calculation in the code is based on total data years, not total trading years.This leads to great errors in calculating the annualized rate of return and CAGR. 

3.I think CAGR and the annualized rate of return should be calculated accurately according to total trading month or day, and then converted into annualized. Just like the calculation of the Sharp ratio is also in this way. I think this calculation is more accurate and convincing.
   This phenomenon is easy to reproduce, so I will not upload the strategy. As we all know, building a portfolio with low correlation is the closest thing to a free lunch.It is strongly recommended that defects related to portfolio be fixed and related features as soon as possible.

b
#7

bentra

20.11.2022 04:41
Voted for this task.

Votes: +4

Drop files to upload

or

choose files

Max size: 5MB

Not allowed: exe, msi, application, reg, php, js, htaccess, htpasswd, gitignore

...
Wait please