KDS v0.1 - more accurate CR and comparisons of any ability in 3.5

[nijineko]

D&D is complex. And trying to figure out how to pick the right mix for an encounter is even more complex.

To make a character, NPC, or monster using the default methods, you have to use the following:



Stat Block items
Method used to pick


Primary statistics (STR, etc)
Random Number Generator


Secondary stats (saves, AC, etc)
Various derived Formulas


Race
Package Select


Classes (base and prestige)
Package Select


Templates
Package Select or DM approval


Skills
Point Buy


Feats
Option Select


Carrying Capacity
Secondary stat, look up on chart


Languages
Option select & point buy


Money
Random Number Generator


Equipment (weapons, armor, etc.)
Purchase from giant lists


Special Abilities
Derived from race, class, or template


Special Qualities
Derived from race, class, or template


SFX (psionics, magic, etc)
Derived from race, class, equipment, or template



That is a huge range of variance in possible options that any given character, NPC, and monster will have. The demerits of CR have been discussed at length elsewhere, so I will not go into that here. As can be seen from the table above, there are so many different aspects to a creature, that it is hard to quantify the value or weight of various different options, especially when some choices are very niche in scope and effect, while others are fairly broad.


I have come up with a variant CR system that potentially applies not only to monsters, NPCs, and characters, but also to items, powers, spells, special abilities and qualities, and other special effects. This variant potentially allows one to roughly compare almost any aspect of the game with any other aspect of the game. It is not perfect, but I believe it is better than the CR system.

The first step in comparing two things is to convert them all to the same unit. Looking at the table above, this at first seemed impossible... until I noticed that there was one unit in the table above that actually included all the effects of all the other units.

Feats. There is a feat that grants the benefit of almost every other part of the stat block. And where there is not a single feat that gives a specific benefit, a group of feats can equal the same end effect.

I call this the "Hidden Equivalency Value", or HEV, a value which is equivalent to every other value on the stat block, that has been hidden in plain sight all along.



Stat
Feat
Equivalent to one HEV


BAB
epic prowess
+1 to attack


Fort
great fortitude or luck of heroes
+2 to fort, or +1 to all saves (thus after first level every two or three levels of saves is worth one HEV)


Ref
lightning reflexes or "
+2 to ref, or "


Will
iron will or "
+2 to will or "


Skill points
open minded or various in PHB
5 skill points or +4 bonus to skills


Armor usage
various armor proficiencies
gain proficiency


Shield usage
various shield proficiencies
"


Weapon usage
various weapon proficiencies (simple, martial, exotic, etc)
"


Unarmed strikes
improved unarmed strike
"


AC
dodge
+1 to AC, or dodge bonus specifically


Natural armor
improved natural armor
+1 to natural armor


Initiative
fire heritage
+1 to initiative


HP
toughness
+3 hit points


iterative attack
rapid fire or snap kick
extra attack


spellcasting ability
magical training
grants spellcasting and 3 cantrips


spell known
extra spell
extra spell known


spell slot
extra slot
extra spell slot


gaining higher level spells
improved spell capacity
grant +1 max spellcasting level


spontaneous casting ability
spontaneous healer/ wounder/ summoner
gain ability to swap out a spell


psionic ability
hidden talent
gain psionic ability, pp, and one power


power points
psionic talent
+2 power points


power known
expanded knowledge
gain one extra power


gaining higher level powers
improved manifesting
grant +1 max manifesting level


character stat
great (stat)
+1 to stat


magic item slots
additional magic item space
grants additional slot


fly
born flyer
gain flying ability


trample
centaur trample
gain trample ability


pounce
catfolk pounce
gain pounce ability


speed
speed of thought or dash
+10 movement or +5 movement


spell-like ability
innate spell
conver spell to spell like


supernatural ability
assume supernatural ability
access to supernatural ability


1d6 of damage
improved sneak attack
use for either sneak attack specifically, or for general 1d6 damage


etc
etc
etc



Furthermore, using the note in the A&EG that states that feats are worth 10,000 GP (in item form) plus 5k-10k more GP per prerequisite of the feat in question, we can assign a GP value to feats/HEV, and could also convert the GP value of items directly into HEV.

This could optionally be used to roughly address the "value weight" question of feats, as not all feats are equal in scope and effect. Instead of using a flat 1:1 regardless of what kind of feat it is as I did above (notice that there are several epic feats listed above), one could consider using the GP value equivalency to calculate the HEV of a high level feat versus a low level feat.


As one can see from the table, it is possible to convert practically every ability listed in the stat block to HEV, which when all the various abilities of a given character or monster or NPC or other item/effect/etc are totaled, it gives a rough total power level of the target in question. And where there is not a direct HEV for a certain effect or ability, one can break down the end result of the ability into the component statistics and convert that into HEV.

Finally, once the HEV is determined, it can be converted to KDS, which stands for Kobold Death Squad: where the HEV value of a standard unmodified kobold from MM1 is equal to 1 KDS. Thus we can now determine just how many kobolds your character is worth. Which is where the name I decided on comes from.

I would be interested in feedback and analysis.



Note: Kobolds themselves use the term "Meeps" instead of KDS.

[Endarire]

Why do you value +1 init as a feat when +4 init from Imp Init is more powerful and popular?

[Biggus]

I've been using feats as a rough way to compare things for a while now. The problem is, some feats are MUCH stronger than others; it can never give you more than a ballpark estimate. For example, Toughness gives you +3HPs, Improved Toughness gives you +1HP per level. Except in games which will never get past very low levels, Improved Toughness is much better (although exactly how much better is impossible to quantify).

One way to get round this to some extent is to specifically choose weaker or lower-value feats as your base unit, but there are still problems there. You mention using the number of prerequisites as way to quantify them, but Improved Initiative is generally considered a far superior feat to Toughness for example, despite them both having no prerequisites.

The nearest I've come to making it work is to choose feats which are generally considered relatively weak as the baseline, such as Toughness, Skill Focus, and Weapon Focus. But then there aren't low-end feats which can be equated to everything, so there's still some eyeballing involved.

Epic feats are typically worth 2-3 times as much as a nonepic feat in cases where they have a straightforward numerical value. While there are some which give the same bonus as their nonepic versions, at the other end of the scale some give many times as much (Epic Speed and Epic Toughness for example). For simplicity's sake I usually treat an epic feat being worth 2 nonepic feats.

FWIW, I'd say as a very rough estimate that gaining a level in a PC class (and therefore one point of CR) is typically worth about 6 (nonepic) feats.

[Endarire]

Epic feats are also either very powerful (Epic Spellcasting, Epic Leadership, Improved Spell Capacity, Improved Metamagic, Spell Stowaway...) or mockingly weak (Epic Speed, Epic Weapon/Skill Focus...) This is more pronounced in epic feats since they're meant to be EPIC! It also felt awkward that some epic feats are weaker than nonepic feats, like Legendary Leaper being weaker than Leap of the Heavens. (Epic abilities were generally made when WotC didn't know how powerful things should be. They got better in the next few years.)

Regardless, nijineko, I understand that we need a baseline, whatever that is. Thankee!

[Kayblis]

First of all, I'd like to commend you on that work. You must have spent a lot of time double-checking to get a consistent list. This is a direction of number-crunching that will end up with a general score for a character, so using your terms, you can have a character be worth 120 HEV or 12 KDS or another equivalent.

I do believe it's a misguided method though.

The thing about D&D is that you don't have activity %s or multi-actions like other RPGs, say Rolemaster or Palladium. Barred edge cases, you get one fixed set of actions. This is obvious, but the implication often isn't. In fact, this whole board is based around the implications of this design and even then it's hard to give concrete answers sometimes. Thing is, in general, you only do what you're good at. You can crunch stats, abilities, spells and feats into a single unit, call it a cool name and turn whole characters into a single number - but in the process you lose the things that make the characters perform differently, so the metric is void. Compare two characters with the same values, for example: two Human Fighters of level 6, same total stats. One of them has 12s all across the board for stats, two separate pairs of weapon focus/specialization, and Toughness. The other is all Str/Con, dumps Cha and Dex, and has charging feats. They can have the same HEV total, the same KDS value, and their performance is like night and day. This isn't a case of a generalist not measuring up in an edge case, it's a case of one character being better than the other at everything that matters. One uses his actions efficiently, the other just does a similar action worse. This is because some options are clearly better than others, and the same feature can resonate with others to a much bigger result. Not only Power Attack is much better than Toughness, but it also mixes well with higher Str scores. That's the whole reason for there being an optimization community in the first place.

3.5 is complex. This is a good thing. Reducing a character to a single number necessarily loses important parts of that complexity. The CR system is flawed not because it fails to account for numbers, but because it lacks fine-tuning. It's not tested well against players that know how to play, and it can't account for future releases that raise the expected power level.

I don't want to shoot down the idea, and I have a different suggestion - create a system to rank performance. Just to give some ideas on execution, you could divide a monster into three main categories - Offense, Defense, Special. Judge how the creature performs with its main tactic, and assign a 'performance value' for each category. A melee brute should be judged on expected performance in melee combat, both ofensively and defensively, and any unique features that further its role or aid without hurting its performance should be thrown in Special. Note that it shouldn't include features you'll never actually use, like Standard action effects that perform severely below its common tactic for no reason. At the end, you can note these three values(like "6/6/3 for Off/Def/Spc") and attribute a value to it with a relevant tag(like "CR: 6 Bruiser"). You can have flexible rankings for creatures with multiple options, like many outsiders and dragons do, so their performance is judged based on two tactics(like "12/10/10/8 for Melee/Magic/Def/Spc", which leads to "CR: 13 Hybrid Caster"). This still requires you to judge performance based on a parameter to be set, so for a community project you'd have to estabilish a measuring stick. For me, it works fine for the groups I deal with, because I can gauge the overall performance of my groups. This is almost the exact opposite of what you're doing - you're looking only at the ingredients, I'm looking only at the finished product's results. Thing is, if you're trying to get something usable, you need to use the results as base.

[nijineko]

I don't want to shoot down the idea, and I have a different suggestion - create a system to rank performance. Just to give some ideas on execution, you could divide a monster into three main categories - Offense, Defense, Special. Judge how the creature performs with its main tactic, and assign a 'performance value' for each category. A melee brute should be judged on expected performance in melee combat, both ofensively and defensively, and any unique features that further its role or aid without hurting its performance should be thrown in Special. Note that it shouldn't include features you'll never actually use, like Standard action effects that perform severely below its common tactic for no reason. At the end, you can note these three values(like "6/6/3 for Off/Def/Spc") and attribute a value to it with a relevant tag(like "CR: 6 Bruiser"). You can have flexible rankings for creatures with multiple options, like many outsiders and dragons do, so their performance is judged based on two tactics(like "12/10/10/8 for Melee/Magic/Def/Spc", which leads to "CR: 13 Hybrid Caster"). This still requires you to judge performance based on a parameter to be set, so for a community project you'd have to estabilish a measuring stick. For me, it works fine for the groups I deal with, because I can gauge the overall performance of my groups. This is almost the exact opposite of what you're doing - you're looking only at the ingredients, I'm looking only at the finished product's results. Thing is, if you're trying to get something usable, you need to use the results as base.

I pretty much agree across the board. That's partly why this is v0.1, and partly why I mentioned some alternate suggestions in the spoilers, included the problem of the "weight" of feats where given one feat slot, there are great options, middlin' options, and crappy options that could be put in there - hence using the GP value as a partial work-around - though it would still need a lot of eyeballing.

Also, I like your ideas of performance based and separation of feats into categories, however, that has a potential problem as well.

While we could certainly rank feats by usage-category and by power level, this fails to take into account actual application. I can hand an uber-character to a sub-par player, and that character is going to still perform sub-par in most cases, because of the player skills, or lack thereof. So I think we should eliminate player input from the equation for the rough comparison.

However, there are still feats that are overall great and feats that are overall horrible, not to mention feats that are great for different things, but horrible at combat, or the reverse. As pointed out, some epic feats are worse than regular feats, while others are crazy useful. Could you display some examples of suggestions for improvement?

***

It was also asked why did I pick the smallest value (initiative)? the idea is to get the most fine tuned break down, thus smaller values are better than larger for that purpose. It may not be the best comparison, that's why I'm asking for feedback. since there is a feat worth +1 init, and another feat worth +4... obviously the player would likely pick the more advantageous feat, but I'm trying to look at the value equivalent... for the purposes of analyzing the overall POTENTIAL of a given monster/character is a +4 to init worth one feat or four feats overall? What's your take on it?

***

So that's what I'm really trying to get at, a value that represents not the actual effectiveness of a thing, but rather the POTENTIAL of the thing. Depending on what feats/options/HEV is actually selected, that potential may not be lived up to, after all, D&D is complex and covers a lot of possible play styles.

Thus I think that this idea is not misguided or off target, but rather is targeting something which has a caveats that need to be made clear in order for the tool to be used correctly.

In an RP heavy campaign, combat effectiveness might be worthless, making a minmaxed-for-combat build a very sub-par character. In a classic dungeon crawl, a diplomacy and uber-backstory-with-lots-of-political-contacts type character might be twiddling their thumbs and feeling worthless and uninterested in the monster of the day and treasure of the week. So the HEV does not represent combat value, or diplomatic value, but simply POTENTIAL value of the character based on how many feats that the thing represents.

That's actually why I came up with the 'convert to KDS' step... after all if you have the HEV, why bother with KDS? One of the reasons, is that most DMs have a reasonable grasp of how (in)effective a kobold is or is not in their campaign, thus comparing them to X number of kobolds will at least give a fuzzy impression to almost any DM. After all, again referring to the person behind the curtain of the character/monster, there are Tucker's Kobolds, and then there is Meep, and/or random Kobold #13.... same stats, very different end results.

Perhaps it should be called HEP for Hidden Equivalency Potential instead of HEV?

And thank you all for the feedback so far. Please continue to chime in!

[Kayblis]

While we could certainly rank feats by usage-category and by power level, this fails to take into account actual application. I can hand an uber-character to a sub-par player, and that character is going to still perform sub-par in most cases, because of the player skills, or lack thereof. So I think we should eliminate player input from the equation for the rough comparison.


Agreed, and to specify, the main suggestion I give is to look at the whole before ranking specifics. Marking some feats as 'always good' and 'always bad' speeds up the process, and it's a secondary effect.



So that's what I'm really trying to get at, a value that represents not the actual effectiveness of a thing, but rather the POTENTIAL of the thing. Depending on what feats/options/HEV is actually selected, that potential may not be lived up to, after all, D&D is complex and covers a lot of possible play styles.


Reforcing my previous point, I feel you should measure the monster's overall effectiveness. You can set up, say, one monster every 2 CRs that you believe is the perfect example of that type for a given level. For example, let's say the Girallon is a perfect CR6 Brute by your standards, and Elephant is the perfect CR8 Brute - they're the measuring stick you estabilished, you tested them on combat simulations and they're adequate Brutes. You can then go and look at the new Brute monsters you want to rank, and pick the 6-headed Hydra. You compare it to the Brutes you already have, and while the Hydra has low damage potential compared to both, its Fast Healing makes the Defensive side much stronger. By comparing it to your measuring sticks, you think the Hydra sits right in the middle, so CR7 is a good number for it compared to your measuring stick.

This approach is for a monster, not a single feat or feature in isolation, because these most of the time are useless. You need the context to judge something. "How much is +1 attack worth?" Well, depends on the context - a creature with 12 attacks will see a very marginal increase, while a creature with one big swing sees a doubling in damage. Same thing for the other statistics, they mean nothing in a vacuum.

[Endarire]

@OP: Pathfinder 1e and 3.5 each have a trait (worth about half a feat) that gives +2 init. Initiative is normally rolled once per creature per fight, and it's important to go before your foes, but the exact number is so variable due to it being d20-based and DEX-based. +1 initiative is something I'd like but probably ignore. +2 init for cheap might get me to take it. +3 gets my attention. +4 or more means I'm likely to get it as soon as practical, since having a high initiative means I still must do something helpful on my turn; otherwise, I wasted my time..