Expertise as exception handling

A few months ago, I wrote a post claiming that expertise was doing difficult tasks consistently and Rif challenged me on that. And I’ve been thinking about it over the past few months and have another model I’m going to throw out there: expertise as exception handling.

One example of this is my experience as both a skier and a snowboarder. I am an expert skier, having skied on and off since I was a kid, and an intermediate snowboarder, having picked it up a couple years ago. I spent most of the winter skiing in anticipation of skiing trips to Japan and Baldface. After I got back from Baldface, I got on the snowboard for the first time in a year, and it was interesting to see how my mindset changed. On my skis, I am confident I can handle any terrain and conditions, even people cutting me off while I’m speeding down the hill. On the snowboard, I can comfortably go down any groomed run, no matter how steep, but as soon as the conditions are uneven (e.g. moguls) or off-piste or if people around me do something unexpected, I freak out because I don’t know how to adjust quickly.

Another example comes from volleyball, where a previous post noted how better teammates put me in a better position to succeed. If I’m given a good set, I can hit it down. If a ball is spiked right to me, I can dig it. However, if the ball is a little off, I’m not as reliable. Meanwhile, the expert players can take a ball hit out of their reach, and if they can get one knuckle on the ball, they’ll pop it up perfectly to their partner. And if the set is five feet off the net, they find a way to hit it down anyway.

A last example comes from bridge. I’m an intermediate bridge player at best, but I am subscribed to the bridge players list at Google. The discussions on that list are often around rare hands, where there’s no standard play or bid to cover the situation. The experts on the list debate about how to handle such situations, and many of them have arcane bidding systems to cover all sorts of unusual hands. These are situations I could never figure out how to handle with my basic understanding and the standard bidding system, but they have played enough to figure out how to handle these corner cases.

In all of these examples, the difference between the intermediate player and the expert is that the expert can handle a wider variety of rare situations. The intermediate may be almost as good as the expert the majority of the time, but in unusual situations, the greater experience of the expert allows them to do something when the intermediate is frozen by uncertainty.

This also explains why mere repetition is not enough to acquire expertise. Mastery requires deliberate practice, where one is continually and deliberately testing the edge of one’s ability. By setting up artificial practice situations which don’t come up normally, one gains the ability to handle these exceptional situations whereas repeating the standard situations would not help. It was described once to me as the difference between ten years of experience, and the same year of experience ten times.

I don’t think my advice changes from my last post on expertise, which suggested that deliberate practice was how to gain consistency, but I like this model better. Expertise is learning about how to handle anything that an activity can throw at you, and do it with confidence because you’ve seen it all before. This is also consistent with Gary Klein’s Recognition-Primed Decision Model. Build up your intuition and expertise by getting oneself into more difficult and rare situations so that you can handle them better in the future.

3 thoughts on “Expertise as exception handling

  1. Interesting. I have some musings. By the way, even though I picked on your last post, my picks were in some sense nit-picky, and I certainly agree with your basic hypothesis that “deliberate practice leads to mastery.”

    Exception handling has a certain connotation of rareness. Left open is the question about what the total probability of “rare” situations is. If it’s sufficiently low, then the advantage gained by mastering them, if it doesn’t translate into advantage in more common areas, will be very small. I don’t know anything about bridge, but my understanding is that in chess, for instance, the intermediate player is NOT as good as the expert “the majority of the time.” Perhaps chess is made up of a large collection of “relatively but not very rare” situations, so becoming an expert is just knowing what to do in all of them.

    Your theory has implications for the shapes of learning curves, right? If intermediate skill is acquired by mastering a small number of relatively common situations, and expertise by mastering a much larger number of relatively rare ones, in fields where you can measure this, you expect to see rapid initial progress followed by much slower progress. Which is in fact exactly what you see, supporting the theory. [I think real learning curves also show long plateaus, which this theory says nothing about, but that’s fine.]

    Actually, the example I am most interested in is music. If you consider some amazing blues musicians at the intersection of blues and jazz [say Aretha Franklin or Ray Charles], this model seems to have little validity. It doesn’t feel like what they’re doing is handling a lot of corner cases or exceptions, instead it seems that they are doing something simple just unbelievably well. Put differently, an intermediate performer isn’t as good as these performers in any situation, which I think points at a limit to your theory. Going further, I note that if you hang out with people who studied jazz in an academic setting, your theory seems a lot stronger, because they’ve intellectualized it further. Suddenly, jazz is more about “can you play faster”, “can you play in this or that odd rhythm”, “can you play the song but modulating by a half step every five bars”.

    So thinking about it more, I conjecture that your theory of exceptions is stronger in intellectual domains than in physical/intuitive ones. Thinking about music, an expert will play even a simple Bach prelude more beautifully than an intermediate player. I suspect the same is true for skiing, although I’d be interested in your opinion.


    ps. Did you find you had to do a lot of deliberate practice for skiing, or was starting as a child sufficient?

Leave a Reply

Your email address will not be published. Required fields are marked *