Improvement Ravine

18 October 2006

If you care about what you do, you care about getting better at it. This involves reflecting about how you do things, and trying out new techniques to see if they make you better. Even if other people recommend new techniques, the only way you know if they work for you is by trying them out yourself and seeing if they improve your performance.

The trouble is that improvement, particularly with new techniques isn't linear. Often there is a ravine that opens up when you try a new strategy. The best illustration I remember of this is from Gerald Weinberg's excellent Becoming a Technical Leader. It comes from his habit of playing pinball.

As I slowly improved on my plateau, I began to suspect that I was missing something. Although my scores mounted quickly when three balls were in play, once in a while, when trying to keep all three balls going, I lost all three at once. It happened only about one time out of four, but it meant I had lost one of three precious turns.

I wondered what would happen if I stopped trying to keep all three balls going and concentrated instead on being sure that one of the three balls was retained on the playing field. When I tried this new strategy, my scores immediately fell, into the ravine. In fact, I was playing against a pretty fair kid at the time, and he started beating me. Unable to face defeat, I went back to my old strategy and put him in his place.

Later, though, when I didn't have an audience, I tried the new strategy again. Once again my scores dropped, but I noticed I wasn't losing a turn so frequently, perhaps only once in five times when I played three balls at once. Steadily, with practice, I improved my ability to ignore the other two balls and keep at least one ball on the playing field. I didn't score so heavily each time there were three balls in play, but my overall scores got larger. I also got to play longer for my quarter, which is one of the major objectives of the game.

The point is that when you try a new technique it'll often make you worse, at least initially. You have to work with something unfamiliar, perhaps also unlearn something else that gets in the way. During that period you performance has dipped. Only with perseverance can you get through this and reach a higher plateau.

While you're in the ravine, it's very tempting to back out onto the familiar high ground. When you're learning a new language it's often tempting to stop when you have a task that you know how to do in a more familiar language.

Often your first ravine is the crucial one, because once you've been through that cycle of ravine and higher plateau once, you're familiar with the emotions involved. That memory can help you with the next ravine. After a few cycles the feelings become familiar and less frightening.

There is a dark side - sometimes there isn't a higher plateau. Sometimes a technique just doesn't work, at least for you. The trouble is that you can never be sure if that higher plateau is just one more week away or if there isn't one there.

There no solution to that problem, but that shouldn't make you turn tail at every ravine you come across. Ravines are a common part of improvement and if you want to make any real progress in anything, you have to work your way through a few of them.

Further Reading

This effect is commonly described as an example of a "J curve". A common example is the Satir Change Model (essentially naming that period of chaos and early integration).

Revisions

2020-11-24: added sketch