Sharpe Ratio Lookback Choice and Manager Skill Assessment
Summary
The discussion asks whether a manager’s Sharpe ratio should use the risk known when an investment was first made, rather than risk measured over its full return history. The answer reframes this as a question about choosing the lookback horizon for estimating a portfolio manager’s underlying Sharpe ratio. It points to the same estimation issue for individual assets and portfolios, since a manager’s portfolio combines assets.
The response cautions against judging performance from the investment date through the present when doing so requires a very long sample. A high measured Sharpe over that span may reflect luck rather than skill. The document offers no formal estimator, empirical comparison, or universal horizon; it directs readers to related discussion about lookback length. Its main lesson is that a Sharpe ratio depends on the measurement window, so conclusions about manager competence require care.
Key ideas
- The Sharpe ratio estimate depends on the return history and lookback horizon used.
- Assessing a manager over the entire life of an investment may require an impractically long sample.
- A high Sharpe ratio over a long lookback can arise from luck rather than skill.
- The discussion does not establish a single correct horizon or provide a formal estimation method.
Tags
Full text
# Using Sharpe ratio as a measure of performance seems misleading # Using Sharpe ratio as a measure of performance seems misleading By defintion, doesn't the Sharpe ratio use a denominator that is the risk, and that is the risk that's taken up until right now, and if only the risk changes and nothing else, the ratio for an investment will change? If that is true, can I argue that it is a misleading measure of the investor's competence because the investor can only be expected to know the risk at the time of the purchase? Therefore the risk that should be used in the Sharpe ratio should be the risk at the time of investing and not change afterwards. Do you agree? I mean if we look at a portfolio and the Sharpe ratio is high or low, we draw the conclusion that the investor is good if the Sharpe ratio is high and likewise, but in fact, it will be misleading and the right measure to show who is a good investor would be to use the return in the numerator of the ratio (as usual) and the risk that was taken at the time of the investment in the denominator, which may very well have been in 1987 or 1914 for an investment that was bought and then hold and a long time. It should be the available risk taken for the investor at the original point in time when the investment was made. Do you agree? ## Answer by KaiSqDist (score 1, accepted) https://quant.stackexchange.com/a/81026 If I understood your question well, what you are actually asking is this - What is the correct way (in terms of lookback horizon) to measure the "true" Sharpe based on the performance of the portfolio manager? This query (in my opinion), helps to resolve the dispute well: Portfolio rebalance - How many data back do I need to perform sharpe ratio optimization Measuring the Sharpe of an asset or a PM is not that much different as a PM is just a compilation of assets. The query above talks about how to estimate the "true" Sharpe based on how far back the lookback horizon. I personally do not think it makes sense to measure the performance of a PM based on the time of investment to the current date. This is because, as the query and as @phdstudent answers - you need a hell of a long lookback. Therefore, if you want to measure the performance of the PM from the investment to the current date, a good Sharpe could just be due to luck.
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.