Originally Posted by Outshined_One
You know, there is one thing I've been wondering about recently. All the time, people talk about how clutch hitters are, such as the ability to get a base hit with RISP and two outs or whatever. The problem is, most statistical analysis has thus far pointed to either inconclusive results or to the notion of clutch being a myth in hitters.
However, there is a second possibility to this whole notion that I think people have not adequately addressed: the role of the pitcher.
In numerous interviews with pitchers I've read, along with scouting reports and the countless number of games I've seen in my life, a good number of pitchers have had different reactions to situations where they are in a pressuer situation (bases loaded and no outs, for example). I've seen some guys reach back and add a few ticks to their fastball while other guys have seemingly collapsed under the pressure, throwing a "get me over" pitch that ends up in the bleachers.
Is this also something that would be hard to prove or that would be inconclusive? If there is something to it, perhaps this could also be applied to the college pitchers in those situations...
You mean that there's more happening during the game than just the hitter? Do tell...
Terrific post above, by the way. I have some bones to pick with it, but that will require some time for me to ponder over some of what you said and also to write yet another long-winded response.
Thanks and same to you. But before you write that "long-winded response" (because, of course, I never write those) I'd ask you to take another look at the correlations. The reason I ask you to do that is that you've shown a propensity to, IMHO, overestimate the value of Strikeouts when incorporated in an analysis. That's not uncommon. I repeat- that is not uncommon because, intuitively, Strikeouts are bad.
That being said, again- take a look at the correlations because we're talking about how to best project performance. What we've been able to identify during this exercise (with you as a primary contributor) is that K-based metrics have a lower correlation with performance and that which is most correlated to peformance than more valid non-K-based metrics.
That doesn't mean you're wrong, mind you. It actually means that you're mostly right. It also means that you might be able to better project performance by dropping strikeouts from the mix entirely. There's real value to that because it might just allow you to move on to what I consider to be a more sophisticated talent projection iteration (IsoD projection rather than BB/K projection).