Do you think the entire MM is full of creatures that only live in secret on one island shrouded in mists?
That would imply no one would have much in the way of knowledge of those creatures. Do you consider the MM bedtime reading for toddlers around the D&D world?
Certainly. If you're in a desert and you get attacked by a White Dragon, there's going to be some head scratching. Your desert character might not know too much about what's going on.
I was referencing "some creatures would be much less well known than their HD might indicate", but the need to have separate modifiers for every character based on background and experience in different terrains multiplies the complexity considerably. But your model suggests that both the Arctic character and the lifelong desert dweller know all the basics of that white dragon, Where it is encountered does not alter those odds, in my view (unless we assume the arctic dweller dismisses the signs it is a White Dragon because they don't live in the desert - but then, that implies he has knowledge of the desert as well as the arctic).
Perhaps because even common knowledge could be missed by a person with no bonus 45% of the time. Or for scenarios where you actually do need to know minute details about Medusas (as opposed to the obvious "don't look at them", which I'd put at DC 10-15 or so depending on context). Perhaps for things other than monster IDs.
Take 10! Perhaps this is an answer for a lot of the randomness. Given the knowledge check cannot be re-tried because "the check represents what you know, and thinking about a topic a second time doesn’t let you know something that you never learned in the first place" , perhaps Take 10 should be automatic in many cases. This does, however, set a pretty strict curriculum - EVERYONE with +3 knows everything with a check of 13- and nothing with a higher DC.
I wonder whether a better answer might be that the check is "what you immediately recognize in the heat of the moment", and you may Take 20 by taking the time to sit down, in calm surroundings, gather your thoughts, and carefully consider all aspects of the issue. Maybe that gets a penalty if you lack the appropriate tools - say, without your notes and reference materials, you take a -5, say, much larger than the usual penalty, but with access to excellent facilities, such as a research library, you get +5. Or perhaps taking a few moments to review the notes you carry with you, and consider the matter, allows you to Take 10 (meaning all common knowledge can be recovered in a minute or so, provided you can focus on the issue without being distracted), while access to proper research materials (your library at home, for example) and a few hours permits you to Take 20. A great research facility might even allow "Take 20 with a bonus", while colleagues might provide a sounding board ("Aid Another").
We'd still be a long way off from "real world" realism. Maybe not as far as I'm off the actual topic, but pretty far nonetheless.
I wouldn't either. However, a bodak has six entries that could qualify as special abilities (and revealed as useful knowledge). I would not require a DC 43 (i.e. epic) check to learn all of them.
It is another flaw in the system that more abilities = less likelihod of recall. Perhaps a random number of facts, with a bonus for making the roll by X (and X could vary with the creature, smaller for those with lots to reveal) ensuring higher skill typically generates more results. That said, for a rare creature with many unknowns, I have no problem with only Epic Success getting all the details, especially with NO CHANCE of misrecall or misinformation resulting in an error, rather than "don't know".
I find this discussion particularly ironic because I did set a group of PCs against a bodak once, and they had no religious types and rolled low and didn't have any idea what it was. They never really figured it out and didn't particularly recall the bodak from the MM even when I explained it later, and the idea that all those angry glares I described were gaze attacks and that the fort saves were against death seemingly never occurred to them. At least, until the ranger's animal companion died after several rounds of this. Even then, they smoked it and its numerous allied undead without too many worries; I think some NPC died but that wasn't a huge deal.
Which just takes us back to the idea that this "broken math" and "guaranteed PC death" nonsense is...nonsense.
Does the fact that I once had a group roll all 1's for saves prove it's not nonsense, or do we accept the potential for statistical outliers? If a bunch of NPC's died, as well as that Animal Companion, perhaps the issue is that you targeted the creature away from PC's, another means of mitigating the "save or die" aspects. That, or they're heroically using the NPC's as human shields to die in their place. The Bodak DC is 15. At CR 8, a strong save is +6 and a weak one is +2. It's CON based, so slap +1 to +4 on and we get a range of success from 12 (45% success) to 5 (80% success).
Now, assuming the bodak took your advice and did plenty of scouting, it should target those FORT weak characters. Three attempts have a 9.1% of not killing the target. Three attempts against the hardiest targets allow a 51.2% chance the target survives.
i think those are very good points, and I did have to think about them, which suggests to me one can go in either direction on that particular point and I prbably would be entering the realm of pixel picking if I were to dig my heels in and defend the position that "no, poison must be different from these other things and the system supports that!" So let me start by saying my chief reason for wanting SoD is aving high stakes lethality at certain points in the game.
-i thing with the sword wound, the difference is i can easily write off the range of damage as a scratch or an impalement depending on the result. It has shortcomings of course (you cannot stab a 20th level character in the heart unless they only have a few hp left). But that is harder to do with a cobra, where you are basically either poisoned or not and cobra venom, once in the system, goes for the organs (my understanding is it can make your heart stop or interfere with cardiac functioning). But also keep in mind many snake poisons in egame work like breath weapons. Some are save or take 2d8 damage, or some are damage plus a special condition. 3E uses ability damage that is often enough to kill. I just like having soe venomes that are death.
So why does the most lethal of venoms either kill the target or leave him unharmed, with no possible result in between? That is what SoD does. Is that a reasonable result that satisfies your desire for verissimillitude?
-dragons breath is a save but you are right falls more inline with traditional damage. This is probably best acknowledged as being somewhat inconsistebt with my claim about poison. But i can still see a difference in play, where I could take in various amounts of the gas and suffer a broad range of damage. Also fire and acid do have different properties in most editions so there is at least some effort to ake a distinction between that and a sword blow. I think the chlorine gas though is hard not to ignore. Probably a good argument formakingit save or die (or at very least con damage and maybe even dex and int).
I'm amazed you can see all those possibilities, but view poison as always having a precise dose delivered (or, perhaps, nothing), with no possibility that a hardier soul might be injured, but not killed, by the lethal venom. If he does save, that exceptionally lethal attack did precisely nothing? "Dead" or "Unaffected" are appropriately the sole choices?
