Games, constraints, and the humanly possible

Posted by ekr on 26 Feb 2022

On Friday's Ezra Klein show, Ezra interviews philosopher C. Thi Nguyen on the topic of games. Nguyen provides an interesting definition of a game (btw, thanks to the Times for providing transcripts so I didn't have to type all this in):

What’s interesting about games for him [Bernard Suits —EKR] is that you have this thing— the finish line—but it doesn’t count unless you did it under specified constraints. It doesn’t count unless you follow a particular path, unless you did it for a marathon on your own feet instead of a bicycle or a taxi. And the fact that the activity would lose its value if you didn’t do it in the specified, inefficient, constrained way, that, for Suits, points the way to what games really are.

And the way I think of them sometimes, after Suits, is that games are constraint-constituted activities. Does that make sense? That what it is to run a race is to do it inside a certain set of constraints. Like what it is to climb a rock in rock climbing is to do it with your hands and feet and not a jetpack, or a chain, or a helicopter. So whatever is valuable about games has to be in the fact that they’re constructed struggles.

There's a lot here that's true. To take the example of the marathon, not only is it rarely the case that running is the most efficient way to get from point A to point B. In fact, it's not even the most efficient way allowed in marathons. Many major races have a wheelchair division and the wheelchair athletes are much faster than the runners. For instance, in the 2021 Chicago Marathon, the men's winner came through in 2:06:12 and the men's wheelchair winner came through in 1:29:07.^[1] Moreover, plenty of marathons actually start and end in the same place (and don't even get me started about 100 mile ultras run on a quarter mile track).^[2]

It's interesting that Nguyen uses the example of rock climbing, as mountaineering and rock climbing are both sports that started out much less arbitrary than they are now and gradually became more arbitrary and rule bound. Mountain climbing is perhaps the purest example here: the tallest mountains are essentially inaccessible by any means other than actually climbing them on foot it's just barely possible to fly a helicopter to the top of Everest, but as far as I know it's been done exactly once, so as a practical matter if you want to get to the top you have to walk up.

That doesn't mean that there aren't arbitrary rules, but the interesting thing is how they have grown over time. Initially, it was just a challenge to climb Everest at all and it took about 70 years of more-or-less serious attempts before Tenzing Norgay and Edmund Hillary's first ascent 1953. At the time, this was an incredible achievement and people took any advantage they could get including supplemental oxygen, teams of porters, etc. After a while, though techniques developed and the mountain was better understood and so people started to find ways to make it harder, for instance by climbing without supplemental oxygen (Reinhold Meissner and Peter Habeler in 1978), solo, without oxygen (Meissner again in 1980), alpine style, etc. Another complication is that there are different routes mountains, some harder than others, so it might be a challenge to do a new route even if you've gotten to the top before. At this point, just getting to the summit by any means necessary is difficult but doable by ordinary people even without large amounts of mountaineering experience (see Krakauer's Into Thin Air for more on this).

The story is similar with rock climbing: the first ascents of a number of the big wall climbs like Half Dome or El Capitan were were done "aided" which means that you use your protection (back in those days, this meant bolts and pitons) for support. Here too, initially it was a challenge just to get to the top,^[3] but after a while it became clear that if you were willing to spend enough time and drill enough bolts you could get up just about anything and so people started thinking about free climbing (using ropes for safety but not support) (El Capitan's Salathe Wall by Skinner and Piana in 1988 and The Nose by Lynn Hill in 1993), or free soloing (no rope) (Alex Honnold up Freerider in 2017). Here too, this is a story of technology (primarily sticky rubber shoes and better mechanisms for attaching your protection to rocks) and better technique.

Under the definition being offered by Nguyen—and as I understand it, Suits—when the first people went up Everest it wasn't a game, but as soon as it became relatively achievable by ordinary people and the challenge became to handicap yourself by doing it without oxygen, then it became a game. This might be right, but on the other hand it seems to me to that Tenzing Norgay and Edmund Hillary's first ascent in 1953 and Meissner's 1980 solo ascent without oxygen are a lot more similar than they are different in a way that Nguyen's definition tends to erase. You could of course respond that the original first ascent was a game—after all, isn't Everest arbitrary?—but then I think you've just redefined almost any challenge to be a game.

I think that the common thread between all these challenges is something Nguyen hints at later, which is that games can be designed to be just difficult enough that you can do them, but only barely:

But in games, because the game designer manipulates what you want to do and the abilities and the obstacles, the game designer can create harmonious action. They can create these possibilities where you’re— what you need to do— the obstacles you face and your abilities just match perfectly.

...

And in games, for once in your life, you know exactly what you’re doing and you know exactly that you can do it. And then you have just the right amount of ability to do it.

This feels a lot closer to me as a description of the essence of the kind of challenge that mountain climbing or running a two hour marathon presents, namely that they are at the very limit of human capability. When people first tried to climb Everest or El Capitan (or the moon!), nobody knew if it was possible, so the challenge was just to do it at all. But then once it was achieved, then the limit of capability shifted and people wanted something harder, which could either mean trying something harder like K2 (or Mars!) or adding new constraints to make it harder, like climbing without oxygen.

What I'm saying is that the core experience here is doing something that is just barely possible for you. Of course at some level, "something" is arbitrary and once you've run a marathon "just barely possible" can be "do it slightly faster" but humans like things that feel like natural anchor points even if they are ultimately arbitrary, hence the appeal of the 40 minute 10K or climbing 5.12 for the amateur or the four minute mile or 2 hour marathon for the professional.^[4] I think this is also behind the appeal of climbing without oxygen, in that it feels like a clear dividing line. From this angle, the nice thing about games is that the games designer gets to set the conditions so that they are at the right level, but those arbitrary tuning parameters are buried inside the rules of the game so that finishing the game becomes a concrete anchor that people can focus on.

Of course, this is all easier said than done, especially if you want everyone to do the same task. Human capabilities vary widely and a challenge that is just barely at the limit of someone's capabilities (say running 100 miles) is easy for others. This is something that Gary Cantrell, the creator of the Barkley Marathons talks about, namely that it's easy to make a race that's so hard that nobody can do, but what's hard is making a race that almost nobody can do. But of course that's exactly what makes people want to attempt it.