JABO: A Spring Training Stat That Might Matter
Because there’s a lot of noise, we tend to look past spring training stats. But recent work suggests that using spring training stats can improve projections, and our own Mike Podhorzer worked with Matt Swartz to show that a pitcher’s strikeouts and walks in the spring are meaningful results.
In other words, if we use our small-sample tools, maybe Spring Training really is a little bit like September: expanded rosters, some meaningless games, but still baseball, being played at a high level. The veterans usually play veterans early in the game, and the minor leaguers don’t get the same number of innings or plate appearances to be grouped with the veterans in any statistical mining we might do.
So, that said, let’s look at the starting pitchers, and their strikeouts. These strikeout rates are basically half-way to the stabilization point — the point at which they represent more signal than noise — and they’ve been (mostly) facing the veterans before they leave early, along the right field line.
With a phone full of pictures of pitchers' fingers, strange beers, and his two toddler sons, Eno Sarris can be found at the ballpark or a brewery most days. Read him here, writing about the A's or Giants at The Athletic, or about beer at October. Follow him on Twitter @enosarris if you can handle the sandwiches and inanity.
Seems like you could get percentages by exporting a CSV from this page:
http://www.baseball-reference.com/leagues/MLB/2015-spring-training-pitching.shtml
and then dividing Ks by batters faced (BF).
I downloaded the 2014 spring stats from MLB.com’s sortable spring stats and ran a regression analysis on 2014 spring K% and BB% vs actual 2014 for the 49 pitchers with most innings in the spring who qualified for the ERA title. R squared on K% was .1319, BB% was .1128. Just too small a sample to deduce any trend.
You might want to read the research,
http://www.economist.com/blogs/gametheory/2015/03/baseball-statistics
before commenting.
Maybe you should read it as well. This article is noting pitchers who had their rates change significantly this spring from the prior regular season. It’s much too simplified of an analysis, and I was showing that there is a lack of correlation between these spring stats and the following regular season’s results (looking at just 2014, I know more data would be better). Now, the article you mentioned does a very different analysis where it appends such results to ZIPS forecasts, and shows that the predictive power is enhanced a bit. Sure, more data should help the correlations, this isn’t too surprising, but it’s quite different than what was offered up here.
Shouldn’t the column title on the tables be 2014 K/9?
While it looks like Drew Pomeranz had a “spring breakout” compared to his 2014 regular season numbers, he was actually far more dominant during 2014 Spring Training, amassing a ridiculous 20 strikeouts over 12.2 innings. Some pitchers and hitters (hello perpetual tease Mike Moustakas) just seem to do particularly well in March.