Judging creature strength

[Cassiel]

I think I'll keep sense motive a large element in all calculations.

On the opposed rolls, I was advocating allowing NPCs/monsters comparable bonuses to PCs. Another thought strikes me:

If you decide to give the +2 synergy bonus to folk with 5 ranks in whichever of Sense Motive / Knowledge (X) they're not using for the roll, could we give the "defending" critter a +2 synergy bonus for 5 ranks in disguise?

The next discussion, of course, is about how to factor the monsters' assessment of their foes into the AI. It'd be one *really*, *really* easy way to encourage people to take ranks in bluff / disguise...

[Fionn]

Sorry Dorn - I probably read more into that than you meant.

[fluffmonster]

let's start from general principles here.

1. First, what needs to be measured here is the *absolute* difficulty of the mob, not relative to the party. Level differences thus are not a good approach to employ primarily because there are too many distortions introduced which complicate the matter far more than necessary. Also, having a variable scale of difficulty measure greatly complicates even talking about how to handle the whole thing at a game-rule level. Lastly, a relative assessment runs the risk of misleading PCs and doesn't even get at how you handle a variety of PC levels in the party. Leave it to PCs to determine where they stand relative to the absolute assessment returned. An absolute measure will make it easier to come up with a system of DCs on the check - just a simple opposed bluff roll isn't necessarily gonna be enough.

A more consistent and simpler system will focus on absolute difficulty.

2. what we're talking about is a kind of intuition. You see the mob, and something allows you to make an assessment (right or wrong) about its uberness given the meager information available from spotting it. Thus, the relevant skills are based on wisdom and/or experience.

3. The wisdom part is best represented as a wisdom-based skill, and sense motive works beautifully for that. Its got a PnP basis, the skill is the best approximation for general intuition of the lot. Further, the contribution of sense motive is very general and should apply to any and all mobs. Classes that have sense motive as cross-class are just gonna have to suck it up and live on their wisdom bonus (or lack thereof).

4. How to represent experience? several dimensions to this have been proposed, based on level or class similarity.
a)) direct experience is going to be creature-type specific realistically. You may know goblinoids inside and out, but that does nothing for assessing dragons. This maintains a measure of simplicity as there is a relatively small number of creature types and all creatures are already grouped this way. You maybe get a +1 for each encounter with that creature type or something. We're being told anything can be coded, but a big question here is whether this data will bloat bics too much.
b)) general experience is probably best captured in level of the PC, but not necessarily like +1 per level. Really depends on how the DCs are scaled I guess, but maybe +1 every 2 or 3 levels. General experience is useful, but slowly gained.
c))class-specific bonuses...conceptually, this would only work for the class levels a monster has. For orcs or goblins, their difficulty is almost always wrapped up in what class levels they are padded with...this is where we get CR8 goblins. What you folks aren't including there are the raw monster levels...a base troll is like CR5 or something, and that's without any PC levels. How do these monster levels figure in? No class affinity should be working on those, you'd have to rely on direct experience. How exactly this part is put together depends on how the mix of monster and PC levels is to be handled.

I think you put those pieces together, you've got a workable and understandable system.

[Mayhem]

let's start from general principles here.

1. First, what needs to be measured here is the *absolute* difficulty of the mob, not relative to the party.

I disagree. I think a relative assesment is better - "is this thing better than me?"

And its not for the whole party - its for the one guy, looking.

It would likely be wise for the entire party to look, and compare notes, of course.

Fig: "It doesn't look so tough to me, I say we charge"
Wiz: "Wait, my friend - it may not be much of a warrior, but I think it has considerable magical power - charging straigh tin would be... unwise."
Rog: "Well, it doesn't look to be paying much attention - I reckon I can sneak up on it, keep it busy long enough for you to get in close without getting fireballed..."

[fluffmonster]

relative assessment however makes it more difficult to share information without going meta. "Hey, its a little weaker than me, but I've got 3 levels on you so it might be tougher for you". Plus relative is less realistic, and presumes a meta basis in the first place...its this number, you're this number, here is the difference. Part of the RP is assessing the difference yourself, I don't see why that should be undone. If relative is what you want, then just unblock the challenge note in the creature description and be done with it. Otherwise, piling on information about specific relative difficulty just sounds to me like an excuse to meta a lot rather than a little.

[Cassiel]

2. what we're talking about is a kind of intuition. You see the mob, and something allows you to make an assessment (right or wrong) about its uberness given the meager information available from spotting it. Thus, the relevant skills are based on wisdom and/or experience.

If we have to pick one, I'd go along with that - happy to abandon "Knowledge (X)" as an alternative roll. I think the synergy bonus is reasonable, though.

a)) direct experience is going to be creature-type specific realistically. You may know goblinoids inside and out, but that does nothing for assessing dragons.

Favoured enemies! What sort of bonus do we think that'd be worth? Can we tie it in to how long the ranger has had 'em?

This maintains a measure of simplicity as there is a relatively small number of creature types and all creatures are already grouped this way. You maybe get a +1 for each encounter with that creature type or something. We're being told anything can be coded, but a big question here is whether this data will bloat bics too much.

If that could be done it would be lovely - although if we have knowledge skills, an argument can be made that (a) PCs can take skill ranks to represent their experience of critters and (b) not to do so encourages relentless farming to get knowledge bonuses for which a game mechanic already exists. Just painting a worst-case scenario...

b)) general experience is probably best captured in level of the PC, but [snip]General experience is useful, but slowly gained.

Agreed. We could actually tie this to monster types - say:
Fighters: get bonuses to monstrous humanoids, giants
Paladins: outsiders, monstrous humanoids
Druids: fey, animals, plants
Wizards: dragons, magical beasts

etc. Some argument necessary over who would get what if this is thought to be a good idea.

[Fionn]

If we go with an absolute scale, I'd recommend straight CR modified by the SM roll. We may wish to either make the roll more accurate if you share classes, or come up with some way of modifying based upon cross-class threats (Tank > Rog, Rog > Wiz, etc), but that is extremely subjective and likely to be more of a hassle.

For simplicity and consistancy, I'd say (if we do anything), to strongly look at a SM based system with bonuses for in-class threats (with tanks being in-class for essentially anything without a dominance of non-tank leves).

This avoids some silly IG ranking system to replace CR. I'd hate to see 'Novice' be the IG speak for CR1, 'Tenderfoot' for CR2, etc. Return the CR (which can easily be modified if you blow the roll) and let the PCs figure out how to translate to Common.

The only weakness I see with that would be preventing repeated SM checks until you get a consistant value. Honestly, I'm not sure that's an exploit - you spend IG time/risk, you have better data.

This would require that builders give some sort of data about the abilities of their mobs. NWN1 models allowed only weapon/shield to differentiate a tank from an archer from a mage (thus reliance on floaty text). Hopefully we get better models (I know scaling will help) so that PCs can recognize a 5th level spoonbender vs a 5th level warrior.

[fluffmonster]

As to the returned value...one way to think about it is that a single value is returned, and it may be right or mostly right or mostly wrong. A single CR works there. That however presumes a specific kind of error. Another possibility is a non-specific error, or only being able to narrow it down to a range. Those "silly IG classifications" are precisely that, reference to a range of CRs. The specific error is easy to do, but the non-specific error is more interesting. This is entirely aside from other returned information like whether there's any spellcasters in the mob.

So there we have 3 possible elements to the returned information. If the returned information is the numerical value [CR + X +/- Y] plus a qualitative description, then

X is the specific error (and could be negative)
Y is the non-specific error which defines the range of precision of the difficulty estimate
the qualitative description might include things like spellcasting, special abilities, whatever.

We must have generation of an X, that much is certain. Y would be completely transparent without X. We do not necessarily need a Y, but I think it would make the assessment far more interesting. The qualitative information is necessary if this effort is to be anything more than a glorification of the CR information already provided in the creature description (which we currently hide). I don't see how just 1 roll could do all that though.

//edit...and actually, what the return values are needs to be determined before the roll bonuses are. If there are to be different rolls, it becomes possible to have different bonuses applied to each.

[ç i p h é r]

What happens on the first instance where this relative or absolute difficulty rating is grossly inaccurate? Will the player ever use it again if they survive the encounter? Will they ever use it if they get killed because of it?

A difficulty rating system just won't work well enough IMO to warrant spending the time writing one. Bioware's rudimentary system is enough evidence to sway my opinion in this regard, and for those who've only played on ALFA where it's masked, I encourage you take an objective look at it elsewhere (fire up the SP module). There are too many variables to consider and not enough statistical information available to draw a conclusion about what matchup favors who.

There are a lot of good ideas here to consider but regardless of the specifics, I still believe we should limit the system to providing factual information for players themselves to evaluate or provide nothing at all (ie "you don't see anything of significance"). Providing FALSE information will render the system just as unreliable as the relative measure of difficulty making it quite unlikely that anyone would invest in, use, or depend on this in any way.

Just my opinion of course, so carry on.

[fluffmonster]

kind of depends on the extremity of the error. extreme errors would result in extreme outcomes, but its not clear that's really necessary.

So what you're talking about though Cipher is a check as to whether it returns a true value or nothing at all? Could be doable certainly. I'd still like the return to be a range rather than a specific number. Only need one roll anyway. How much you beat the roll could determine the amount of qualitative information delivered.

[Ronan]

Alright, enough with the feature creep

• Its not going to be terribly complicated.
• It won't record any extra data on bics, or anywhere else for that matter.
• It won't be and shouldn't be totally accurate.
• It could and should be wildly inaccurate at times.
• BAB will be used to judge a PC's expertise at melee fighting.
• Sense motive will be used heavily. Even experienced pros underestimate opponents, especially ones with more ego than wisdom.
• Knowledge skills will be used at their appropriate creature types, as per 3.5 rules.
• All of this is relateively simple.
• I do need to figure out the best way to turn an object id number into a seed for a d20 roll.
• It will prevent Bob the Adventuar from suiciding himself on that CR 12 goblin chief who just looks a whole lot stronger than he is.

[ç i p h é r]

So what you're talking about though Cipher is a check as to whether it returns a true value or nothing at all? Could be doable certainly. I'd still like the return to be a range rather than a specific number. Only need one roll anyway. How much you beat the roll could determine the amount of qualitative information delivered.

Yes, that's what I was saying. From my own personal perspective, any system which introduces intentional inaccuracy or one in which the accuracy is suspect is destined to remain unused.

Ronan, I don't think anyone would dispute that players can (and will) underestimate their opponents from time to time or that it should not happen. What I'm discouraging is for us to write a system which does that for the player. I don't think that's a well devised system as I think players would rather make a guess themselves than use a computer's guess based on unknown variables. Especially knowing that it can also be "wildly inaccurate". Especially in a perma-death environment.

[Fionn]

ç i p h é r - for simplicity I agree with you. OTOH, I want to see a serious reason to play (or hire) a Scout. We can certainly weight it so that you are more accurate with things close to your value (e.g. a Ftr15 might only know the Kobolds are less than CR5, while the Ftr3 might know that they are CR2). This would make it more accurate where it counts.

If a PC gets burned by the system because they are *not* a skilled scout, then they should learn to rely on other data sources - like a skilled scout :)

Ronan - I'm not sure if I *want* a seed value for the roll. If we do so, I'd avoid having it be 'wildly innacurate', or even often somewhat innacurate. Say miss the (modified) SM by 5, you get +/-1, by 10 you get +/-2, and by 15+ you get +/-4 (the max possible innacuracy).

[Dorn]

Sorry Dorn - I probably read more into that than you meant.

np mate. Easily done.

[Ronan]

Ronan - I'm not sure if I *want* a seed value for the roll. If we do so, I'd avoid having it be 'wildly innacurate', or even often somewhat innacurate. Say miss the (modified) SM by 5, you get +/-1, by 10 you get +/-2, and by 15+ you get +/-4 (the max possible innacuracy).

A seed value will ensure the same roll outcome on a creature every time without having to store any information anywhere. When I said "wildly inaccurate", I ment a PC with no SM ranks, 8 wisdom, judging something he is totally unfamiliar with.

Players can ignore it if they'd like, but it adds value to the grizzled rogue scout who can size up an enemy for his party, or the learned mage who can look at a dragon and tell its age.