One of the most important concepts in finance and investing is risk-adjusted returns: how much return you got per unit of risk that you took.

If I go bet everything I have on 23 in Roulette at Caesar’s Palace and I win, that doesn’t make it a good bet with a good risk-adjusted return, it just means I got lucky.

Any game where the house has a guaranteed edge such as roulette has a negative risk-adjusted return – if you keep playing you can only lose.

However, many investments have a positive risk-adjusted return, the question is how to figure out what that is.

The return part is pretty straightforward, you can look at the performance (past or future) to see what returns were achieved. There is rarely much debate about calculating returns.

It’s a lot trickier to measure the risk being taken. The human tendency is to assume if the outcome was right then it was low risk and if the outcome was wrong then it was high risk. However, this is not really true.

Maybe I think my roulette wager is low risk because I got a 23 in my fortune cookie the night before and so it was a prophecy that I knew what was coming.

People that made a bunch of money off crypto in the last decade tend to think that it was low risk and they were just better at analyzing the technology and its implications than others so they derived an “edge” that they were able to exploit.

People that didn’t make a bunch of money off crypto tend to think that people who did got lucky, the equivalent of winning by betting everything on 23. (The truth is no doubt somewhere in between).

There have been many attempts to quantify risk in investing. The most common is the Sharpe Ratio, a measure of historical volatility. The thinking is that something which has been very volatile in the past is likely to be volatile in the future and so should be considered risky. Many (including myself) have argued this isn’t the best metric because it penalizes upside volatility as well as downside volatility and metrics like the Sortino Ratio which focuses exclusively on downside volatility may be more appropriate.

However, I think it’s helpful to step back and think about what risk is at a little broader and more abstract level.

A sort of abstract but useful definition of risk is not getting what you want.

As my friend, Kris Abdelmessih pointed out, if a loved one was kidnapped and you needed to come up with ransom money in one day that was more than you had, the riskiest thing you could do is NOT bet everything on roulette.

Even though you are likely to lose, it’s the only chance you have at getting what you want most: the safe return of a loved one.
By this definition, choosing a career which maximizes your long-term wealth but also makes you hate your life is not a good risk-adjusted return. The return may be high, but so is the risk: you are very likely not going to get what you want.¹

Last Updated on June 6, 2025 by Taylor Pearson

1. This opens up the much larger can of worms of how to know what you want. On this point, I will mostly just defer to The Last Psychiatrist.