Gain-to-Pain Ratio: Schwager’s Metric for the Trader Who Wants to Sleep

Sharpe needs a standard deviation and a risk-free rate. Sortino needs a target return and a denominator most software computes incorrectly. Calmar needs a high-water mark and three years of history.

The gain-to-pain ratio needs addition.

Jack Schwager popularised it in Hedge Fund Market Wizards, and its appeal is that it asks the question a human being would actually ask. Not “how much return per unit of variance.” Just: for everything I lost along the way, how much did I end up with?

The formula

Monthly returns are the convention, and it matters. Compute it on daily returns and every ordinary noisy Tuesday enters the denominator, producing a much lower number that is not comparable to any published figure. If you change the frequency, say so.

Worked example

The same twelve months used across this cluster:

Red bars are the pain: −1.8, −2.6, −0.4 and −3.9 percent.

Read it in plain language: this trader earned 1.33 units of net result for every unit of downside endured. That sentence is intelligible to anybody. No other risk-adjusted metric can say that.

The hidden identity: GPR is profit factor minus one

This is not widely pointed out, and it makes the metric much easier to hold in your head.

Split the numerator into gains and losses. Total return is the sum of the up months minus the sum of the down months. So:

Check it against the example. Positive months sum to 20.3%. Negative months sum to 8.7%. The monthly profit factor is 2.33, and 2.33 minus 1 is 1.33. Exactly the GPR.

So the gain-to-pain ratio is a monthly profit factor, rebased so that breakeven sits at zero instead of one. Which means the mental conversions are trivial:

It also explains why GPR benchmarks look so much lower than profit factor benchmarks. They are measuring the same thing on a shifted scale, at a different frequency.

What the numbers conventionally mean

The guide that travels with the metric, for a multi-year monthly record:

Our example scored 1.33, which sounds excellent until you remember it is one year. Twelve months is not a track record. It is a sample of twelve numbers, four of which do all the work in the denominator.

What it does better than the alternatives

It survives non-normal returns. There is no standard deviation anywhere in the calculation, so fat tails, skew and kurtosis cannot distort it the way they distort Sharpe. A single enormous winning month raises the numerator and leaves the denominator untouched, which is the correct treatment and the exact opposite of what Sharpe does.

It is unambiguous. No target return to argue about, no risk-free rate to omit, no denominator convention to get wrong. Two people computing a GPR on the same monthly returns will get the same answer, which is not true of Sortino.

It maps onto how traders actually experience risk. Nobody remembers their standard deviation. Everybody remembers the four months that hurt. GPR puts precisely those months in the denominator and nothing else. This is the same reasoning behind the argument that volatility is the fee, not the fine, applied to a scorecard.

Where it falls down

It cannot see path. A year with four separate 2% losing months and a year with three flat months and one 8% loss can produce a similar GPR. They are not similar experiences. Four scattered dents are survivable; one crater is the thing that ends accounts. GPR is blind to sequencing, which is exactly the gap that Calmar and MAR exist to fill.

It flatters the same strategies everything else flatters. A short-volatility book that clips small premiums has almost no losing months, so the denominator is tiny and the GPR is spectacular. Until it is not. No ratio built on realised returns can price a loss that has not occurred yet.

Small denominators break it. With only one or two losing months in the sample, a single number is doing all the work and the ratio swings wildly with each new observation. Under three years, treat it as a description of what happened rather than a measure of what you can expect.

The same trader, four verdicts

Run all four ratios on the twelve months above and see how differently they read the same year:

Four numbers between 1.28 and 3.02, all correct, all describing one identical set of twelve returns. None of them is the truth. Each is a different question, and the ratio you quote reveals which question you think matters.

Run your own numbers

Also available on its own page: the full trading performance metrics calculator.

Frequently asked questions about the gain-to-pain ratio

What is a good gain-to-pain ratio?

Above 1.0 over a multi-year monthly record is very good, and above 2.0 is exceptional enough to warrant checking the sample length and looking for hidden tail exposure. Between 0.5 and 1.0 is a solid, realistic result for a working trader. Below zero simply means you lost money.

Should I use monthly or daily returns?

Monthly, if you want a comparable figure. Daily returns put every ordinary red day into the denominator and produce a much smaller number that cannot be benchmarked against anything published. The metric was defined on monthly data and the thresholds assume it.

How is it different from the Sortino ratio?

Sortino squares the shortfalls before averaging them, which weights one large loss far more heavily than several small ones. Gain-to-pain simply adds them, treating a 4% loss as exactly twice as painful as a 2% loss. Sortino also requires you to choose a target return; GPR does not. GPR is the simpler and more literal of the two.

Can I compute a gain-to-pain ratio on trades instead of months?

You can, but you have just computed a profit factor minus one, which is a well-established metric with its own benchmarks. Use the profit factor framing for trade-level analysis and reserve gain-to-pain for the equity curve, where the monthly convention makes it comparable to other people’s numbers.

Why does Schwager prefer it to the Sharpe ratio?

Because the traders he interviews tend to have return distributions that Sharpe handles badly. Concentrated conviction bets, long-volatility structures and opportunistic macro books all produce large winning months that inflate a standard deviation and depress a Sharpe ratio without doing any harm to the trader at all. Gain-to-pain does not make that mistake, and it happens to be the ratio you can compute for a manager from nothing more than a column of monthly numbers.