So two things: One, you’re probably right about the F-35’s engine. It runs too hot (which is two things: one, it makes you more suspectible to IR detection, and two, with the F-35B, the exhaust was hot enough to damage the deck below), its too slow (transonic capability has really been reduced compared to baseline), and its too expensive. (On the other hand: From what we know, it seems to have been pretty reliable so far, which is important). The F-35 should probably have had two engine options all along, but it looks like it may get one from GE: https://www.ainonline.com/aviation-news/defense/2016-07-01/ge-pratt-whitney-win-contracts-next-generation-engine . At the same time, P&W has outlined a few pretty near-term upgrades to the F135 engine that might make a significant difference: http://aviationweek.com/defense/f-35-engine-upgrade-gathers-pace

Two, as the F-35 has moved towards entry into service (for real, not what the Marines did last year), the public has learned more of the tactics that the plane might use, and that’s focused on the F-35s advanced sensor suite and its networking capability, both with other F-35s and down the line, other air assets. Whereas now, if a SAM launches towards an F-16, the pilot reports it and then work is needed to share the potential location of the SAM site, it’s done more or less ‘by hand.’ With the F-35, it happens automatically — so other assets know where the SAM site is.

This makes the kinetic performance of an individual aircraft less important, since you’re essentially moving from “can this plane hit this target” to “who on the team can get a weapon at this target,” even when the team is spread out. This software was plagued by development delays, but it looks like we’re moving towards having those sorted out.

It doesn’t “excuse” poor kinematic performance — obviously, you want a plane that can fly AND do those things, but it does mean that even without it, the F-35 could be a highly effective weapon.