Annual Return (CAGR)
Annual Return, also called the Compound Annual Growth Rate (CAGR), is the constant per-year rate at which a portfolio's starting value would have grown to its ending value over the backtest window. It is the fairest single number for comparing strategies that were tested over windows of different lengths.
What is Annual Return (CAGR)?
Annual Return is a smoothed compound growth rate. Suppose a portfolio grew from one hundred thousand dollars to two hundred thousand dollars over ten years. Total Return is one hundred percent. CAGR is roughly 7.18 percent, the rate at which a balance compounding once per year would land at the same ending value. Annual Return collapses every up year, down year, and quiet stretch into a single steady growth rate that, applied consistently, would replicate the actual ending balance.
The reason this metric exists is comparability across time. Total Return scales with the length of the window, so a 100% total return over five years is very different from the same 100% over twenty years. Annualizing strips that effect out. A strategy compounding at 10 percent a year shows the same 10 percent annualized return over one year, five years, or twenty years, even though its total returns over those windows are roughly 10 percent, 61 percent, and 573 percent respectively.
CAGR is a backward-looking summary, not a forecast. It describes the rate the portfolio actually compounded at over the tested period. It does not say anything about how steady the path was, when most of the return arrived, or what to expect going forward. Two strategies with identical CAGR can have wildly different paths to that same number.
Formula
This is the standard geometric-mean form. The expression takes the total growth multiple (VT divided by V0), raises it to the reciprocal of the number of years to get the per-year growth multiple, and subtracts one to express the result as a percentage rate. Both endpoints are taken net of all transaction costs already deducted by the simulation, so the figure reflects what the strategy would have netted in practice rather than a gross theoretical number.
The number of years is computed from the exact day count between the first and last dates of the backtest window divided by 365.25, so partial years are handled cleanly. A window that runs from January 15 of one year to October 7 of a later year produces a fractional-year denominator rather than rounding up or down to a whole number.
CAGR is a geometric mean, not an arithmetic mean. An asset that loses 50 percent in year one and gains 100 percent in year two has an arithmetic mean of plus 25 percent but a CAGR of zero, because the ending value equals the starting value. When you want to know how money actually compounded, the geometric form is the correct one. When you want a measure of the typical year's return, an arithmetic average has its place, but it does not match what the portfolio actually delivered to its owner.
Why Annual Return (CAGR) Matters in Backtesting
CAGR is the metric used to compare strategies fairly. Two strategies tested over different historical windows cannot be compared on Total Return without rescaling, and any single year's return is too noisy to tell you much. Annualizing puts them on the same per-year basis and makes apples-to-apples comparisons possible.
It is also the foundation for almost every other annualized metric. Annualized volatility, the denominator of the Sharpe ratio, sits on the same time scale. Excess return in a Sharpe calculation is annualized portfolio return minus the annualized risk-free rate. The Calmar ratio is annualized return divided by max drawdown. Once a backtest reports CAGR, the rest of the risk-adjusted toolkit fits together cleanly.
Where CAGR can mislead is in implying steadiness. A portfolio with a 12 percent CAGR did not actually return 12 percent every year. It returned whatever it returned, and the compounded outcome happens to match a smooth 12 percent line. A strategy that crashed 40 percent in year one and then ground out 30 percent a year for six years will show a CAGR that hides the early collapse. CAGR belongs on the results page next to max drawdown, win rate, and volatility, never on its own.
How SledgeKey Implements Annual Return (CAGR)
Annual Return appears as one of the four headline metrics on the results page, alongside Total Return, Sharpe ratio, and max drawdown. It is computed from the simulated portfolio value series used for every other return metric. The first day of the configured backtest window provides V0, the last day provides VT, and the number of years is the calendar gap between them expressed as a decimal, so a window of seven years and four months reads as roughly 7.33 years.
The same CAGR is reported for the configured benchmark over the same window. Comparing the two is the cleanest way to see whether the strategy added value on a time-normalized basis. If the strategy compounded at 11 percent and the benchmark at 9 percent, the strategy added two percentage points a year of compounded outperformance, a figure that scales with time the way a real investor experiences it. Over a twenty-year window, two extra points of CAGR is roughly a 49 percent larger ending balance.
The cumulative-return chart on the results page provides the visual companion to the CAGR number. A constant-CAGR line would be a smooth exponential curve. The actual portfolio line shows where the strategy outpaced its own average and where it fell behind. The bigger the visual gap between the smoothed average and the actual path, the more work the CAGR number is doing to hide volatility.
Common Pitfalls
The first pitfall is comparing CAGR across windows with very different market regimes. A 14 percent CAGR from 2010 to 2020 is not the same accomplishment as a 14 percent CAGR from 2000 to 2010. The market environment matters, and the headline rate alone does not tell you whether the strategy delivered genuine alpha or simply rode a powerful tailwind. Pair every CAGR with the benchmark's CAGR for the same window before drawing conclusions.
The second pitfall is treating CAGR as a forecast. A backtest's annualized return is a property of one historical sample. A strategy's true expected return is not observable and is almost always different from the realized CAGR, particularly in shorter windows where one or two good years can dominate the result. The shorter the test, the more the reported CAGR is a guess in expensive clothing.
The third pitfall is reading CAGR without reading max drawdown alongside it. A 15 percent CAGR with a 60 percent drawdown along the way is a strategy almost no real investor would have ridden to the end. A 9 percent CAGR with a 20 percent maximum drawdown is a strategy a real investor could have actually owned. Annual Return summarizes the destination; max drawdown describes the worst stretch of the journey.
CAGR is a smoothed average. A portfolio with a 10 percent CAGR did not return 10 percent each year. It returned whatever it returned, and the compounded outcome happens to match a smooth 10 percent line.
See Annual Return in your own backtest
Run a backtest on any screening strategy and see CAGR computed on point-in-time data, alongside Sharpe, drawdown, and benchmark comparisons, free.
Run a Backtest