ERA+ together with innings pitched is probably good enough. It's true that it doesn't account for inherited runners, but, first, after enough innings pitched, that difference will usually come out in the wash, and, second, in extreme cases, you can do a little mental adjustment. The adjustment shouldn't be that big - pitching well is more important than holding runners.
There are many (many many many) problems with the save statistic, but one of them is that it doesn't tell you anything about middle relievers, who are sometimes quite good. An ERA-based metric won't ignore them.
Whether to pay attention to WPA depends on what you want. If you want to know how skilled a pitcher is then WPA just introduces noise that you don't want. A pitcher who gives up a meatball with the bases empty is just as bad of a pitcher as one who gives up a meatball with the bases loaded. If you want to tell the story of a team, or a player, or a pennant race, then it's useful, because it'll tell you who swung the odds the most (even if there was a lot of randomness involved).
(ERA+ takes ERA, adjusts it for the park in which the player was pitching, and compares it to league average, which is automatically set at 100. Higher is better. The normalizations allow for cross-era comparisons.)
|