ERA and SIERA differences
Much in the same vain as my BABIP/xBABIP piece, I will look into the pitchers’ luck with ERA and SIERA. SIERA is Skill-Interactive ERA, which basically takes the general points of FIP and adds hidden correlations and park factors to more closely analyze performance. For example, pitchers who strike out a lot of guys generally allow a slightly lower BABIP and HR/FB%. There are way too many such factors to stuff into this post, so you can go here for the primer.
Here are all the Braves pitchers who threw at least 70 innings in 2011.
| ERA | SIERA | |
| Hudson | 3.22 | 3.44 |
| Hanson | 3.60 | 3.23 |
| Jurrjens | 2.96 | 4.43 |
| Beachy | 3.68 | 2.94 |
| Lowe | 5.05 | 3.75 |
| Minor | 4.14 | 3.76 |
| Kimbrel | 2.10 | 1.62 |
| Venters | 1.84 | 2.56 |
| O’Flaherty | 0.98 | 2.78 |
| Martinez | 3.36 | 3.30 |
As expected, O’Flaherty was extremely lucky and Jurrjens was nearly as charmed. Venters and Hudson were a bit lucky too. Hanson, Minor, Beachy, and Kimbrel were unlucky, while Lowe was not as bad as his ERA and reputation pins him. BABIP is a prime cause of differences, and pitchers do not have nearly the control on BABIP that hitters have. The other major difference is HR rate. SIERA uses the xFIP method of adjusting all pitchers to allowing the league-average HR/FB%, plus or minus a couple minor factors. O’Flaherty was lucky to only allow 2 HR, while Hanson was unlucky to allow 17. These rates usually hover around the league average over the years, so expect O’Flaherty to allow more and Hanson less.
As with xBABIP, the key to using SIERA for predicting ERA is finding the true rates for the pitchers. If you don’t think Beachy will continue such a high K rate or you think Venters will reduce his walks, you have to adjust your prediction accordingly. SIERA is great to use if you think a pitcher’s peripherals were in line with his true talent, but small sample sizes and unnatural performance can create a divergence between the two stats.