Yes, good point. Trends are hard to develop with only four games, and a loss or two is easy to write off based on the small sample size. Trends are far clearer with 32 games (bigger sample sizes, by definition, lead to better trends). Thus, Payton's point is probably even more pertinent to baseball.
That said, the real point he was making, I think, is that preseason performance is relevant to at least some degree, as opposed to completely and totally IRrelevant.
Good catch.