Originally Posted by
RedsManRick
Steel, I appreciate your post. Few people really understand fully the metrics being tossed around (the assumptions, implications, history, etc.), and I consider myself on the periphery of understanding at best.
We usually know enough to make the case or counterpoint, and often that's not a problem. But we should be careful about just how far we run with our conclusions. I think that generally speaking, we'd all be better off using data to make "observations" than "points".
It's interesting to me that a few tenants of analysis are constantly violated in thread after thread. Things like sample size, variance, etc. are routinely given lip service but not wholly informing the points being made. I know it's not fun to come to the conclusion that "Homer Bailey's 2007 MLB performance tells us very little about what he's likely to do in 2008". "I don't know" is usually a conversation ender. We should be willing to do that more than we seem to be.