Get updates via e-mail:

« UC Berkeley StarCraft Class, Week 14 (Final Week) | Main | UC Berkeley StarCraft Class, Week 13 »

Malcolm Gladwell and Playing to Win

Gladwell strikes again. He's wonderful at summarizing a field of research, often pulling from boring sources that he makes interesting by synthesizing an overall picture of what's going on. Normally, I would have great appreciation for this, but he tends to do it for fields that I myself have researched, then beats me to the punch, which I find highly annoying. This is of course that's no fault of Gladwell's, only mine. This time though, he covers ground that I have written about for years (, summarizing it in this skillfully written New Yorker article:

How David Beats Goliath

He is talking about the "playing to win" mindset, familiar ground for me. I summarize the mindset this way:

If you want to win, then you should do whatever gives you the best chance of winning...within the rules of the game.

Ten years ago, when people first read my articles about playing to win, many objected that the message was just too obvious to write down. The huge popularity of those articles and enormous amount of debate they spawned show that it's not so obvious after all. Gladwell's article does the same. In basketball, stock market software, war, and computer simulations of battle, Gladwell shows that most people are appalled at playing to win. I mean it's just not done that way! There are social conventions you know, and even though what you're suggesting isn't against any rule, and even though it is highly's just not done!

That idiocy shows up in pretty much every "game" you can think of. It's amazing how bad people are at figuring out that if you want to win, you do what gives you the best chance of winning (within the rules). Gladwell's examples show basketball coaches who reject this because it leads to not learning "basketball skills." Funny, I thought winning was a basketball skill. If your "basketball skills" aren't letting you win against my team's "non-skills," then the last thing you should be doing as a coach is throwing chairs and getting mad at me. Try playing to win instead.

Even more pathetic are the participants in Gladwell's example of the military simulation game. A tournament gave each player $1 trillion to buy fictional military equipment, pitted against the equipment of other payers in a mock battle. One player discovered a very unusual strategy that's "socially horrifying" in the real world, but is highly effective given the rules of the game. Predictably, the people involved are not able to grasp the simple concept of playing to win here. Players are outraged at the winner because the way he played does not fit with their social conventions, even though all agree that he broke no rules. Then the judges chime in with that familiar cry of the scoundrel: "but it broke the spirit of the game!" They said they would not hold the tournament anymore if that player was going to keep entering with outlandish strategies.

These people have managed to misunderstand not only the "playing to win" mindset, but also the concept of game design. If the rules of your game allow players to play in game-breaking ways that make the game less interesting strategically, or that make the game simply feel terrible or unfun, then you should seek to design better rules. Blaming the players for making legal moves that help them win is lunacy. They should have thanked this unusual player for exposing the flaws of their game design so they can create a better ruleset. (Sidenote: thanks to all the SF HD Remix testers and Yomi card game testers who discovered game-breaking stuff that I was able to fix.)

Enjoy Gladwell's article on this subject of playing to win, and remember how difficult it is for people to understand:

If you want to win, then you should do whatever gives you the best chance of winning...within the rules of the game.

Reader Comments

There are no comments for this journal entry. To create a new comment, use the form below.
Comment in the forums
You can post about this article at