Returns, NAV and drawdowns
Understand returns, nav and drawdowns through definitions, interpretation and the assumptions used in KabuForge research examples.
What it means
NAV is account value: cash plus marked holdings. The simple return between adjacent valuation dates is NAV_t / NAV_(t−1) − 1. A whole-window return may use initial capital instead. Check the starting value before comparing returns.
How to interpret it
Drawdown measures the decline from the highest NAV reached so far. With NAV_t / running_peak_NAV − 1, the underwater chart shows zero or negative values. Maximum drawdown is the deepest observed decline in the window; it does not bound future losses.
Check the KabuForge example
Check price basis, dividends, fees and observation windows. KabuForge’s jQuantStats view uses N−1 adjacent-NAV intervals and excludes initial capital to the first NAV. The reference TOPIX series is a price index excluding dividends.
The 100 → 110 → 99 example
NAV moving from 100 to 110 to 99 produces adjacent returns of +10% and −10%. Cumulative return is 99/100−1=−1%; drawdown from the peak of 110 is 99/110−1=−10%. Returns compound rather than simply adding.
Variability and comparison measures
Volatility
Volatility usually expresses return variability as a standard deviation. Check sampling frequency, window and annualization. It measures variation, not only losses.
Sharpe ratio
A common form divides mean excess return over a risk-free return by the standard deviation of that excess return, using matching observation frequencies. Check the reference return, costs and sampling frequency. Zero standard deviation makes the ratio unavailable; annualization needs assumptions. See William F. Sharpe’s original discussion for the definition and time dependence.
Benchmark and TOPIX
A benchmark is a reference series for comparison. Price indices differ from dividend-inclusive return indices; this course’s TOPIX series is price-only. Match valuation dates, currencies and treatment of costs and dividends.
Related tutorial and sources
Applies to: main source candidate (unpublished) · Public source docs/en_US/GUI_RESEARCH_COURSES.md
Reviewed · KabuForge