2 years ago Gobert didn't help his team in +/- either and carried a massive -6.0 ORTG +/-.
There's nothing magically bad about Noel's offense anymore than any other defensive center, and in fact his faceup game opens possibilities for increased utility, certainly over the Ben Wallace's of the world. He's young, and he's been playing for a chaotic franchise where every number has to be taken with a grain of salt.
My general guess is that we'd like a massive player of Marc Gasol's ilk in there. Vlade was looking at Pau even before Joerger got here. but those guys are a rare bird, and maybe an aging breed. Though of Gortat too, but another older player with a big contract. So our realistic available choices are really about whether we bring in a defender with adequate offense or an offensive player with adequate defense.
Your Gobert argument is a classic use of using on/off without including context. On/Off does not do the best job at accounting for teammates. If the replacement sub for Player A is much weaker than the replacement sub for Player B, Player A has an advantage to receive a better on/off. Considering Noel's replacements the last 2 years have been players like Henry Sims & a rookie Jahil Okafor, it should allow Noel a much easier opportunity to have a positive On/Off, but despite the favorable situation to post a good On/Off, Noel fell short. Meanwhile, Gobert has a guy of the name of Favors to replace him along with Withey & Kanter in the mix. It's pretty fair to say that Gobert's "competition" was much better than Noel's. So comparing their On/Off numbers without that context is leaving you susceptible to misinterpreting the data.
As I have already pointed out before, RAPM does a good job at accounting for what On/Off doesn't (teammates). It's able to factor in a teammates skill/talent to provide a more accurate read. Having that knowledge, what would you expect? Considering Gobert's replacements are much better than Noel, you would expect Gobert's RAPM to be higher while you would expect Noel's RAPM to be lower. Let's take a look...
Gobert
ORAPM: 0.03 (159th in the league)
DRAPM: 2.15 (15th in the league)
RAPM: 2.18 (39th in the league)
Noel
ORAPM: -2.79 (470th in the league)
DRAPM: 0.28 (173rd in the league)
RAPM: -2.51 (433rd in the league)
Would you look at that? The stats fall right in line with what you would expect when adjusted for teammates. As you can see, your comparison of Noel to Gobert is feeble at best. Gobert's impact is far and away better than Noel.
No way should we be investing our cap in a player with such little impact. You want to stick to the potential argument? Fine. Perhaps one day Noel will be able to help his team win, but considering how far away he is, it's really not worth it to pay him while also giving up a 6th pick in his sophomore year.
To your very last point about bringing in a defender with adequate offense or an offensive player with adequate defense, I agree. However, Noel fits neither description at this time, and considering we're trying to be competitive immediately to help persuade Cousins to stay, locking up a lot of cap space in a player who has not proved that he can contribute to wins is a very, very, very risky move. Some might even consider it reckless.