Negative Sortino Ratio Anomaly from Using Decimal Inputs

Summary

The document records a reported anomaly in comparing Sortino ratios when both are negative: an asset with poorer returns and greater downside deviation appeared to have the better ratio. The author later identifies an input-unit issue in a modified Sharpe-style formula attributed to Craig Israelson.

The formula uses excess return divided by standard deviation raised to a power determined by excess return relative to its absolute value. The author says the calculation requires percentage inputs and that using decimal inputs caused the apparent problem. This is a brief self-correction, not a general treatment of Sortino ratio interpretation: it gives no worked example, independent verification, or discussion of other possible causes of counterintuitive rankings.

Key ideas

• The reported ranking anomaly involved comparing assets with negative Sortino ratios.
• The author attributes the issue to applying a modified Sharpe-style formula with decimal inputs.
• The author states that the formula requires percentage inputs.
• The document offers a self-reported correction without examples or independent validation.

Tags

# Interperting Negative Sortino Ratios


# Interperting Negative Sortino Ratios












I noticed when both assets sortino ratios are negative the asset with the worse return and worse downside deviation has better Sortino Ratio. Is there a way to fix this?

## Answer by Anon9001 (score 0, accepted)

https://quant.stackexchange.com/a/71909

I finally figured out what was causing the issue here. The formula of Craig Israelson's Modification to Sharpe Ratio ie ER/(Standard Deviation^(ER/ABS(ER))requires usage of percentages and I was inputting decimal values. Consider this issue to be fixed.

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.