A few days ago, I asked how important it was for an effective pitcher to have an effective fastball. I constructed some graphs using the top 50 starting pitchers in baseball in the 2010 season and found that perhaps we overvalue the fastball. Fastball velocity and effectiveness was less of an indicator of pitcher success than wins, and we all know how useless they are.
After posting that, I talked to my good friend Ben who owns an amazing little coffee shop with his wife here in Windsor and we pondered whether or not fastball velocity and effectiveness was more important for a reliever.
Relievers tend to throw fewer types of pitches and therefore throw a higher amount of fastballs which would suggest that it is a tad more important for them then say a starter who likely throws at least three types of pitches and sometimes more.
We also tend to think of successful relievers, and especially closers, as having a dominant fastball that they can use to blow away the competition.
So, is a great fastball more important for relievers, or is there just as little a correlation as for starters?
In the previous analysis I used the top 50 starters, for this analysis I’ll use the top 75 relievers (as there are more relievers on team than starters).
I’ll compare fastball velocity (FBv) and fastball effectiveness (wFB) to Wins Above Replacement (WAR) and park-adjusted Fielding Independent Pitching (xFIP), just as I did before.
If you need a primer on advanced statistics, how they’re calculated and how they’re used, WAR can be found here, FIP is explained hilariously here, and a general overview of many advanced stats can be found here.
The top three relievers in FBv in 2010 were Daniel Bard (97.9 mph), Santiago Casilla (96.6 mph) and Neftali Feliz (96.3 mph). The bottom three were Brad Ziegler (84.3 mph), Darren O’Day (85.7 mph) and Andy Sonnanstine (86.6 mph).
The top three in fastball effectiveness, measured in runs above average (wFB) were Matt Thornton (20.3), Neftali Feliz (19.3) and Hong-Chih Kuo (17.9), while the worst fastballs among relievers belonged to every one’s favourite New Year’s celebrator Alfredo Simon (-11.8), every Jays’ fan’s favourite left-hander Brian Tallet (-11.6) and Chad Qualls (-8.7).
Carlos Marmol led all relievers in WAR at 3.1, while Rafael Betancourt led the league in xFIP at 2.29.
The graphs, as before, will have a series of dots, each representing a pitcher, and a trending line. The steeper the trending line, the more of a correlation there is between the two stats being represented.
Last time, I was unsure if I was interpreting the graphs correctly. This time I know I am as I consulted a friend who has two master’s degrees in advanced math and analytics. I’m going to go ahead and assume he’s right.
In the first graph, I compare xFIP to WAR to give you an idea of what a strong positive correlation should look like. Generally speaking, the higher a pitcher’s WAR rating, the higher their xFIP will be and the better the pitcher they are.
As you can see, the correlation is very strong as the trending line is at a near 45 degree angle.
In the second graph, I compare fastball effectiveness to fastball velocity. As you can see there is still a correlation, but it isn’t nearly as strong; it’s very weak in fact. Just as with starters, how good your fastball is has very little to do with how hard you throw it.
In the next two graphs, I compare fastball velocity to WAR and xFIP. Again, only a slight correlation exists, suggesting that fastball velocity is of little importance to reliever success.
Interestingly enough, when you compare WAR and xFIP to fastball effectiveness, there is a much stronger correlation than there is for starting pitchers; as suggested by the next two graphs.
So there appears to be some credence to the thought that fastball effectiveness is tied to the success of a reliever. Velocity is still mostly uncorrelated, but effectiveness is at least of some importance, more so than a starting pitcher.
To re-conceive these graphs, just take this chart, dim the lights, serve it some brandy or ice wine and coax it to your bedroom after getting it nice and tipsy. Nine months later, POW, baby graphs.