As I’ve mentioned before, we often play Burnout Revenge after lunch at work. Not always—we play board games once a week, we recently got a nice Rock Band setup, and several people have started playing ping pong—but Burnout Revenge continues to be our go-to game.

A month or two ago, we unlocked a Formula 1-style car; it’s faster than any other car we’ve unlocked, it accelerates much faster than the others, and controls surprisingly well given those qualities. So we’ve all switched over to using it: at first, we crashed enough more while driving it that other cars would frequently beat it, but once we put in the time to get used to its handling, it was clearly the way to go.

Its behavior on straightaways (or shallow bends) didn’t require much adjustment. You have to be on your toes a little more, and your ability to avoid cross traffic is reduced, but in general you pick your path and follow it. On curves, though, the car felt rather different: I almost always went wide while turning in the new car, and on right turns in particular I’d end up in dangerous situations.

Slowing down would, of course, be an option, but it’s rarely the recommended option in Burnout: instead, skidding is the best way to deal with high-speed turns. So I decided to work on my skidding skills.

It’s a little odd working on a new technical skill in a game for a few minutes at a time over the course of months. It takes a while for the skill to become ingrained: in fact, I still have to consciously think when approaching curves about how I want to approach them. Otherwise, I would forget to skid at all, going wide instead; and when I started skidding, I would often hold down the skid button too long, with the result that I would turn too far, typically crashing into the near wall.

But when I’m alert, I have an approach that works pretty well. I start skidding, but release the skid button quicker than my instincts would have me do. I’m out of control at this point, so I keep on turning; eventually, though, I regain the ability to steer, and after a bit of fishtailing, I’ll generally end up pointed up where I want to go, maintaining my speed in the process.

I’m not used to intentionally losing control in a video game like this. I’m used to games where I don’t know what good strategy or good tactics is; and I’m used to games where I can’t parse the action quickly enough to respond well. This is different, though: I’m happy with my strategy, and if I’m alert then my ability to parse oncoming traffic is generally acceptable. I’m just consciously making a choice that involves stepping back from detailed control for a bit in the belief that, once I regain control, the range of likely outcomes is one that I’ll be happy with.

And it’s a tactic that cries out for analogization. Take software development, for example. Normally, when developing software, I try to maintain rather tight control of how it’s going: test driven development can be thought of as reifying my constant feeling of control over my software. (And letting me know as soon as possible when that feeling is ill-founded, so I can re-establish my control!) And, of course, when playing Burnout, I normally try to maintain control: if I’m going mostly straight, then I try to plot my course as tightly as possible, I don’t just remove my hands from the controls and let the car meander as it will.

But, in software development as in racing games, there are times when you want to make a sharp turn. And I suspect that, in software development as in racing games, if you try to maintain tight control while doing so, you’ll end up not turning enough, instead going wide and (at best) slowing yourself down or (at worst) being obliterated by oncoming traffic.

Which doesn’t mean that I should relinquish control willy-nilly, even when I’m at a turn. In Burnout, it’s not a good strategy to stay skidding, hoping that you’ll end up in the right place. Instead, I get the best outcomes if I start skidding and then release the skid button as soon as I’ve lost control. If I do that, I’ll start drifting across a range of possible angles to take the turn at; the early angles are too shallow, but my tires wouldn’t stick if I tried to accelerate then anyways. So I sit and watch my car turn more and more; by the time it ends up pointed in a useful direction, it’s willing to respond again (after a bit of fishtailing!) if I try to reassert control.

And the same happens with experiments in software development. If you need to make a sharp turn, then throw yourself into it, and accept that you’ll lose control while doing so. But that doesn’t mean that you have to give up control for a long period of time: instead, make a short, focused experiment, and get ready to zoom off in the correct direction (also with a bit of fishtailing!) once you’ve turned enough to be able to see something promising.

Software development has its straightaways, and it has its curves. Know when you’re in the one, know when you’re in the other, switch your strategy to acknowledge those two situations. But, in either case, leave yourself able to respond to feedback as quickly as possible.

Post Revisions:

This post has not been revised since publication.