What's Luck Got to Do With It?

In the story of David and Goliath, David won by changing the rules of the game. Rather than meeting this giant in a head-to-head battle with spear, shield and sword, David used a slingshot. And this is a good allegory for the kind of strategy you should pursue in business competition. According to Michael Mauboussin, “When competing one-on-one, follow two simple rules: If you are the favorite, simplify the game. If you are the underdog, make it more complicated.”

The reason this is the most effective strategy is that luck plays at least some role in virtually every business conflict. If you're the underdog, making the game more complicated will give luck more of a chance to triumph, so you may end up the winner (this one time, anyway) even though the odds are against you. If you’re the superior adversary, on the other hand (more scale, more resources, better research, whatever), then you should simplify the game to make sure that your superior skill is more likely to determine the outcome. David was clearly the underdog, ergo the new and unusual weaponry.

Michael Mauboussin’s marvelous new book, The Success Equation: Untangling Skill and Luck in Business, Sports, and Investing (2012, Harvard Business Review Press) takes a straightforward look at discerning how much of an event’s outcome should be attributed to skill or ability, as compared to how much should be attributed to luck or randomness. Most things in life depend on both luck and skill, but in varying amounts. Some things are more subject to randomness than others. And some things involve little skill at all.

It's easy to tell whether an event is entirely based on luck: Just ask whether you can intentionally lose. You can’t intentionally lose a coin toss or a game of roulette (because each is 100% luck), but you can intentionally throw a football or basketball game (because at least some skill is involved). Similarly, you can intentionally invest in an unwise manner, you can intentionally lose business to a competitor, and you can intentionally manage your company in an unprofitable way. Mauboussin's point is that even though the outcomes of each of these endeavors still involve a good deal of luck, skill does play a role.

The Success Equation is the subject of this week's Friday Book Share. Mauboussin is the eclectically interesting chief investment strategist at Legg Mason Capital Management. His books are fascinating far beyond their implications for finance. I’m a big fan, and this is one of his best.

To help us understand how important luck and skill are in any situation Mauboussin starts with the law of large numbers. Toss a coin in the air 3 times in a row and it's not impossible for you to get 100% heads. The odds of that happening are in fact one out of eight. But the likelihood of getting 100% heads on 20 straight coin tosses sinks to less than one in a million (0.5 to the 20th power). It's much more likely that only about 50% of your tosses will be heads. That's the law of large numbers at work. The larger your sample size, the more likely your results will be about average, and in a coin toss the average is 50%.

Mauboussin takes this and other commonly known properties of chance and randomness and teases out of them some very intriguing and compelling conclusions about how much luck rules our existence in a wide variety of endeavors. In sports, for instance, the outcome of a single basketball game tends to be more skill-driven than a football game, primarily because basketball teams each generate a larger "sample size" of potential scoring events during a single game. And football more than hockey, etc. He backs these assertions up with convincing quantitative evidence.

Mauboussin argues that one of the fastest ways to improve your skill is to pay attention to good, reliable feedback. In sports the feedback is immediate and obvious, but in business and many other fields the feedback isn’t always so reliable. Business executives often don't take the time to gather feedback on the wisdom of a prior decision, in order to study whether it could have been improved, so that the next decision will be better.

In fact, business people seem to think they can succeed solely by studying success, rather than by analyzing failure, as well. And this is where Mauboussin makes an important point about books by various business gurus: Because business involves a great deal more randomness and luck than sporting events do, you need to be careful to use as large a sample size as possible when you make inferences, and not to leave failures out of your sample. He cites Jim Collins’ iconic book Good to Great as an example. If you recall this book, Collins reviewed thousands of companies to identify eleven whose performance went from good to great, and then tried to analyze the strategies that went into their successes. But as Mauboussin says, “The trouble is that the performance of a company always depends on both skill and luck, which means that a given strategy will succeed only part of the time. So attributing success to any strategy may be wrong simply because you're sampling only the winners. The more important question is: How many of the companies that tried that strategy actually succeeded?”

Why does this matter to businesses? Because risky strategies may be more likely to succeed, but they are also more likely to fail. “Going for broke” inevitably will succeed sometimes for a business, and when it does it will likely do so in a spectacular way. But if those companies that went for broke and didn’t make it are removed from the sample (either because they’re no longer around or because their lack of success meant they weren’t included in the study), then what are we really saying?

(And by the way, Martha Rogers and I are always careful to write about business failures as well as successes, to the extent we can get access to details about them. In one of our first books, Enterprise One to One, we devote an entire chapter to a case study on MCI's failed customer loyalty initiative, for example, while our latest book Extreme Trust critiques failures at AOL, Nestle, Netflix, and a few entire business categories, as well.)

I posted a comment not long ago about Geoffrey Moore’s extremely thoughtful and well-argued recommendation to HP that bears directly on this topic. Moore recommends that HP should concentrate on just one innovation per half decade, because that’s all a CEO can truly focus the company on, and if the whole company isn’t focused on it then its go-to-market strategy will fail. But I'm not so sure about Moore's idea, and Mauboussin would almost certainly be on my side in this. Moore's logic is persuasive and, as usual, his knowledge of the venture and tech innovation business shines, but I think success at innovation inherently involves a great deal of luck and randomness. No matter how much skill you have (and no matter how much attention your CEO focuses on the go-to-market strategy), you're still going to need a large sample size before you can be confident that your superior skill will shine through. I don't think one mega-effort per half decade will do it. Fifty might, but one is just a crapshoot (IMHO).

Yet another useful lesson from Mauboussin's enthralling book is what he calls “the paradox of skill.” He says that as any field attracts more and more talent, the general skill level will rise, and as this happens the statistical variance of results will decline. As people get better and better, in other words, the differences among them shrink. Mathematically, this gives us the paradox of skill: As skill levels improve generally, luck plays a greater role in determining outcomes.

To illustrate this, he asks why Ted Williams was the very last baseball player to achieve a batting average greater than .400 for the whole season (in 1941). The answer, he suggests (and he's not the only one to suggest it), is that “the variance of batting averages has shrunk over time…as the skill of the hitters has improved…. This decline in variance explains why there are no more .400 hitters. Since everybody gets better, no one wins quite as dramatically. In his day, Williams was an elite hitter and the variance was large enough that he could achieve such an exalted average. Today, the variance has shrunk to the point that elite hitters have only a tiny probability of matching his average.”

To prove his point he includes a graph showing just the standard deviation and coefficient of variation for batting averages over the last 140 years:

(Note that this graph doesn’t say that batting averages have been going down, but that the differences between themhave been going down.)

The same exact logic applies to business and investing. As skills increase across the board (because of the competition for talent, better education, training, and so forth), the role of luck in determining outcomes also increases. Over the last several decades, for instance, “investing went from being dominated by individuals to being dominated by institutions. As the population of skilled investors increased, the variation in skill narrowed, and luck became more important.”

And in business, Mauboussin suggests: “Recent research shows that while some companies do sustain superior economic performance, the rate of reversion to the mean today appears to be accelerating". [Translation: The role of luck is increasing.]

So as randomness plays an ever greater role in business success, what should the well-managed business do? He suggests several things:

• Find and use more immediate feedback.
• Focus more on good processes, and less on specific outcomes.
• Balance your company’s emphasis on exploration of new opportunities with exploitation of current ones (when the rate of change in your category increases, more attention to exploration is called for).
• Use checklists to ensure that your skills and advantages aren’t unnecessarily undermined by random mistakes.