I think he's asking if we should value people with a high "home run" percentage in their output or those with a lot of output.
I think there is an easy way out of this and that is to look at citations, not raw output. After all a rebound is a rebound is a rebound, but not all articles are created equal!
It's possible to have a very long vita without making any discernible impact on the profession and (this should go without saying, but sometimes doesn't) a very short vita is not necessarily a sign of high quality.
So on the citation front, it's no contest. According to Google Scholar, Ronald Coase has over 31,000 cites to his two classic articles. Nolan Ryan only has 9!!
A more serious comparison might be to someone like Alberto Alesina. When I look at the first four pages (40 articles) of his cited articles in Google Scholar, I count roughly 27,000 cites, while his two most cited articles ring up a total of around 4900 cites.
I am somewhat already agreeing with Tyler in that I'm using total citations rather than citations per author here (i.e. not penalizing Alberto for having a lot of co-authored articles).
Hard to say really in this case, but I think total citations are the metric to compare scholars with different vita lengths.