I fully disagree. You both can and should predict as much to a worthwhile degree, and both can and should design towards what will more likely be fun.
Since we're disagreeing with each other, I'll share a bit of my thought process with you in hopes of showing you where I'm coming from. When I put any of my designs together, I never ask myself if something will be "fun" because, as I said, fun is a subjective thing. This is what I ask myself:
1) What's the concept of the job I'm trying to build?
2) Are there any precedents to how this class has been implemented?
3) Are there any solid foundations that can be used as a starting point?
3a) Are there any ideas/mechanics from other media/games that can be used as inspiration for this design?
4) How many additional systems (if any) should be attached to the foundation?
5) What do I want the final gameplay to look like?
6)
Do the pieces I've picked come together into a cohesive whole?
This has been the thought process for pretty much every writeup I've done here and elsewhere. Do I hope someone will enjoy the design? Definitely. What I don't do is go into these with the mindset that it's going to be fun or that people are gonna love it; that would be delusional at worst and limit one's receptiveness to feedback at best. And as you may know, I like receiving feedback.
inb4 "lol you're not a developer so your methods don't matter"