I really don't think a guy who relies on control and change of speeds has to worry nearly as much about losing velo as a guy who only knows how to throw, not to pitch.
Yeah, when a guy like Chapman starts losing 3 to 5 mph off his fastball, it's time to hang them up. But all Hendricks needs to do is take another 2 to 3 mph off of his changeup, and he still has that same separation.
And as Beck pointed out, it wasn't the velo that was hurting him when he was battling through his hand issues last year, it was the control. Just like with Arrieta -- Jake lost his fine control, so he took 3 to 4 mph off his velo to try and get the control back. He was a really effective pitcher when he was able to find his control, even with 3 to 4 mph off his velo, but got knocked around when he lost it, and got into long counts, as well as running up his total pitch counts early and thus going only 4 or 5 innings. So, again, it wasn't the velo that bit him *or* Hendricks last year; it was the control.
As for Hendricks getting knocked around because his FB was topping out at 85 and not 87-89 -- are you really trying to tell me that major league hitters, who sit back and wait to just crush fastballs, can't crush 'em at 89 just as easily as at 85? Again, it was control, and being forced to come over the plate with less movement on 3-1 and 3-2 counts, than it was FB velocity, that was getting to Hendricks last year. His change-up was commensurately slower than his fastball right about then, too; it wasn't the speed separation that was hurting him.
To bring this back to Darvish, if velocity was all that was important, then why would a Hall of Famer like Jim Thome discuss Darvish's Eephus pitch, which floats in at about 55 mph, as a "jelly-knee" pitch? If you can spot up all of your pitches, you don't need to throw *any* of them past *anybody*.