Regarding
@Charlaquin @Helldritch @Crimson Longinus and
@Bacon Bits (and others) next level Ability scores tangent:
The problem is that common sense and verisimilitude are at odds with gameplay.
10 foot minotaurs stomp 3 foot halflings for sport.
But, if that's the case, a game where minotaurs and halflings are both featured as frontline fighter types isn't going to be fun for most people. (Yes, yes, half-giants in Darksun, you say. I know. But please acknowledge the general principle, even if you are immune to it.)
This thread's arguments about Ability scores are a proxy for said disconnect. If a minotaur is
clearly stronger than a halfling, shouldn't that be represented in some significant way by the minotaur's Strength score. Well... no. With bounded accuracy, Strength ability score is a heavily abstracted game concept that makes only the barest effort to conceptually simulate strength. Trying to get Strength Ability scores to represent "reality" would break the game.
And that's fine. At bottom, the game needs to work well as a game, not as a simulation.
The degree of abstraction just needs to be better communicated by the PHB. (See also every argument about Hit Points ever.)
So... to contribute on this tangent rather than just guttersnipe:
Maybe it would be better if physical differences between species were represented with a minimum bonus rather than a maximum bonus. If we're trying to capture that half-orc PCs are strong relative to human-elf-halfling-etc. ones, give them a minimum Strength value of 12 or a +1 to Strength, whichever is higher, and a floating +2 that can be flavored as anything from cultural background to personal proclivity. Thus, half-orcs are always meaningfully stronger than average, even 17 Int half-orc wizards. Gnomish barbarians can be just as strong as half-orc ones even from level 1, gnomish wizards... not so much. (Standard array and point buy would need to be fiddled with somewhat for it to work, of course)