Protective skirmishers truly are asinine. I'm not sure how a person in MP that isn't using inca can stop them from choking. They are odds-on vs chariots and even fight well vs axes with shock, obliterating spears. Of course, an opponent will be lucky to get any metal at all...
Archer rushing with PRO under restricted leaders makes no sense unless you are using CKN which are of course fairly solid units if you prioritize getting them (and since china doesn't start with fishing, they can bulb machinery with a GS).
Counter promos do in fact matter for gunpowder though. Only AGG and PRO can take counter promos on drafted units (with theo) and cover, pinch and shock can all be significant choices at 2 xp. AGG is still better due to higher odds + ability to get amphibious at 5 xp in that role, but PRO is a significant step from standard gunpowder regardless and the CG can actually matter in those situations too.
CG III longbows are ridiculous and will maul the AI. On a hill, a handful of them will slaughter dozens of units belonging to the AI. Of course, this is a marginal luxury given that stock CG II longbows hold hill cities pretty well too.
And that's just the problem. PRO isn't worthless. You'd take it as a freebie 3rd trait on anybody who doesn't have it, and you'd take it over nothing. It doesn't HARM you to have it and does some things that can actually contribute to victory. The problem is that compared to other traits, they're marginal bonuses. Walls/castles for "espionage econ"? Really? Maybe with no tech trades, but otherwise going EP so early is awkward and slow unless you sell out on gwall and really plan...and econ comes up awfully fast if you research the best tech trade bait. Walls aren't useless but you build them much more rarely than other buildings and the

savings aren't as high. When you wait until engineering to build walls + castles, the

savings are both later and fewer than some of the other traits...ouch.
Of course, pre-patch 3.19 where failaxis introduced a bug in overflow, protective was probably closing in on a top trait if you took advantage of the wall whip + chop gold overflow. Arguably, that was "too strong" although nobody really provided a clear basis for that assumption. It's not like failaxis tested it. If you want to use the "seems cheap" or "immersion breaking" excuses, then why do vassal states, AI resolution logic, and quite a few other esoteric and immersion breaking features/rules survive that patch untouched? No, some people whined to failaxis and just like the idiots who decided that shotguns were too strong in mw3 with the extended mags "glitch" (ridiculous that a weapon is outclassed at the only range it is effective. That nerf is like nerfing ironclads in civ IV), they knee-jerk nerfed protective. Yeah. Read that again and iron it into your mind:
failaxis nerfed the protective trait, which was regarded by most as being weakest, and didn't give it any compensating buffs

.
Pre 3.19, protective WAS actually useful and good though. Often very much so.