Comments on: Can Previous Year Pitch Counts Help Identify Risky Pitchers?

By: Rudy Gamble

Rudy Gamble — Mon, 02 Feb 2009 03:57:28 +0000

@zeppelinkm: No problem, man.

By: zeppelinkm

zeppelinkm — Sun, 01 Feb 2009 23:02:35 +0000

My bad Rudy – there were so many numbers and so many options I didn’t catch that separation. Thanks for pointing it out.

By: Rudy Gamble

Rudy Gamble — Sun, 01 Feb 2009 05:44:47 +0000

@zeppelinkm: Welcome to the boards! The two events were mutually exclusive as I didn't count the FIP increase if there was < 2000 IP. See the line "If a pitcher threw less than 2,000 pitches the next year, we ignored their FIP so as not to add insult to injury (ha!)." @Iamweasel: I hear what you're saying on postseason pitches but I don't think it's a huge factor in this specific analysis. I doubt many pitchers in 2005-2007 fell just under 2700 pitches and the postseason pitches would boost them over the edge to qualify. Same for defining pitchers who dropped off with < 2000 pitches - doubt there are any false inclusions of those who snuck over that number based on postseason. There could be a couple that jump over the +700 mark between years if you factor in postseason. Hard to say. Just don't see this being a big enough factor to invest the hours of digging it would take to total this up.

By: zeppelinkm

zeppelinkm — Sun, 01 Feb 2009 03:34:18 +0000

Great article. This could definitely be the foundation for something a little more econometric.

One correction: “In total, 59 (24%) of the 247 qualified pitcher seasons were followed up by seasons of less than 2,000 pitches and 53 (21%) were followed up with FIP increases above 0.50. That means that about 45% of starting pitchers either pitched significantly less or had a measurable decrease in performance in the following year. ”

You cannot add those two percentages together. This is a common error in statistics. The two events are not “mututally exclusive”, as they like to say. They could have considerable overlap between them, and most likely do as you would (and pointed out) probably find a pretty decent correlation between increases in FIP and a decrease in innings thrown just as a result of the pitcher getting so much worse that they stop using him.

By: Iamweasel

Iamweasel — Sat, 31 Jan 2009 01:47:12 +0000

I enjoyed the article and appreciate the work. You mention yourself that there is some failing with this in that you didn’t add post season pitch counts as well. It would seem to me that these are very important pitches thrown.

I realize that Sabbathia is a horse, but I think he threw enough pitches down the stretch and in the post season that it will have some effect this coming season. Zambrano may have felt similar effects down the stretch this season.

Alternately, it could be effective if you removed pitchers from your analysis that had X amount of post season pitches. Often times these are guys who are simply extrodinary and may help to skew the data anyway.

By: Rudy Gamble

Rudy Gamble — Sat, 31 Jan 2009 00:46:10 +0000

@Andy: Great point. I have been thinking about adding an age component to this although I'm not convinced it'll be a major factor except on the extremes (e.g, Randy Johnson less likely to repeat 2700 pitches than Chad Billingsley). Will keep it in mind...

By: Andy

Andy — Sat, 31 Jan 2009 00:00:55 +0000

There seems to be a long period trend here that you’re not accounting for in that pitchers are not getting any younger. Wouldn’t you expect as a function of age the # of pitchers someone throws to be in general less than the previous year. Interesting study either way!

By: Rudy Gamble

Rudy Gamble — Fri, 30 Jan 2009 17:01:22 +0000

@Nick: Ha...I think you can outargue me on this too. I think the FIP part is less important, though, than the < 2000 IP mark. Not saying the < 2000 IP is more important for fantasy purposes (i'd suggest that the worst case is last year's Harang who pitched but just ineffectively enough that you hold onto him vs. just ditch him for a Nolasco) but it just seems to be easier to predict.

By: Nick

Nick — Fri, 30 Jan 2009 14:58:53 +0000

Yeah, FanGraphs only has FIP at the moment. I think I’ve heard that they’re considering tRA (from StatCorner), which would be super-cool, as in my opinion FIP is good at measuring *what* happened, but some of the other DIPS metrics are better at measuring what *should* have happened.

xFIP is at the Hardball Times, (http://www.hardballtimes.com/thtstats/main/index.php?view=pitching&linesToDisplay=50&qual_filter=1&season_filter0=2008&league_filter0=All&Submit=Submit&orderBy=xFip&direction=ASC&page=1)

and QERA is from Baseball Prospectus (http://www.baseballprospectus.com/glossary/index.php?search=QERA), although it’s not available as a sortable stat. But even calculating it yourself is easy with a spreadsheet, as it only uses three inputs, K%, BB% and GB%.

Since you’re looking at sort of predicting performance here, I think either of these might be slightly better metrics to use. Of course, I’ve been known to argue about pitching metrics longer than most…

By: Rudy Gamble

Rudy Gamble — Fri, 30 Jan 2009 13:40:14 +0000

@Nick: Yes, I went with those two as they seemed to maximize index and sample. Re: FIP, that was the best one available in the FanGraphs data pull I was doing. Didn't see QERA or xFIP.