Reply to thread

Message

<blockquote data-quote="pemerton" data-source="post: 6628611" data-attributes="member: 42582">I am assuming that the adventuring party is not a random sample of inhabitants of the gameworld. I am likewise assuming that the process of PC construction (via array or point buy) is not a model or emulation of any ingame causal process. Hence, the PC stat spreads tell us nothing about trait distribution in the gameworld per se, and nothing about what traits may or may not be correlated. All it tells us is what the distribution of traits among these four to six individuals is, and there is nothing anti-verisimilitudinous (that I can see) in the fact that all of them have either specialised depth, or a degree of breadth, but not both.If you build in additional premises - eg that point buy is a model of an ingame causal process, or that the PCs are a random and representative sample of the gameworld inhabitants - then I can see the potential damage to verisimilitude. But then letting the players roll 4d6 (if the campaign world norm is 3d6) or letting the players assign their stats (which corresponds to no ingame process) would equally damage verisimilitude, I would think, on those strong simulationist premises.EDIT:They existed in the 80s - I can't remember off the top of my head, but I've encountered heavy process-sim systems where STR and CON have to correlate (or where a low STR gives a penalty to CON).You get other approximations to these, too, like in RQ where hit points are a function of size and CON. Or 1st ed AD&amp;D, where a fighter has a CON minimum as well as a STR minimum.</blockquote>

[QUOTE="pemerton, post: 6628611, member: 42582"] I am assuming that the adventuring party is not a random sample of inhabitants of the gameworld. I am likewise assuming that the process of PC construction (via array or point buy) is not a model or emulation of any ingame causal process. Hence, the PC stat spreads tell us nothing about trait distribution in the gameworld per se, and nothing about what traits may or may not be correlated. All it tells us is what the distribution of traits among these four to six individuals is, and there is nothing anti-verisimilitudinous (that I can see) in the fact that all of them have either specialised depth, or a degree of breadth, but not both. If you build in additional premises - eg that point buy [I]is[/I] a model of an ingame causal process, or that the PCs [I]are[/I] a random and representative sample of the gameworld inhabitants - then I can see the potential damage to verisimilitude. But then letting the players roll 4d6 (if the campaign world norm is 3d6) or letting the players assign their stats (which corresponds to no ingame process) would equally damage verisimilitude, I would think, on those strong simulationist premises. EDIT: They existed in the 80s - I can't remember off the top of my head, but I've encountered heavy process-sim systems where STR and CON have to correlate (or where a low STR gives a penalty to CON). You get other approximations to these, too, like in RQ where hit points are a function of size and CON. Or 1st ed AD&D, where a fighter has a CON minimum as well as a STR minimum. [/QUOTE]

Verification