Fancy performance statistics should carry a health warning when it comes to picking a fund manager. These measures (Sharpe Ratio being one well known example) can be comforting because at first glance they appear to give concrete answers to key questions, such as: ‘Which of two managers is better and thus which should I select for my portfolio?’.

While these statistics will tell you which manager’s performance track record was ‘better’, this isn’t at all the same as which manager is better, for the very simple reason that performance track records are highly influenced by luck. Good stats may reveal skill, but they may also be entirely the result of previous good luck. And because you can’t expect luck to last (whereas a good manager’s skill should persist), probably the most important thing to understand and account for when picking a fund manager is luck itself.

A lot of great investors have tackled the topic of luck and randomness. From Howard Marks, Co-Chairman of Oaktree Capital (the heading of this piece is a nod to his outstanding memo of the same title), Credit Suisse’s Michael J. Mauboussin plus Nassim Nicholas Taleb. However, as an illustration of the distorting effects luck can have on performance appraisal I’m going to adapt an example used by the Sage of Omaha himself, Warren Buffett.

Suppose everyone on the planet was handed a regular pound coin and each day told to flip it and guess which way it would come up. Those that guessed correctly would receive the coin (and any winnings from previous days) from those who guessed incorrectly. Based on data from the United Nations and the US Census Bureau there are about 7.4 billion people alive today. So after one flip, there would be 3.7 billion people now with £2 each. After three flips, there would be 1.85 billion people with £4 each. And so on.

Now, human nature being what it is, after five flips those with five correct guesses in a row (and £32 in total) would probably be starting to feel pretty special. All 231 million of them.

After 10 days and 10 flips, those with 10 correct guesses in a row (and £1024) would certainly be bragging to their envious friends, offering them tips on the ‘art’ of on guessing coin-flips (although by this stage there would also be fierce debate on whether guessing coin-flips was really an art or a science, or both). There would also still be 7.2 million such winners.

After 20 flips, those with 20 correct guesses in a row (and over £1million in the bank) would almost certainly believe that they were supremely gifted in predicting coin flips. Some of them would undoubtedly be giving speeches to packed crowds and writing books with titles such as: “How I Turned a Pound into a Million in Twenty Days Working Thirty Seconds a Morning” and jetting around the world to seminars on coin toss prediction. There would still be over 7,000 of such people.

Even after 30 rounds, there should be about seven people who had guessed right every single toss. Imagine that.Even the most rational, stable person is going to have trouble believing that their coin isn’t magic/blessed and/or they aren’t psychic.

That is the power of luck and randomness. Get enough people taking part in an activity subject to a lot of randomness and luck is going to play a big role in who wins and who loses.

Investing is just such an activity, subject to a very large degree of randomness. And with over 400 funds in the UK focused on UK equities alone, sometimes picking a manager can feel a lot like a trip to the local casino, minus the fun.

But, before we lose all hope in active management, we need to pause here. Just as it’s a big mistake not to realise that luck plays a very large role in investing, so too is it to assume that luck is the only factor at work. Whilst it can sometimes seem that way, investing actually isn’t the equivalent of flipping a fair coin. Rather it is more similar to flipping a biased one, with the degree of bias linked to the skill of the investor and how much hard work they put in to choosing their investments. A random process, but a random process with unequal odds.

Is it possible to identify the managers with the most biased coins in advance? Yes, we think so. To start with, longer track records are less likely to be purely based on luck than shorter ones and so are more reliable in identifying skilled and hardworking investors. And unlike the owners of lucky coins, skilled investors tend to share characteristics that can be, tentatively, identified through in-depth due diligence.

All in all, when it comes to choosing who to invest with, while it’s never going to be possible to guarantee that you’re getting a skilled rather than lucky manager, it’s at least possible to take a bit of the casino out of the process.

*An edited version of this article appeared in New Model Adviser magazine on 30 May 2016, entitled Why a good fund manager often requires a bit of luck.