I'm not clear on your basis for comparison. The original complaint was that the rogue was denied his most basic, goto, rules-as-written attack option because of verisimilitude concerns; whereas the mage and the cleric were not.
Whether the mage is a great simulation (of whatever mages simulate) seems to be a completely different question altogether.
The thesis is a simple and familiar one to us all. See the OP:
However, because Hide is not a supernatural ability, it becomes subject to the dreaded unwritten verisimilitude rules. Rules which seem clear per the RAW are now subject to whatever the group decides is believable for heroic characters in a world full of dragons and wizards to accomplish.
Certainly, it's an understandable tendency. However, it often has the unintentional side effect of neutering martial characters in comparison to spellcasters.
The implication is:
1) Rules construction should attempt (to whatever degree) to model mundane processes or exploits.
2) Supernatural abilities are exempt from this inclination "because magic."
3) Where rules are opaque or absent, there is an "understandable tendency" to have mundane abilities become "subject to the dreaded unwritten versimilitude rules" when GM rulings on mundane action resolution are made.
Hence my curiosity.
We've had thieves with the most ineffectual basic competency in there speciality possible (such that they are a laughingstock at their profession). So ineffectual that they were a laughingstock at their primary shtick and basically disposable with a standard life expectancy of a few sessions due to grotesque failure rate. We've had Fighters without basic competency in things like Athletics and Endurance such that they fail on trivial tasks.
Also, we've had failure within the mundane components of Wizarding. We've had failure to learn spell percentage. We've had the outright inability to cast in armor and we've had Arcane Spellcasting Failure % in various armors. We've had various concentration and OA mechanics. These add up to meaning that the mundane mental process of spell casting is (and should be if 1 - 3 above are universally applicable) quite difficult and the somatic component in spellcasting is intricate/precise/demanding and burdened by harraassment.
We have people losing their minds over DoaM and CaGI because of their perceptions of 1 and 3 above; OMG FIGHTERS ALWAYS HITTING. However, simultaneously in the same ruleset legacy we have all of this "stuff" that says the mundane components of spellcasting (concieving and memorizing formulae, speaking in an opaque, eldritch tongue, and performing the intricate somatic gestures) is hard (presumably more difficult than the 25 % failure rate in freethrows for "good" practitioners) but there is no base % chance to fail to cast a spell (or a failure continuum based on spell level).
Its just a little odd. Its odder still that no one cares about it nor loses their mind with rant after rant decrying OMGHOWCANWIZARDSNEVERFAILSPELLCASTINGWTF!!!? I would think there should be dusk till dawn keyboard mashing and hand-wringing over such perfection in the mundane components of a very difficult craft.