Good link, Craig! There’s a great piece about early box scores over at Walk Like a Sabermatrician:

http://walksaber.blogspot.com/2009/04/keeping-score-with-henry-chadwick.html

It mentions the evolution aspect of the modern box score design. He then goes into a (valuable) digression about how things might have been different if the early box score designers had made just one or two decisions differently. Here’s a good excerpt:

I wonder if or how the development of conventional statistical wisdom would have been altered had the box score column remained “outs” instead of “at bats”. The move to at bats obscured the fact that at bats = outs + hits. By hiding the outs in the broader category of “at bats”, which are easily dismissed as simply a playing time measure, they were hidden from the view of casual fans.

Of course, unless you also consider walks, the result might be the same; valuing hits in relationship to outs is ultimately the same as valuing them in relationship to at bats. But if a hits/outs ratio was used rather than a hits/at bats ratio, there would be many more people who would inherently accept that outs are the proper denominator for an overall offensive rate stat. One of the big steps in converting a lot of people to a sabermetric outlook is to convince them that using outs and not just total opportunities (AB or PA) results in the best model of true offensive value.

I don’t necessarily think that that would’ve been the best change, but I definitely agree that it may have changed the way we think about the batter-pitcher matchup. It’s an interesting thought exercise, at least.