Sorting optimisation results by overall gain makes little sense, not least because overall resulting gain may have resulted randomly. Similarly, sorting by number of trades, or by hit rate alone gives little benefit. These are not metrics one would use to make choices based on strategy “robustness” or risk management considerations.
Sadly, there is no ability to add custom fields to this list, fields which may directly correlate between their value and the robustness or reliability of a strategy. One would have to screen shot the window, export it to an OCR tool, dump the results in excel, and attempt to derive meaningful metrics from a limited set of fields.
I would like to see alpha (α) and beta (β) fields in the optimize report, along with sharpe ratio and sortino ratio in particular, to help determine not “where the backtest ended up“, but “how it got there“, which is obviously the most important factor.
- How did the equity curve compare to a straight line between the start and the end?
- How did the equity curve compare to the underlying instrument’s returns in the same period?
These calculations are used to measure real world fund performance, particularly against a benchmark (or underlying instrument) and tell you how smoothly, or with what volatility the results were produced. Was the theoretical equity curve chaotic, volatile, or predictable? Gain, number of trades, and hit rate do not tell you that alone.
This feature (custom fields, and above mentioned ratios) would allow PRT back testing features to be taken more seriously.
.
