Rolling 12-Month Sharpe

What is Rolling 12-Month Sharpe?

The headline Sharpe ratio collapses an entire backtest into one number: average excess return divided by volatility over the full period. That single figure hides timing. A strategy with a lifetime Sharpe of 1.0 might have earned it smoothly, year after year, or it might have posted a Sharpe near 3.0 in two extraordinary years and roughly zero the rest of the time. Rolling 12-Month Sharpe answers the timing question by computing the Sharpe ratio over the trailing twelve months at every point in the test, then tracing those values as a line through time.

Each point on the line is a self-contained Sharpe calculation that uses only the most recent twelve months of returns. As the window slides forward one month at a time, the oldest month drops off and the newest is added, so the line shows how the strategy's reward per unit of risk drifted, spiked, and sagged across market regimes. Reading it is closer to watching a heart-rate monitor than reading a single resting pulse.

Where the headline Sharpe is a verdict, rolling Sharpe is the evidence behind the verdict. It tells you whether the edge was durable or episodic, when it broke down, and how violently it swung between good and bad environments. A strategy whose rolling Sharpe rarely dips below zero is behaving very differently from one whose rolling Sharpe oscillates between +3 and −2, even when the two share an identical lifetime average.

Formula

For every month in the test, once at least twelve months of history exist, the trailing twelve monthly returns are collected. From each month's return the risk-free rate for that month is subtracted, giving a monthly excess return. The mean of those twelve excess returns is multiplied by 12 to annualize the numerator, and the sample standard deviation of the twelve monthly returns is multiplied by the square root of 12 to annualize the denominator. The ratio of the two is that window's Sharpe. Because the window holds exactly twelve monthly observations, the expression simplifies to the mean monthly excess return divided by the monthly standard deviation, scaled by the square root of 12. The standard deviation uses the unbiased (n minus 1) divisor that is standard in finance software.

One convention deserves emphasis: which risk-free rate gets subtracted. SledgeKey uses the Treasury rate that actually prevailed during each trailing window rather than a single average for the whole test. A window sitting in 2013 is measured against a baseline near zero; a window sitting in 2023 is measured against a baseline near 5 percent. This keeps each point honest to the opportunity cost an investor faced at that moment, and it is a deliberate difference from the headline Sharpe, which applies one average baseline across the entire period. Returns themselves are simple percentage changes in the simulated monthly portfolio value, net of the transaction costs already charged in the simulation.

Why Rolling 12-Month Sharpe Matters in Backtesting

The decision rolling Sharpe informs is trust. A backtest with an attractive lifetime Sharpe is only investable if that performance was reasonably stable; a number built from one remarkable year is a coin flip dressed as a track record. Rolling Sharpe exposes the distinction directly. If the line spends most of its time comfortably positive, the edge looks structural. If the line is mostly flat or negative under one towering spike, the lifetime figure is an artifact of a single window and should be discounted heavily.

The failure mode of ignoring it is over-trusting an average. Investors routinely allocate to a strategy on the strength of one backtested Sharpe, then abandon it during the long stretches when the trailing figure was negative, stretches the headline number gave no warning about. Rolling Sharpe is also a regime detector. It tends to fall during volatility spikes and bear markets and to rise during calm trends, so the shape of the line tells you what kind of environment the strategy needs to thrive and how badly it suffers when that environment disappears. A momentum strategy that looks brilliant on a full-period Sharpe but shows a deep, multi-quarter trough around early 2009 or early 2020 is telling you precisely where it will hurt next time.

How SledgeKey Implements Rolling 12-Month Sharpe

Rolling 12-Month Sharpe appears on the results page as a line chart plotted over the backtest timeline, below the headline metrics. The line begins twelve months into the test, because a trailing-year Sharpe cannot be computed before a full year of returns exists. Each point is the annualized Sharpe of the prior twelve months, computed from the simulated monthly portfolio values after the transaction costs applied at each rebalance.

The risk-free baseline for each window comes from SledgeKey's daily Treasury data, averaged over that specific window, and the baseline used is reported alongside the Sharpe figure so the comparison stays transparent. Read the chart together with the equity curve and the rolling returns line. A stretch where cumulative value is still climbing but rolling Sharpe is falling means the strategy is making money while doing so with steadily worse risk efficiency, an early warning the equity curve alone will not give you. When the line sits mostly above 1.0 and seldom turns negative, the strategy earned its keep across the cycle rather than in a single burst.

Common Pitfalls

The largest misreading is treating a single low or negative window as a verdict on the strategy. Twelve monthly observations is a small sample, and one bad month inside the window can drag the figure down sharply, then drag it back up when that month rolls off twelve months later. Short-window Sharpe is noisy by construction. Judge the line by where it spends most of its time and how often it crosses below zero, not by any one reading.

A second pitfall is forgetting that the window's edges create artifacts. A single crash month influences the rolling Sharpe for a full twelve months and then vanishes from it abruptly, which can produce a cliff in the line that looks like a regime change but is only the bad month aging out of the window. The same overlap means adjacent points share eleven of their twelve months and are highly correlated, so they should not be counted as independent pieces of evidence.

A third pitfall is comparing rolling Sharpe across strategies without remembering the baseline. Because each window subtracts the prevailing risk-free rate, the same raw return produces a lower rolling Sharpe in a high-rate environment than in a zero-rate one. Two strategies that look different on this chart may simply have lived through different rate regimes, not had different skill. Always check that the windows being compared cover the same calendar stretch before reading anything into the gap.

Related Concepts