Hi,

I recently noticed a potential issue with qs.reports.metrics/html when a value is provided for the benchmark parameter. Specifically, some metrics in the results seem to show inconsistencies.

To better understand this, I conducted a comparison by passing both benchmark and return together, and then using only return to call qs.reports.metrics. The "Same?" column in my analysis highlights the comparison between the "Return with Benchmark" and "Only Return" outputs.

It is worth noting that metrics such as "R^2" and "Information Ratio" are missing when using only return. Since these metrics rely on both return and benchmark return for their calculations, I would like to suggest displaying "-" for benchmark, similar to the method used for Beta and Alpha.

Additionally, metrics like "Kelly Criterion," "Expected Shortfall (cVaR)," "Payoff Ratio," "Outlier Win Ratio," "Outlier Loss Ratio," "Serenity Index," "Avg. Up Month," and "Avg. Down Month" yield differing results when a benchmark is included compared to when they are calculated individually. This discrepancy could potentially relate to the issue I previously mentioned in "The calculation result of CVAR in the HTML report appears to be questionable #467." It seems there may be variations in the logic applied during calculations, depending on whether a DataFrame or Series is used.