Let’s remember what sample sizes are for. They are for testing hypotheses, whether it be that the data are representative of something (normality, for example), that two things differ, or other inquiries.

Now, sample size influences your confidence that what you are seeing is representative, or that a difference is really not due to chance, and so forth. But other things besides sample size also influence that confidence. For example, you might have a small absolute sample size but a large proportion of the overall population (e.g., Adam Kennedy has only made 8 errors all season, but he has only played in 8 games). Or you might have constructed your argument ahead of time (a one-tailed test) and need less sample to have confidence that the data support or don’t support it (Rick Ankiel strike out once per Major League game all season….and true to form, he has one in each of his first 20 games). Or - and this is lost on a lot of folks - you have so little variability in your data that you need less sample to support your hypothesis (PJ Walters always goes at least 6 innings and never allows less than 2 runs - after 8 such starts, you might have better confidence in his success than after 12 highly variable starts that yield the same peripherals).

What we are really discussing with Walters is whether his body of work in AA is (yet) a reasonable predictor of how he will perform the rest of the season (and eventually in the majors). And Azru is right that 3 starts is a small sample size…but Azru also cannot dismiss the optimism of other people if that small sample size comes with some other attributes. I see this “sample size caveat” from a lot of folks who got their stats training mainly from Sabermetrics, and they make me both happy (at increasing quantitative sophistacation in baseball) and sad (because real statisticians often make fun of sabermetricians) at the same time.

So Azru, we get that sample size matters - but, like the saying goes, it’s only a part of the package, and one size does not fit all. Go, PJ!

I think the second portion is referring to what I said earlier and it’s less a matter of not having any scouting reports than the pitcher being new to the league and the team not having their own scouting reports.

I have heard that his change is well above average and of course his fastball is below average. My point is that we often relate major league stuff with fastball velocity, which doesn’t matter as much as pitch movement. Really velocity for a pitcher can be related to one of the 5 tools for hitters–it is just a tool not a skill…Control is a skill…Deception is a skill.

Second, someone said that there are no scouting reports on Walters. Every team in the FSL had one and they shared that info with their TL affliates.

It is a rule that you can’t trade a prospect until he has been in your ogranization for a year. So Jesse Tood will not be the PTBNL since he was just recently drafted and signed. Thought that would help!