If we got it that wrong when we don't have a host of real monsters in our world, why would the "real monsters" world have greater accuracy?
Well, partially because D&D people have in some ways better (magical) methods of traveling, learning, and disseminating knowledge than we do today.
So write a background that males you a well-travelled cosmopolitan academic - that way, you get the best skills.
To some extent, yes. However, DMs (and players) are expected to self-police themselves to some extent. Saying that your 1st level character is the guy from the Dos Equis commercials is not forbidden by any rule, but is BS.
And there is not one clear "best". If your background is "I'm an elven druid, I've lived in these woods for a hundred years and I know them like the back of my hand", your common or assumed knowledge about that area is probably much deeper and more accurate than that of some dilettante adventurer. But you may be pretty clueless about other environs; effectively raising the DC for unfamiliar settings. Conversely, a well-traveled adventurer may be a jack of all trades, master of none.
To a large extent, I find the 3rd Ed skill system was an attempt to codify use of character resources (skill points, feats) to be knowledgeable, persuasive, etc. and not leave this to who can best leverage their background and their personal abilities as a player, rather than the abilities of their character.
I don't know about that. 3.0 was very open-ended in what a Knowledge skill could be and did not have the monster ID rule; it was codified more for 3.5 (not one of the better updates, IMO). The skill system is a mixed bag; not all skills seem to mean the same thing. Something like Jump has a very clear objective meaning (roll X, Jump Y feet), but Knowledge (as well as many other skills) seems more subjective to me. AFAICT, a 20 Knowledge check means whatever the DM says it does.
As a somewhat random sidebar, the entire skill system in 13th Age works off of players saying what their background is, leaving it to the player to define what they're spending their skill points on, and the DM to decide when they apply. Those guys seem to trust people to be reasonable about it (and one of them did write 3e).
Much like hit points, we allow playability to override realism or even verisimilitude. That's hardly new.
Except that in this case, the 10 + CR rule causes more difficulties in play than if you just ignored it completely, stuck with the general guidelines for Knowledge checks, and let the DM make a quick call.
Hit points aren't worth it either.
You want knowledge beyond that, pay for the skill. Again, character resources dictate character abilities. I have no desire to see the Wizard with no ranks be more knowledgeable of religion than the Cleric, or beat out the Ranger re Nature. Not even at 1st level!
Well, that would not happen (assuming the other characters maxed their main knowledge skill) unless the wizard had an Int mod 4 higher than the other character. If that's the case, I think Mr. Wizard the Genius deserves it. And so what if he does know more? Training quickly outpaces these ability score differences. If anything, I think the genius deserves more of a relative advantage. As another random sidebar, one of the minor class abilities I've added is to give several classes extra bonuses in their main knowledge skills to make sure that druids are the nature experts (nature sense expanded).
But why does this trained only line of reasoning apply only to Knowledge? A character who rolls a 20 on a Jump check jumps as far as the check result dictates. A character who rolls a 20 on Diplo gets the results of a trained diplomat. Why does a character rolling a 20 on Knowledge get no more than a character rolling a 10? It seems perfectly reasonable to me that even a person of average intelligence can answer a DC 20 question 5% of the time. People hear things. The whole Knowledge being trained only thing is another easy ignore in my book.
Precisely. Just as your example means that, when people get lucky, SoD becomes less challenging.
When people get lucky, any part of the game gets easier. When they get unlucky, the game tends not to go well. That's why we roll dice for everything, because we want random, unpredictable outcomes. No problem here.
Not sure how the raising of an NPC of negligible consequence fits with resurrection being restricted (doesn't Raise Dead fail on death effects?).
The NPCs importance to the PCs and his importance in general are different things. To explain, this NPC was a rich guy the characters met only due to a teleportation accident, resurrection in my world requires that a life be sacrificed in exchange for any resurrection, and his traveling party brought along a condemned criminal and resurrection scroll for that purpose, because they knew it was a dangerous trip.
If I had a death gaze and were inclined to engage in combat, I'd be targeting the fellows most effectively cutting down my lesser undead minions to prevent them getting to me - and getting out of Dodge if it looked like they were getting through. But then, Bodaks have INT 6 - so they should be no more tactically savvy than, say a 7 INT dump fighter.
Also, how are they going to get out of dodge with that speed? This bodak had 10 Int, and was played as such. He did the best he could. I believe the lesser undead were mowed down by fireballs from the party wizard, who most likely was smart enough to stay far away. Not much a slow-moving bodak can do.
And I suppose if you have L6 characters powered up to the point where DC 20 saves and hitting AC 31 is simply routine, and the resources to raise (resurrect?) an NPC are easy enough to come by that we'll use them on an NPC we have no real alliance with, then this creature isn't that big a deal. Let's add in a monster with SoD at a DC that is actually as challenging as a 15 would be for a typical L8 character and let it fight for a few rounds.
I suspect they killed it without hitting its AC. Magic missiles and such. And, the point of all this detail is that my PCs were a good bit more powerful than your standard array, standard gold, standard rules chumps (as I would hope most are), but so was this monster. It all balances out.
I also find it odd your initial arguments was that it's all about learning about the enemy, but now the enemy isn't really a threat anyway.
My initial arguments were about SoD not being an inordinate threat, in some part because knowledgeable PCs can mitigate that threat. This is your tangent.