Mike Napoli sure started out strong. He hit .327 in May, then .281 in June, with a 1.055 OPS. Since the All Star break, though, he's hit only .148. His power has dropped off as well (.296 SLG), but he's still taking walks, giving him an OBP of .298 which, while bad, is not that far away from what the Molina brothers put up in their good years.
I am confident that Napoli is not a .148 hitter, or anything near it. I know he strikes out a ton, and often looks like he's swinging with his eyes closed (maybe he's listening to coach Kenobi and trying to use the force). During his slump however, he's struck out slightly less than 1/3 of the time - not much different than when he was going good. The difference is in batting average on balls in play (BABIP). In the slump, his BABIP is a mere .176 - and that is just not sustainable.
BABIP, among all the components that make up a player's batting line, needs to be regressed to the mean the most. Walks and strikeouts need much less regression. What this means is we're fairly confident in how much Napoli will walk and whiff, but his BABIP is still up in the air.
I looked at career BABIP for all players with 2000 or more AB since 1980. The average for all players is generally around .295-.300. The top 10 players are between .345 and .360, and include Reggie Jefferson and 9 of the top hitters of the last quarter century (Jeter, Abreu, Ichiro, Boggs, Kruk, Helton, Gwynn, Puckett, Carew). The bottom group is around .240 to .260 and includes low average sluggers like Kingman, Darrell Evans, Gorman Thomas, and Graig Nettles. Even with these players, its likely they weren't true .250 BABIP players - even with long careers you still have to account for some regression.
With Napoli, my projection system has him at a .286 BABIP for next year. Marcel is at .281, and he's at .277 this year (including his April MLE). If that's correct, he should hit between .220 and .235 next year. At this level, with his other skills, he's a very good catcher.
If however, he's one of the real low BABIP guys (and as a contact challenged, slow moving uppercutter that's a real possibility, then he might have a BABIP around .265, and an overall average of .210-.220
Worst case scenario is that he's a historically bad BABIP guy, around .250. At that point he's a .200-.210 hitter. If that's the case, he's still probably our best option at the position, and probably helps you more than Jeff Mathis would hitting .260 (not that Mathis has convinced anyone he can do that yet).