This post was provoked by a conversation about Chasen Shreve in which someone I was talking to claimed that rWAR (or Rally WAR) and RA9-WAR were better measures of a reliever's performance than fWAR. By those measures, Shreve has fared better than on the fWAR measure, posting a 1.1 RA9-WAR, campared to a .5 fWAR.
It's my understanding that the primary difference bw FIP-based WAR, such as fWAR, and FDP-based WAR, such as RA9-WAR and rWAR is that the FDP-based WAR calculations don't strip out batted ball results. So pitchers that outperform their FIP-based WAR in a small sample, such as Chasen Shreve, will fare better on the FDP-based WAR measures. This, I imagine, is true for starters as well as relievers. So why is RA9-WAR or rWAR a better tool for measuring relievers than fWAR? Do they have different methods of weighting the leverage index? I believe fWAR assigns relief pitchers some credit, but not full credit for the higher leverage situations in which they pitch. Do rWAR and RA9-WAR do this differently? In other words, is the difference bw Shreve's rWAR and fWAR (as well as the difference bw his RA9-WAR and fWAR) due to the fact that he has outpitched his peripherals or bc of a difference in the way performance of relievers is weighted? Thanks.
