CoolBreeze44 wrote:Stats are what they are, numerical facts or pieces of data drawn from a specific sample of larger sets of numerical data. Stats themselves really can't be noisy. What becomes noisy is the way the statistics are interpreted within a more comprehensive analysis. Trust me, you can find a statistic to back up just about any point you want to make. Those interpretations can be more or less noisy depending on the quality of the comprehensive analysis being performed. If you're going to say Towns is a great rebounder because he averages 12 per game, to me that is very noisy. Who does he play with? What position does he guard? What kind of shooting percentage does the defense cause? How many minutes does he play? However, if you say KAT is a great rebounder because through film analysis he consistently puts a body on his man, he's a quick jumper, has great hands, and averages 12 boards per game, the stat has a lot less noise. I know it's a pet peeve of mine, but statistics are not arguments in themselves.
As it relates to individual metrics - like RPM - I agree that one can not make an argument by that metric alone without any other information or observation. But there is such a concept as "preponderance of evidence" and I think that applies and stats are a really big part of it.
Let's start with Zach's poor DRPM. The first question I'd ask is, do I ever actually observe him playing poor defense? The answer is yes, I do. Anecdotally, he appears to get beat on back door cuts a lot and he tends to over-react to fakes and change of direction moves such that he has a difficult time recovering into a good defensive position. But that's my flawed and biased brain watching - I don't have a perfect memory of every defensive possession he's played in!
The second question I'd ask is, are there other individual stats that indicate Zach might be a poor defender? The answer again is yes. He gets very few deflections (steals and blocks) according to box score data and he is a mediocre defensive rebounder.
The third question I would ask is how does the team perform defensively when he is on the court? The answer is, very poorly!
So now I have multiple data points to form an argument. And the aggregate evidence doesn't really contradict what the original stat (RPM in this case) told us.
There may be other players where the data is much more conflicting, so perhaps Zach was too easy of an example!