Not worthless but certainly weaker than the W3 stat which uses expected runs rather than actual runs. ERA shows very weak correlation from year to year for a pitcher, FIP shows a much stronger correlation and pulls out variables that don't belong to the pitcher like fielding etc. Its still not the greatest stat though, something like Shandler's xERA is probably the best. Take a look at Joe Blanton once. He pitched better in 2006 than he did in 2005 from pretty much every stat he can control but his ERA was over a run higher that year. The reason his ERA was so high in 2006 was BABIP and LOB% which are things pitchers have little control over and are predictors of his true talent level, that combined with a good bit of luck in BABIP in 2005. The difference is that things like K/9, bb/9, GB% etc are predictable, they correspond year to year and most pitchers don't see big changes in them. ERA is a result that the pitcher only partially controls, it is all over the place and you really cannot predict it year to year. If you want to see the results of his pitching you look at ERA, if you want to predict future results you look at indicators. Thanks for the explanation. Is there anything out there that's reliable other than pitcher abuse points that could be used to predict an arm injury? Some pitchers can just take the abuse, it's not common but some can do it.