How can they know it wasn't a spell? Because a spell wasn't cast. I don't think there is a single spell in the game that doesn't at the very least have a verbal component. And before we get into what exactly they are saying and if your paladin always declares "For Helm" before activating Divine Smite and how that could be construed as the verbal component for a spell, again I'll say that for the paladin example I was talking about Orcus. An immortal being who has fought many, many paladins as well as clerics. I think he might even have truesight (will need to confirm) which could allow him to see enough to know the difference between a spell and a magical ability. Also, he would know that while he normally can effect people with his spells, he historically has a harder time of it when Paladins are around, law of averages plays in here by extension of my example being that you are fighting a Demon Lord.
Was it a spell? What if it was a subtle spell cast by a sorcerer? What if it was a spell-like or supernatural ability? What if it was an NPC whose abilities don't map directly to a character class; there are infinitely more NPCs than PCs? What if it was a multiclass PC? What if you just didn't notice the components?
And 'Law of Averages'? I've worked in casinos for over quarter of a century, and that little myth is what keeps us in business. The fact that more PCs than usual made their saves (how is
that observable!) does not indicate that there
must be an aura around which gives them a bonus to their saves, therefore there
must be a paladin around! What about good or bad rolls? What about other abilities/spells which increase saves? What about magic items? There is absolutely
no way for Orcus or anyone else to reason that
because four out of five targets made their save that
therefore there
must be a paladin about!
The point is that if it is real, concrete, then beings can observe it. Even if it is only those beings whom are so massively powerful that they are end game fights, if they can make these connections and observations then classes are real and concrete.
But class abilities are not always concrete. 'Saving throw bonuses' are not observable, and frequency of saves being made/failed is not only difficult to observe (if at all) but impossible to pin down to the bonus. When dealing roulette, I know that there is an equal chance of a red number or a black number, but if I didn't know that but tried to work out the odds by simply tallying the results, you might
expect that there would be an equal number of red and black, but you'd be a fool to believe that. I've spun 17 black numbers in a row before now, and it was just a mild curiosity because we have an electronic board which shows the last 17 numbers and they were all black. For any period of time short of infinity, your mere observations are extremely unlikely to see an equal number of reds and blacks, therefore you could not reach the conclusion that there
must be an equal chance for each result.
The only reason we
know that the chances are equal is that we can see the numbers in the wheel! 18 red and 18 black. The point is that although the game rules can tell us that there is a definite +2 to this save, the creatures in the world cannot
know the game mechanics, only observe the results, and the results do not reveal the game mechanics with any certainty.
In real life, physics is what it is, but even we cannot
know what it really is for a certainty, only to have our guesses get closer and closer.
Damage is observable through more than effect. Rolling a 1 on a swing of the greataxe would have to be a different swing than rolling a 12. A fireball that does 40 damage is different than one that does 10.
Rubbish! Rolling a 1 on damage is as much about the target trying to twist out of the way as the force of the blow. Luck and unpredictable chaos is such a huge factor that the game never said that weapons do, say, 7 damage; they made it a random number to account for this. The orc isn't swinging with random strength or skill, but the result is random anyway. You cannot observe a single swing and
know the game mechanics behind it, and therefore you cannot know them even by observing multiple swings. Even if you get a set of results which seem to prove a point, what point? You can say that these swings did more damage on average (although that is unreliable; you just might have rolled well lots of times), but you cannot say that the increased damage
must be because he is raging, therefore he
must be a barbarian! There are
many, many ways to increase damage!
If the 10 hp
fireball kills 5 people but the 40 point one only scorches 5 others, how can the creatures in game
know which fireball was more deadly?
However, none of that matters. When I brought up the Fireball it was in direct reference to the Evoker Wizard's ability to create pockets of safe space within the fireball, where targets he/she chose were not effected. No matter how you flavor the creation of those pockets, they exist and have the same effect, and are also only possible to be created by Evoker Wizards.
How do you know that only evoker wizards can do this? What about NPCs? Magic items? multiclass PCs? Classes, spells yet to be written? Don't sorcerers have a metamagic ability to do this? If not yet, they might in a future supplement.
And all this doubt, even when you have the PHB in your hand! The creatures in game cannot
know any of that!
Yeah, critical hits are based on luck and a bad example. Does not invalidate the other parts of the discussion, because even though they are difficult to observe, all Champions do get crits more often than normal people, so it still has to have some meaningful impact on the game world.
No they don't! Champions have more
chance of critting, but this does not mean that they actually
get more crits! And as you pointed out, crits are not observable.
I acknowledge that some class abilities
are observable, but I'm saying that those observations cannot allow the creatures in game to
know what 'class' a person is, because they can never be certain that what they are observing is the result of a class ability, because 'class' is a game mechanic and they cannot know about game mechanics.
Does he turn into a bear by casting a spell? Or does it just happen? Can he also turn into an Elemental? Is he capable of becoming a dragon? There is a difference between casting Polymorph and using Wildshape, even if it is just the use of the bonus action instead of an action and the fact that a wizard, sorcerer and bard all approach magic differently than a druid.
First, 'bonus actions' are not observable as being 'bonus actions'! Next, although wizards/sorcerers/bards/druids in some ways (but not all ways) approach magic differently, so does each individual wizard/sorcerer/bard/druid! Every character is an individual, and an observer cannot
know that the differences he observes are because they are different 'classes' (a game mechanic) or just different styles. Sure, a guy might
say he is a Druid, and 'Druid' might very well be an in game title, but the guy may be lying? He may be pretending to be a Druid by using similar abilities and be skilled in Deception. He may legitimately have the in game title without actually having any druid class levels.
That's the other part of the argument by the way. It isn't just each individual ability. It is the fact that these abilities tie into other ablities. If you can do one thing you can most likely do this other thing or you will in the future.
Since PCs can multiclass, can acquire magic items, fluff their abilities in many different ways, and since NPCs are not limited to the class system and vastly outnumber PCs, there is no way for an in game observer to assemble the observations into 'class' that matches the info in the PHB. He can't even tell if a person he meets is a PC or an NPC.
The entirety of my point is that class effects the game world. Since it effects the game world it is real. Since it is real, as per my understanding of the term, they are concrete. It is more than fluff, it is real in some way, shape or form
And my rebuttal is that so many things affect the game world in similar ways is it not possible for an observer to observe the game mechanic of 'class' through observing what happens in his world.