Many RPGs work in a somewhat confused way- they have a small set of core mechanical rules that are applied generally, and then a whole bunch of exceptions to those rules.This is a long standing tradition in RPGs dating back to original D&D and before that to wargames where you had the basic mechanics of the game and then 'Chrome' rules added on top for flavor.
In the days of wargames, designers were careful to keep the amount of Chrome limited. Chrome was to be little highlights to make the core more interesting, i.e. basically the same purpose chrome on a car design served (and hence the name). It wasn't suppose to be key to the game,as the game was to fun and involving itself. The law was simple: Thou shall not use Chrome to support function.
RPGs changed this, basically making Chrome the game itself. Some of this is unavoidable, if you're to have a extensive list of spells for example (as opposed to a spell design system) you end up with more pages describing that 'Chrome' than you do core mechanics. Besides, the very birth of D&D was in Chrome added to the end of a wargame (Chainmail by many accounts).
D&D however took things much farther, with individual rules for individual classes. And then things got worse with various skill or weapon packages that modified things even further. And then on to Feats. As of today, the list of exceptions far outweigh both in number and importance any core values such as stats or basic resolution mechanics.
The reason for this is rather simple, the basic game is... boring. Always has been. So lacking any real interest in themselves, the rules needed Chrome to inspire at all. And over time, it just plied higher and deeper. It becomes increasing difficult to know how any combination of Chrome will react to another combination. Both the GM and Players ability to judge and thus control results are degraded.
Attempts to avoid this include the OSR method of burying one's head in the sand, i.e. go back to an old edition (with less Chrome) and spend as little time as possible in combat because it's now boring. I suppose a victory dance of some type helps one deal with this self-defeating option.
Another example of this type of design can be found in the 40K RPG products like Only War and Dark Heresey. Limited to a closed ended d100 system, things are only vaguely interesting when effective chances for success hover around the middle of the random number span- say 30-70 percent chances.
This results in a game that represents the mighty genetically enhanced superhuman Space Marines with stats only 10 points higher than normal men... so maybe a base 58% to hit instead of 48%. The mind is underwhelmed by such power. Instead it's 'Talents" (i.e. special abilities, i.e. Chrome) that's suppose to make up the difference. There are more Talents in the game than all the core rules (stats and skills) combined. And once again, the core rules are by themselves, rather boring to play.
I believe that Design by Exception is by its nature poor design. Instead the core mechanics themselves should scale from the lowest creature to the mightiest. Chrome should be just that, little things that makes those core mechanics pop just a bit more, not be something that buries them.
To pull this off those core mechanics need to be interesting of and by themselves. A fact that is basically impossible individually, but quite obtainable in combination. Consider the individual chess pieces, by themselves their movement rules are simple and boring, but in combination against the board the result is vastly more engaging.
I don't see any modern designer attempting such outcomes. Raised on endless D&D editions, they are only recycling failed concepts while wondering why their market is shrinking. They have only themselves to blame.
Pity me, for I am weak
1 week ago