Anomaly or Signal

Yesterday, reported SB Nation, tennis legend Novak Djokovic lost his first-round match at the US Open to unseeded Mariano Navone: 39-year-old champion against 25-year-old admirer. It lasted four and a half hours. Djokovic vomited, cramped, struggled with back pain, and eventually lost the final two sets 6–2, 6–1.¹ The news story was obvious. Djokovic is getting old. This is the beginning of the end.

»Perhaps living is less about discovering the true story than maintaining a good enough one, and knowing when to let it go«

Except we don’t know that.

One match is a single data point. It may be a signal of decline. It may be an anomaly. It may be illness, heat, fatigue, or simply a bad day. Djokovic himself didn’t seem to know. »I’m just trying to address really what’s going on and how far I can go this way«, he said at the post-match press conference.²

And now Djokovic has to decide what to do next.

A few days ago I played tennis for the first time in a year or so, against people who play every week. I was predictably terrible – evidence, it seemed, that I can’t play tennis. But one match isn’t enough to establish that. I might have been ill. I might have had an unusually bad day. Or I might be one of the rarer people who are competent six days a week and a catastrophe on the seventh.

We do this constantly: see an event and build a story around it. Polymath (and, controversially, eugenicist) Francis Galton found a version of the problem in the nineteenth century, measuring parents’ and children’s heights. Exceptionally tall parents had children who were tall, but less so; exceptionally short parents had children taller than themselves. Extreme observations tend to be followed by less extreme ones, a pattern Galton named »regression towards mediocrity«.³ There doesn’t need to be a causal mechanism behind it, but our minds instinctively look for one.

Psychologist Daniel Kahneman described a related tendency: when we see A followed by B, we instinctively invent a mechanism connecting them. A dog runs down the street, so it must have escaped its owner. Djokovic loses, so he must be declining. The story isn’t necessarily false. We just don’t know yet.⁴

Humans turn sequences into narratives because narratives compress enormous amounts of uncertainty into something we can carry. We write the story as we read the book. We are audience and playwright at once.

Bayesian reasoning offers a more disciplined alternative to the story-building instinct.⁵ Instead of asking whether a story is true or false, it asks how likely it is, and revises that likelihood every time new evidence arrives. Before the match, you might have put the odds of Djokovic’s decline at one in twenty. After this loss, the odds should move – to one in five, say – but they shouldn’t leap straight to certainty. A single match moves a belief. It doesn’t replace it with a final verdict.

More strangely, the decision changes what happens next. Djokovic can play another tournament, change his preparation, get some rest, or retire from tennis altogether. Each choice produces different evidence. Sociologist Robert K. Merton called one version of this a »self-fulfilling prophecy« – a belief or action can help bring about the conditions that make the belief appear true.⁶ Our actions do not just respond to uncertainty; they partly determine what we can know next.

A scientist choosing which experiment to run does something similar. So does a person deciding whether to stay or go.⁷ The decision isn’t merely a response to uncertainty. Sometimes it’s what moves the uncertainty forward.

Perhaps the cycle isn’t simply observe, then decide. Instead we construct plausible stories, pick one, act as if it’s true, observe what happens, then revise the story. The difficult part is the revision. We need to remember the story well enough to learn from it, but forget it enough that the world may change it. The story that got us here was, by definition, useful enough to get us here. Abandoning it means abandoning something that worked.

We’re constantly forced to make decisions before we know which story we’re in. Certainty is a future gift, not a present condition. Perhaps living is less about discovering the true story than maintaining a good enough one, and knowing when to let it go.

Djokovic is caught between anomaly and signal. He doesn’t yet know which.

The next match will revise the story. Or it won’t.

References