Application Corner: Choosing between forward and backward problem-solving

The previous article included discussion of how working a problem backward from an artifact to find the cause of a problem can create considerable uncertainty. Nonetheless, this approach is sometimes the best — and often the only — way to obtain information (such as when studying dinosaurs).

Now let us consider the opposite approach of working my (previously) invented problem of determining the cause of the splat on the sidewalk. Suppose that you found witnesses who were in a second-floor office located directly above the splat who saw a person drop a brown egg out of the window a few minutes before the splat was found by a passerby. Now we can work the problem forward and reasonably expect that the egg should cause a splat on the sidewalk immediately below that would be consistent with the observed splat. This can be confirmed by dropping other similar eggs and comparing splats.

The point is that working the problem forward from known events generally creates less uncertainty than working the problem backward from an artifact, understanding that working the problem backward is often the best and only method to address certain problems.

Knowing which approach is appropriate for the problem at hand — or if combinations of the two are warranted — is more an art rather than science. Interestingly, we often find ourselves embarking on one of these approaches without formally thinking about it.

More next month.