If one wants to forget real-world physics and have the game world run on its own different physics as defined by the game rules, those rules are going to have to include an awful lot of pretty dry stuff about how said game-world physics actually work...think about it.
Actually, a lot less than you might think if you have always assumed a "real world process" system basis. A lot of that "physics" is actually not relevant at all to the system; it's part of the "colour", or the way the players sat around the table visualise the game. The system really only cuts in when an actual resolution of a contested action must be made - i.e. when something the system defines as "a thing that doesn't happen automatically" is attempted by either a player (and yes, I do mean "player", not "character" - even though they will likely be attempting to have their character do something) or the GM.
All the rest is "fluff" - which is not to say it's not important; fluff is very important. But it doesn't need "system" to make it work.
You can't really have it both ways. Either you use real-world physics as a basis for what the characters are experiencing in the game world (e.g. you fall because of gravity, fire doesn't burn without oxygen, solid-liquid-gas are the three states of matter, the characters are carbon-based lifeforms, etc.) and tweak them a bit to suit what magic does; or you start completely from scratch and rewrite physics to suit your game world and the universe that contains it.
You may very well "write" the fluff to resemble real world "physics" (though I might call them "tropes", rather than "physics" at this point, since the cause need not and usually is not either specified or fixed).
For the rest, I find that the pitfalls are manifold. Take "fire doesn't burn without oxygen", for example. Does this mean that a fireball doesn't work underwater? If yes, then I would throw back that "fire doesn't burn without fuel", either, but the fireball still works just fine without the caster carting around cans of gasoline. Does this mean that the fireball spell doesn't work out of water, either? Or does it mean that the "laws of physics" just don't apply to fireballs? If the second, then I would argue that you have just proven that "fire doesn't burn without oxygen" is not a "rule" that need apply in the game world at all, so, if you insist on it, it's purely arbitrary.
Conversely, if a DM intentionally ignores a rule and a better game (i.e. more fun) results because of it, isn't that a good thing?
Sure, but how will they know that up front? And "more fun" for whom?
My experience is that on-the-fly "rule-zeroing" hardly ever achieves this - certainly not in the longer term. Carefully considered houseruling might - which is just one reason why I'm a lot more accepting of houserules that are published by the GM in advance than making up rules on the run.
Problem is, for most of us pretty much any game system will have some aspects to it that some of us like and others that we do not; and they'll be different for every one of us.
Sure - good job there isn't just one single game system!
To clarify, my position is this:
- Considered houseruling of a system is OK where needed; on the fly rulings and
deliberate rules ambiguity are not because they gimp players of understanding they should have of how the (game) world works and frequently cause unintended consequences.
- Every group, for every game they play, should ideally pick a ruleset closest to (or most easily houseruled to) what they want out of the system for the game they want to play. If houserules reach a book-full, you might want to consider other start points.
- D&D, specifically, is one just possible ruleset that might be chosen. As such, it would do best to aim at those features of play that are best fitted to its "core assumptions or features"; I see these latter as character classes, hit points, armour class, experience points, levels, magic by "spells" as things rather than processes and doubtless other stuff I've overlooked right now. A real-physics-based, real-life-process based system really is not what I think fits with these things. That is not to say that a game based on these things is "wrong" or even "not great fun" - just that a game focussed on such considerations would be much better off starting with a different system than what D&D has ever been.
I think that a good GM generally won't do this in a way that won't work with long term players (because if he does, the players with which this isn't a fit will leave his group). Additionally, poor GMs who abuse this rule are more likely to be poor in other areas, from my experience (Rule 0 abuse from the GM is commonly found with GMPCs or railroad plots, from my experience).
This, again, seems to be talking about "houserules" set up beforehand rather than making up rules
in media res because the rules themselves are deliberately ambiguous.
I agree that Rule 0 is often used for enabling railroad plots and GMPC "neat tricks", but I think there's a thin, thin line between this and building situations and NPC abilities based on how the GM imagines the "real physics" of the system will work. If I build an NPC with a spell that is loosely and ambiguously defined,
of course I design them using
my interpretation of the spell as a basis. If I set up an in-game situation involving ambiguously defined elements,
of course I set the situation up assuming
my interpretation of those elements - how could I do otherwise? But, if the players don't share my understanding/interpretation of those rules, it will naturally seem to them that the NPC is advantaged, or that the situation is contrived - even if that was far from my intent.
In part that's due to how it is presented; and in principle the idea of the game rules being in the DM's control is not a bad thing. Where it goes wrong is when rules are changed on the fly (as opposed to after some forethought ahead of time) and-or the rules whether changed or not are not applied consistently.
This seems to be very close to my position. Houserules notified in advance and consistently applied are fine (with a caveat you phrase so well I'll quote it below); on the fly changes or rules that demand interpretation on the fly aren't.
If a DM sees a problem and thinks she knows how to fix it she should try.
The challenge is admitting after a while that the fix is worse than the original problem, on those occasions when such turns out to be the case.
So good it bears repeating just for truth. I agree completely; if you must change the rules:
- Consider it carefully beforehand, including figuring out
why you want the change and checking that you are fixing the bit that actually needs fixing (for you).
- Notify all involved about the change in advance and entertain objections and reservations (maybe allowing some character changes or tweaks to address those).
- Review the changes as time goes by and be ready to rethink if the change has adverse consequences that are worse than the original issue.
I'd also like to see a section devoted to resolving conflicts between the DM and the rules. In other words, a primer on how to kitbash the system.
I think it's also worth being clear about different types of addition/change. I don't for example, consider creating new "monsters" to be a system change - any more than creating NPCs or locations is. If the world has orcs with one hit die, I don't consider adding in orcs with two hit dice to be a "system change", but changing weapon damage so that it reduces CON instead of hit points would be a different kettle of fish!
For "creating new elements" - be they monsters, traps, social tasks or even spells - I think guidelines and design tools are what is needed. Something similar to 4e's monster charts (though perhaps explained differently, since some seemed to think they were "rules", as such).
For
system changes, however, I don't think there's all that much the designers can do, except explain why the systems are designed as they are designed and thus what might break if you start to fiddle with them.
I've always held that magic items found in an adventure should either be a) somewhat random, or b) things the dungeon inhabitants would likely possess and-or use.
I would be a little careful, there. If magic items are supposed to be only "found in treasure" or "hideously expensive to buy", a bunch of monsters having just what they can usefully use can look a lot like some sort of "GM's creatures are all in the conspiracy" set up. Are we considering that the monsters - but not the characters - can exchange items freely in a worldwide "magic mart" network? Or that the world-spanning conclave of black-hat mages cooperate to supply all the world's monsters with what they need?
I've seen this become a slippery slope. I once saw a player suggest interrogating a random goblin (or was it a kobold?), on the basis that "all the monsters seem to know what every other monster knows about us - let's assume they do likelwise about the boss guy's plans!"...
Problem still exists, though: how is the stuff going to be divided so everyone gets a fair share?
The players I run for handle it pretty simply (in D&D 4e); each item goes to whoever can best use it. After the run, the monetary values are totalled out (using a spreadsheet), but the values are used merely as a guideline for who gets to make choices about spending fungible wealth.
I use the original 4e item rules, mind you, and I find they have several features that support this approach. For one thing, rituals allow for items to be made, remade, modified and unmade by the characters. Basically, magic items become a character building resource that is shared by the party as a whole, rather than received by each character exclusively for their own use. This seems to lead to some interesting "team optimisation" going on, which helps enhance teamwork generally. It's certainly a lot more interesting than the "how do we split this fairly" sums I remember from 3.X, but then maybe we could handle 3.X treasure distribution based on this "teamwork" experience? Maybe, but I doubt I'll ever find out for sure.