First off as I’m sure RJ knows all this is totally meaningless unless RJ provides us with the actual regression numbers.

Secondly and much more importantly, the problem with WHIP is that it ignores the fact that not all hits are singles and that not all outs are the same. A walk is preferred to a double, a double is preferred to a home run, a strike out is preferred to a ground ball to the right side that moves a runner, and a ground ball is preferred to a fly ball with a runner on third and less than 2 outs. Does this analysis take into account any of those very obvious facts? No it doesn’t.

Pitchers that have high walk rates, that miss a lot of bats, and that do not get hit very hard generally do very good in MLB. Edinson Volquez is a prime example of that this year. We would all like those types of pitchers to cut down on their walks, but in the long run they will do better than a strike thrower that gets hard all around the ballpark. This analysis just ignores that entire component.

I actually did some research like this once before and strike out and home run rate need to be included, too. When you include WHIP, K-rate, and HR-rate in unison, you can get a pretty solid estimation of what a pitchers ERA should be.

http://baseballanalysts.com/archives/2008/03/categorizing_pi_3.php

I think it gets at the factors that make a pitcher successful better than most other approaches.

Note how Lederer ends it:

“When it comes to evaluating pitchers, I would rather know their strikeout and groundball rates than their ERA. Throw in walk rates and you have almost everything you need to know about a pitcher. Focusing on these components gives one a much more comprehensive understanding of a pitcher’s upside and downside than looking at a single metric such as ERA.”

Looking at the list of pitchers in each category can lead to opposite interpretations. Either the claim that high K & GB rates indicate a likelihood of success or that there are a few failures in that category and some successes in the low K & GB rate sector and therefore it is not a useful analytical tool.

I think the second view is mistaken for 3 reasons:
1. You have to look at how close to the cutoff lines some of the pitchers are. For example, Shields falls into the second category, but only because his GB rate is ever so slightly below the cutoff. He really is in the first category, especially when his BB rate is included.

2. The chart is really not projecting. It is looking at past performance and grouping pitchers based on two factors. In reviewing the list, it seems clear that the greater number of successful pitchers in 2007 fell into the first 2 categories. The only inference we can draw from that is “IF a pitcher can sustain those ratios he has a good chance of sustaining his success.”

Someone like Kyle Lohse, for example, can have a good year despite his low K & GB rates, but can he sustain that success? If any other part of Zito’s game slips, can he remain a good pitcher with those rates?

From the Edwin discussion, that should lend credibility to both positions. On the one hand, he lands in the top group, and given that he does not usually give up lots of homers, that bodes well. On the other hand, considering the additional concern about walks and a sliding K rate, that provides ammunition for those who are dubious. And of course, we should, in my view, consider other observations as well, especially when the data has some contradictions.

3. One cannot argue from exceptions. The chart is not intended to demonstrate a perfect correlation between K & GB rates and success. What it shows is a pattern, and referring to that pattern is the starting point for projecting, not the final answer. It provides one more data point in the analysis, I think a very useful one, but anyone who uses the exceptions as an argument is misrepresenting what the chart is all about.

Incidentally, with a nod to another thread, anyone who reads Lederer regularly knows that he not only attends games but is an astute observer and devoted admirer of the history and beauty of the game. And that he also is well aware of the complexity of baseball, as his last 2 paragraphs in this article indicate, framing his statistical analysis within a broader appreciation of this complexity, just as the moderators of this site do.