May 1, 2009

Fantasy Beat

Tweaking QuikERA

by Marc Normandin

Printer-friendly

With one month of the season behind us, it's time to take a look at one of my favorite statistics for pitcher performance. QuikERA, or QERA, was developed by Nate Silver a few years back, and is explained in this article as a way of separating skill from luck in order to analyze playoff pitchers. Here's the definition, straight from the BP Glossary:

QuikERA (QERA), which estimates what a pitcher's ERA should be based solely on his strikeout rate, walk rate, and GB/FB ratio. These three components-K rate, BB rate, GB/FB-stabilize very quickly, and they have the strongest predictive relationship with a pitcher's ERA going forward. What's more, they are not very dependent on park effects, allowing us to make reasonable comparisons of pitchers across different teams.

Since these components stabilize quickly relative to other pitcher stats, we can use QERA almost immediately during the season in order to get a read on who is over their head, and who deserves to get a second look or accumulate more innings to straighten out their actual ERA. This year, I want to try something a little different, though; QERA normally uses BB rate, which encompasses both unintentional and intentional walks. Since the components we are accounting for are under the pitcher's control, and intentional walks are issued by the manager, I've made a slight tweak to the formula for today's article, and changed BB rate to UBB rate. The formula is now:

QERA =(2.69 + K% × (-3.4) + UBB% × 3.88 + GB% × (-0.66))²

This early in the season, that doesn't mean much for changes, but let's take a look at what a few pitchers' QERA look like using straight BB rate versus UBB rate:

Pitcher         ERA  QERA(BB)  QERA(UBB)
Derek Lowe     3.10    4.03      3.66
Ricky Nolasco  6.92    4.38      3.99
Johan Santana  1.10    2.63      2.44
Ian Snell      3.72    5.93      5.63
Gavin Floyd    5.52    5.01      4.76

Those are the five starting pitchers with at least two intentional walks in April. Derek Lowe and Ricky Nolasco see the most significant changes, both of about half a run, while Santana, Snell, and Floyd all see smaller differences. If the point is to be more accurate though, then I feel like this is a better representation of what the pitcher should be doing than using regular BB rate, which for whatever reason also encompasses intentional walks. As more innings add up and more intentional walks hit the board, there will be more players that see a third of a run or half of a run shaved off of their QERA.

With that out of the way, let's see what we can do with QERA. Here are the five starters (minimum 20 innings pitched) who are outperforming their QERA by the most significant amount:

Pitcher         K/9  UBB/9   GB%    ERA   QERA    Dif.
Jair Jurrjens   4.6   4.0   40.2   1.72   5.88   +4.16
Tim Wakefield   5.3   3.7   34.5   1.86   5.61   +3.75
Dallas Braden   5.4   3.3   35.7   2.10   5.34   +3.24
Matt Cain       6.2   3.1   35.9   2.08   4.86   +2.78
Daniel Cabrera  3.3   6.3   45.7   4.44   7.21   +2.77

Jurrjens has been lucky thus far, as he lost two strikeouts per nine from last year while increasing his walk rate, but has still managed to strand an above-average number of baserunners (the timing of your outs can be everything-ask Dice-K about that one). Jurrjens has also posted a G/F ratio of 0.8, which is disappointing given his grounder tendencies last season (1.9). He's throwing his slider a little more often, and throwing fewer fastballs (with a loss of about one mph average velocity), though I'm not sure that this would account for the significant dip in G/F ratio.

I do, however, think that we can see why Jurrjens isn't inducing grounders (or picking up as many punchouts) as in 2008. Last season, Jurrjens got hitters to swing at 27 percent of his offerings outside of the strike zone, which was above the average of 25.4 percent. This season, hitters are swinging at just 18.3 percent of the pitches out of the zone. In addition, Jurrjens threw first-pitch strikes nearly 62 percent of the time in 2008, which was a great way to get ahead of hitters and set himself up to control the plate appearance. This year, he's at 56.8 percent, which is roughly one percentage point below the league average, rather than comfortably above. Pitching from behind in the count more often makes things problematic, and also gives the impression that the pitches that the hitters are laying off of that are outside of the zone are the pitches that should be turning into grounders on contact.

If anyone has picked up Daniel Cabrera because they're desperate for either innings or ERA, now might be the time to wake up from your dream. Cabrera has never become the ground ball-machine we thought he could be based both on his stuff and his 2005 season (1.8 G/F, 53 percent GB%) and though he has actually been unlucky this year, stranding just 60.7 percent of his baserunners, he's also managed to get away with handing out free passes to nearly twice as many hitters as he has whiffed. His fastball velocity is down-again-this time from the 92 mph range that worried me last season (Cabrera used to hit the mid 90s with regularity) to just 90, with plenty of fastballs coming under that number. Even his slider, which used to be nearly as quick as his current fastball, is coming in under 80 mph regularly, and he's using it more often than he has in the past despite this. Now, instead of a power pitcher with high walk totals due to control issues, you have a guy with the fastball of a control artist, sans the control.

*- Dallas Braden pitched decently enough last season, posting a 4.14 ERA over 71 1/3 innings, so it's easy to assume that he's improved this year based on his April performance. His ERA is 2.10, and he's cut his homer rates slightly from last season, but there are two problems with this. First of all, his 4.14 ERA was below where it should have been, and secondly, he has stranded nearly 90 percent of his baserunners. Based on his peripherals, he has been nowhere near as solid as his ERA suggests, and has been just about the same as last season. The one change he has made that may allow him to improve is the addition of a cut fastball to his repertoire; though he hasn't thrown it much overall (less than six percent of the time), he has had his pop-up rate increase at the same time his home runs have dipped. It's too early to tell how closely related all of this is, but it's something to keep an eye on if you're desperate for some pitching help (or just now realized you need to replace Daniel Cabrera).

QERA serves as a solid base from which further analysis can be made. It lets you quickly determine who might be performing above their heads and who may need a few more innings under their belts to show how good they actually are, and when used in conjunction with more advanced information, can help you make informed decisions regarding your team's roster, be it real or fantasy. QERA is not something we currently have in our Sortable Statistics database though, so I've created a spreadsheet for you to download.

Here is how you access the information. Follow this link to the QERA Report, which was created using the custom sortable statistics. From there, you can highlight the 100 names that show up (along with the top row of headers for stat names) and paste it into the spreadsheet (right-click, Paste Special, Paste as text). You can also access the data by running a Data Query in Excel using the URL above, but for those who don't normally use Excel for that, the above explanation offers a simple alternative.