The Score: How to Stop Playing Somebody Else's Game

This may seem bizarre, but striving play is perfectly ordinary. Your friends come over, and you decide to play a silly party game, like charades. There is a paradox at the heart of charades. Most people play the game, not to win, but to break the ice and get goofy. But when you play charades, you have to actually try to win to have fun. A game like charades is fun only when you're moderately invested—when you're genuinely trying to communicate some silly little phrase through gestures, when you're truly frustrated that people aren't getting it. Getting emotionally invested in winning charges the whole experience, and gives it spice.

But most of us don't actually care about winning charades. If you lost but everybody had lots of fun, you probably wouldn't feel disappointed. Because deep in your heart, you already know that winning isn't the point. Having fun is.

* * *

THE STANDARD SCORING SYSTEM OF climbing was good for me, for the first few years. It made me stronger and more graceful than I'd ever been. But over time, it stopped working. I'm not naturally very athletic, and I didn't have the time to train hard. The standard scoring system said there was one goal: Climb more difficult climbs. After a while, I got stuck. The difficulty game just had me throwing myself against gruelingly difficult climbs, and mostly failing. The whole thing turned tedious. It took me a while to remember that I had never really cared about being a great rock climber. I loved rock climbing because it showed me the beauty in my own moving body. I loved it when it changed my experience of the natural world—when it sensitized me to microscopic details in the rock and the landscape.

I eventually found a way to love rock climbing again by tweaking the goals. I stopped being laser-focused on pure difficulty. I looked for middlingly hard but interesting climbs, which pushed me into new kinds of movement. And I didn't declare myself finished with a particular climb after the first time I climbed it. I took the Sherwood path. I aimed for smoothness, flow, and total mastery of the climb. What I most wanted was the feeling of complete organized attunement—and I got that from seeking elegant mastery over moderate climbs, and not from the perpetual quest for more difficult climbs.

In other words, I was turning into a game designer, by creating my own personal version of the game. I was starting to play around with what counted as winning. I was starting to tune the game to my own purposes, to give me that specific mental state of ecstatic organization and flow.

Games tell you what to desire, through the scoring system. But you can also take charge of which games you play, and choose which scoring systems you adopt. You can tweak your games, shift to new ones, jump from ordinary Super Mario Odyssey to speedrunning it. Or you can quit Mario entirely and learn archery, or start a Dungeons & Dragons campaign.

You can use games to explore different ways of thinking and caring. You can become a planning, optimizing, resource-maximizing being. You can become a goofy gesticulating charades being. You can become a twitchy being of reflexes, dodging and whirling and sniping with angelic grace. You can merge yourself into a telepathic group mind.

In a game, you can plunge into an alternate self. And scoring systems are at the center. They make it easier for us to change what we target and what we care about. Scoring systems are an instruction manual for new values.

* * *

ONCE, A FEW YEARS BEFORE the rise of ChatGPT, I was in a room with a bunch of computer scientists who were working on building a creative artificial intelligence—an AI that could make good art. I asked, "So what target data are you using? What counts as success when training the AI?"

One team explained that they were using the Netflix database about engagement hours. Their AI was "successful" when its output most resembled the Netflix shows that got viewers to watch the most hours of TV. The problem is that engagement hours are not the same as good art. Art can be good in so many ways. It can move us; it can make us more sensitive and empathic; it can enrich our understanding of one another. But none of those qualities are necessarily measured by engagement hours.

I said all this. I said that they were training their AI not to make good art, but to make addictive shows. They were optimizing for total engagement hours, and not for the quality of those hours. The AI wranglers responded, "Well, that's the dataset we have. Show us a better dataset for good art and we'll optimize for that instead."

And that's the core problem: We don't have a massive dataset for good art, but we do have a massive dataset for engagement hours. And this is no accident. Some kinds of things are systematically harder to measure because they are more variable, more personal, or more delicate. This is what a lot of this book will be about: why so many of the important things in life seem to consistently defy measurement. They vanish from sight when we insist on using the measurement tools of large-scale institutions and bureaucracy. What's meaningful is intimate and unpredictable; it eludes easy classification. If we let institutional metrics set our values and drive our lives, we end up chasing what's easy to count, and not what's really important.

When we play games—when we are in charge of them—we are touring a whole ecosystem of deeply different rules and scoring systems, choosing and tweaking them to fit our own purposes. But our relationship to metrics is different. Our institutional lives are usually ruled by a very small number of pervasive metrics. We have very little power over which ones dominate our lives, and what they measure. And these metrics have been engineered not to give us a meaningful life but for the convenience of vast bureaucratic information systems.