Inconsistency in WFM stability
I’m trying get to know how exactly works WFM stability.
On this simple example with 3 runs we can see different approach with calculating WFM stability with different metrics. NetProfit, DD, and R/DD.
Net profit seems like to be calculated as
Sum(NetProfits(OOS))/Sum(Days(IS)) / Sum(NetProfits(OOS)/Sum(Days(IS))
In Excel and as well in SQ results on left side screenshot are equally 128.47%. That’s seems like avg. daily NetProfit in IS / avg. daily NetProfit in OOS, which makes sense. Cool.
For Drawdown is used same logic as on NetProfit. In XLS and SQ is same value: 249.56%. But I’m not sure if now make sense doing something like a avg. daily DD in OOS/IS as DD is not so linearly increasing as NP.
What is weird is system of counting stability of R/DD. Before was avg. daily DD in IS/OOS, but it seems now we don’t use something like a normalization per 1day as before (result should be 396.35% – red field in XLS), but only formula
Sum(RDD(OOS)) / Sum(RDD(IS)) which we see 63.1% in SQ. Practically using is complicated as it is strongly depending on how big is %OOS. So bigger %OOS should have automatically better RDD stability result.
More severe is Inconsistency with method of calculation. Stability on NP or DD is same, but RDD is different. Is somewhere defined/described which metrics is in WFM stability calculated which way?
if you see an issue using StrategyQuant WFM please use our tool to report bugs and requests: https://roadmap.strategyquant.com/projects/sq4/tasks/new
When reporting a bug, a screenshot is usually not enough, you should provide also your strategy .SQX (if the problem is related to the strategy), or optimization configuration as saved CFX file.
We appreciate your feedback. Developers will check that
Note: to start using the roadmap you might need to register first using our homepage -> Login -> Registration
Viewing 2 posts - 1 through 2 (of 2 total)