Deconstructing the Problem: Monster Skills, Monster Perception, and the Skills DC Table; Fixing the Item Bonus Assumption Problem

[Zman0]

Warning: This is a long and in depth thread digging into the foundational math of skill DCs, their Scaling, Monster Skills, Monster Perception, and consequently Skill Items. I apologize if any of this is difficult to digest, without access to the exact terms, tables, etc the Devs have my verbiage and jargon are a bit varied and it can be difficult to determine some of the exact details. Enjoy!

As many threads before have illustrated and pointed out, there are some problems with PC competence compared to the Skill DC table, Monster Skills, and Monster Perception. This thread is about deconstructing exactly where the problems are coming from on the back-end math. There exist some systemic issues relating to the ~50% competence calibration etc. ie where that design goal is not being met as intended. This thread isn't about the merits of the ~50% competence design decision and is about fixing the back-end math to fulfill that design goal.

We'll start by looking at what problems exist, and how the back-end math was designed and structured. Then we'll look past the symptomatic problems to isolate the root causes. Finally, we'll look at a couple of possible solutions for those problems.

Problem 1: Monster Skills

Monster's skills are better than PC skills. Monster's names skills come in varieties of specialization and are no longer determined by their ability scores. Monster's best and worst named skills rarely vary by more than 2-3 pts. Specialized Monsters scores are at least as good as a fully optimized PC and their aggregate is significantly better than PCs man manage. Unnamed Skills +X are generally equivalent to trained or better compared to a PC's and PCs cannot directly compare to them due to how they are structured with their added in ability modifiers. Monsters also do not suffer the equivalent of Armor Check Penalty.

Problem 2: Monster Perception

Monster perception vastly outpaces PC Perception and generally keeps pace with moderately optimized and focused PC skills. It is usually only a 3-4 point range among Monsters of a given CR. PCs, even optimizing Perception, have difficulty reaching Monster Perception levels or flat out are incapable of it due to their class's limitations of Trained/Expert and the dearth of higher end perception skill items.

Problem 3: Skills DC

Skills DC's according to Table 10-2 on pg 337 start at ranges that adhere to the definitions of their difficulties listed on pg 336. But, as a character levels the DCs quickly outpace all PC skills except for moderately or optimized PC's focused skills. This effect is especially noticeable for Low and High DCs. By mid to high levels most PCs will have fallen behind their rate of success on skill difficulties that they by definition should have retained competency at.

-----------------------------------

Monster Calculations

Named Monster skills are not affected by a Monster's ability scores. Only their Unnamed skills +X are affected by their listed ability modifiers.

The fundamental maximum range for the monster skill calculations assume four things: Character Level, Proficiency, Ability Modifier, and Item Modifier of a possible PC. They use some kind of tiered set of tables ie a Monster might have their effective rough equivalent of Untrained/Trained/Expert/Master/Legendary. Untrained is their Skills +X, but still scales upwards and can reach near or occasionally exceed trained levels depending on the Monster's ability scores. Untrained skills also vary significantly. Trained is the Monster's lowest possible skill which at high levels is usually three points lower than its highest skill tier. What I am calling Expert is the Trained value +1, Master is +2, Legendary is +3. Now, I'm giving these skills those names, but it may not exactly correspond to a PC's proficiency, but I'm guessing it was indented to exhibit that kind of competence.

When a monster is being created the named skills select a proficiency track and get the modifier for that proficiency at that level. Ability scores do not matter. Ie the best skill is usually equal to an optimized character of that level. Due to named skills having such a narrow range they often represent very high degrees of competence in those named skills by PC standards.

Monsters occasionally get a special modifier representing extremely specialized skill. This can put a Monster well beyond a PC's optimized capabilities. An example of this would be a Shadow's Stealth.

Skill DC Table Calculation

The 10-2 Skill DC table is built with certain assumptions: Level, Ability Modifier, Proficiency, and Item Modifier.

Trivial and Extreme DCs set the range with trivial being a DC10 check scaled only by level representing no investment in the skill and Extreme representing an unchanging check for a completely optimized PC(Ability, Proficiency, and Item) at approximately a 40% success rate.

Severe is the same as Extreme but scaled for ~55% success rate.

Low and High are our problem children; they scale quickly and stay closer to the Severe and Extreme ends of the table than Trivial representing a significant improvement outside of level. This, observably, quickly outpaces most PC's competence and ability to succeed on them. Essentially, character lose pace with these DCs despite meeting the description of those DC difficulties on Pg 336. They are calculated with scaling assumptions for Proficiency, Ability Mod, and Item Bonus, and those assumptions amount to increases nearly as high as those for Optimized characters.

-----------------------------------

Problem Causes

Smoothing

We know that when tables 10-2 were created and monsters were created they were using a now defunct progression table that due to the levels of expected Master and Legendary Proficiency and smoothing resulted in a ~+1(+0-2) discrepancy in Monster Skills, Monster Perception, and Skill DCs at many/most levels. This is universally working against PC competence. I believe this is the result of in an earlier version they had Master and Legendary proficiency at different levels ie Legendary at 13th levels, this means that there are levels that with the smoothing is likely thrown off by 1 and possibly 2.

This is a known issue among the Devs and is going to be fixed for the final release. But, we need to keep that ~1 discrepancy in mind for everything else as it acts as a magnifier.

Perception's Dual Nature

Monster Perception fulfills a weird gap between skills and what Perception is. As it is not a PC skill, most classes cannot move it beyond Trained or Expert, outside of Ranger and Rogue. But, it also acts like a skill for opposing skills etc. Monsters treat it very much like a skill and it progresses like one. This is doubly bad for PC competence. Essentially PC's have a significantly harder time sneaking against monsters than they do against them in addition to having pretty significant initiative disadvantages.

Scaling for Low and High DCs

Low and High DCs scale quickly and outpace PC's ability to compete. This is directly related to the assumptions for Proficiency, Ability, and Item bonus factored with with a big emphasis on Item Bonus. Any PC that does not have some kind of reasonably item bonus, or hasn't focused on proficiency and ability sans item bonus will quickly fall behind.

Trivial/Low/High over the course of lvls 1 through 20 goes from 10/12/14 to 29/38/41. That +0/+7/+8 scaling without level is what we need to zero in on.

"A high-difficulty skill DC can be overcome by a character who has increased their proficiency rank in a skill but doesn’t have a high score in the associated ability"

So, that +8 over the course of 20 levels assumes a modest increase in Ability score ie and or increase in proficiency. You'll notice that those modest ability and proficiency improvements only account for about half(+3-4) of that expected +8 improvement. Even if it is high by 1 because of the aforementioned smoothing problem, we're still looking at a +3-4 expected Item Bonus. Low is even worse and the situation gets extremely bad for any untrained skill. Essentially, with out high to max item bonuses in most skills PCs fall behind the Low/High DCs when they shouldn't. It looks like Devs are assuming ~+3 for Ability Mod improvement from level 1-20.

Any PC that doesn't horde skill items, or is caught without the right one invested, is going to be at a severe disadvantage. PCs that horde skill items have an outsized bonus compared to other PCs. Also, Resonance becomes a limiting factor.

When each skill is looked at individually, we can find ways for most PCs to keep up, but when aggregated and the whole breadth of PC skill competence is examined we see that most PCs, even with magic skill items, are falling behind Low/High DCs when they really shouldn't be.

Item Bonus

This is the crux of the real problem. Both Monsters Skill/Perception and the Low/High DCs on the 10-2 table assume significant item bonus scaling on virtually every single DC track outside of Trivial. Any PC that does not acquire item bonuses to their skills will fall behind. And often, they will need these bonuses to be reasonable ie +3 or higher by late levels to remain competent.

One major problem is skill item availability. There isn't universal availability of items across all ranges ie +1-+5, or the items are not available to all character types, or locked behind rarity for all skills. Some skills ie Perception have very limited item availability, limited at +2 for reasonably accessible item, but that limitation isn't factored into Monster's competence.

Any character that does not invest in skill items will fall way behind not only the Skill DCs and Monsters, but also fellow PCs who are skill item hoarders. Currently, the back end maths assumes all characters are effectively item hoarders who have all the skill items as default.

Also, one of the assumed bonus sources is Mundane skill items, but there are limited availability of these items and often they have no combat uses when compared to Monsters which can give the appearance of PC competence that falls behind Monster comparisons.

One other compounding problem is that PC's suffer armor check penalties that can be quite steep. This can really tax the expected competencies at low levels before higher quality items are acquired to mitigate ACP.

Monster Scaling

Monster's Skill/Perception scaling suffers the same item assumptions as the 10-2 Skill DC Table mentioned earlier. Worse, it assumes essentially maxed item bonus progression to each and every Monster Skill tier including their generic Skills +X. This puts Monster's capabilities for their named skill essentially hitting High to Severe DC at PC competence levels, or equivalent to an Optimized PC to Optimized -~3. So Monster Skills effectively assume an optimized PC's ability Mod, max Item Bonus for level, and effectively Trained to Legendary Proficiency. Monster's "Untrained" catch-all Skills +X +Ability level is enough to match Low DCs and has a substantial assumed item/proficiency contribution.

With all of the aforementioned problems regarding PC's abilities and the innate assumptions about item bonuses we see just how far Monsters outpace PC's abilities to compete on a skill footing. When you look at Perception, this gets even worse.

It does get worse, Monsters don't suffer an Armor Check Penalty adding insult to injury.

-----------------------------------

How do you fix these problems?

Now that we've isolated the problems and determined the root causes of those problems our next step is to look at options for fixing them. We can ignore the smoothing issue as that will be addressed on its own.

We need to address the Low/High DC scaling problem resulting in general PC incompetence in all but a few skills.

We need to address the Monster Skill problem, ideally in a way that addresses the Perception imbalance at the same time.

Suggestion 1: Lightest Touch: Change only what is needed to address the core problems:

This approach addresses the underlying issues with as light of touch as possible.

Low/High Skill DCs: The scaling of these DCs needs to be shifted to be more in line with PC expectations for those skill DCs ad defined on pg336. Currently the lvl 20 DCs for Trivial/Low/High/Severe/Extreme are 29/38/41/44/47 representing scaling from lvl 1 DCs of +0/+7/+8/+10/+10. A more viable and effective set of DCs for lvl 20 would be 29/34/39/43/47 representing scaling of +0/+3/+6/+9/+10.

Skill Items: Left largely unchanged. Fill out the gaps in availability of skill items ensuring a more equal availability of both skills and magnitude of static bonuses.

Monster Skills/Perception: Monsters effective proficiency tracks for their skills need to be revised. Monsters "Legendary" track, ie their highest ability track doesn't need to change, but their lower proficiency named skills need to be addressed. Right now named Monster skills exhibit ~4pt range, expanding that to a 6 or 8 point range would diversify named skills. Monster unnamed skills would need to be addressed down correspondingly with the bottom of lowest named skill tier. Perception could be addressed by using the Master and Legendary monster perception tracks more sparingly, focusing most Monster Perceptions off of the Trained and Expert Tiers. When coupled with more available PC Perception options this would narrow the existing gap.

Potential Problems: There still exist some substantial pitfalls of this approach including the large gulf between item hoarding PCs and those PCs who do not buy skill items as soon as they reasonably affordable. Resonance remains a problem, the weight of the required skill items on resonance is still quite high. Monster Perception is still likely to outpace PC perception by a couple of points.

Suggestion 2: Moderate Touch: Static Skill Item Rework to +3 Maximum with additional required core fixes.

This approach is largely the same as the lightest touch method, but it it limits the static Item Bonus to skills to a +3, shrinking the rank and scaling of all problem areas. Coupled with the core fixes it further reduces the existing problems.

Low/High Skill DCs: DCs are re-scaled similarly to the lightest touch fix, though with the reduced static modifier the Exteme DC is scaled down with the others below it. A reasonable set of DCs under this change would be 29/33/37/41/45 representing a level 1-20 scaling of +0/+2/+4/+7/+8.

Skill Items: Skill Items can be left largely unchanges with their magnitudes changing from +1-2 becoming +1, with +3-4 becoming +2, and with +5 becoming +3. At the end of this post I will include a pdf of skill item templates, the second page referenced default item templates adhering to this style. It offers a variety of different kind of skill item functioning.

Monster Skills/Perception: As the lightest touch method, except it closes things down further. A scaling of monster skill tiers of a 6pt range instead of the current 4pt range would work better. As above, the skills+X would need to be addressed downward to compensate.

Potential Problems: As the lightest touch, though smaller.

Suggestion 3: Heavier Touch: Proficiency Based Skill Items

This approach changes how skill items function and structurally changes how DCs are calculated. It shrinks the absolute range maximum end of the range to +2, but offers potentially higher bonuses for the least skilled characters.

Proficiency Based Items: If a character has a Proficiency bonus lower than the of the item, their Proficiency Bonus is calculated as if the character had that item's level of proficiency. If the character's proficiency is equal or greater than the items they receive a +1 Item bonus instead. If the character and item are both of Legendary Proficiency, they receive a +2 Item bonus instead.

Skill Items: Items follow the basic outline above with +1-2 Items becoming Expert Items, +3-4 Items becoming Master Item, and with +5 items becoming Legendary Items. In the pdf below I've made Item templates using Proficiency.

Low/High DCs: These DCs are shifted to be more in line with the moderate touch DCs. Reasonable DCs for this modification would be lvl 20 DCs of 29/33/37/41/44 representing scaling for levels 1-20 of +0/+2/+4/+7/+7.

Monster Skills/Perception: Scaled similarly to the moderate touch. The Tier tracks will shift down 1 due to the reduced maximum skill item bonus. A level 20 6pt range in skills should be sufficient to vary Monster skills as well as the corresponding reduction in Skill +X. This variation eliminates completely virtually and problematic distinction between Monster Skills and Perception in relation to PC's capabilities.

Potential Problems: If implemented this method has the fewest potential problems as far as the existing fundamentals are concerned. It has the greatest ability to affect the outlined problems. Potential new problems created are similar to the moderate hand set of fixes which is reducing the total scale and range of potential skill modifiers at all levels. Some will view this as a bug, others will view it as a feature.

Skill Item Templates

[b]Bonus Fix: Armor Check Penalty[/roll]

ACP is extremely punishing at low levels for heavy armor wearers and is needlessly complex.

Problems: ACP is hugely punishing to PCs at low levels. It is quite varied, changes often, and is ultimately greatly reduced to inconsequential by high levels. Heavy Armor and especially Full Plate is some of the worst armor available long term. Heavy Shields are vastly more desirable than Light Shields.

Suggested Fix: Light Armor has an ACP of -0. Medium Armor has an ACP of -1. Heavy Armor has an ACP of -2. Light Shields have an ACP of -0. Heavy Shields have an ACP of -1. Superior Armor quality no longer reduces ACP, instead it adds to AC in the same way weapons attack modifier is increased. Like Magic Weapons, that modifier would be overwritten by potency runes.

Potential Problems: This cramps some of the design space around armor differences. Virtually all are fine with these changes except for Splint which needs to be modified. Something along the lines of a 100sp cost and only a 5' speed penalty would suffice. Mithril Armor will become more desirable to reduce the ACP of of Medium of 0 as well as reduce that for Heavy Armor to -1.

[RazarTuk]

Some bonus statistics from 1e, for comparison:

Monsters gain 1.2 HD/CR, 1.15 AC/CR, 0.9 Will/CR, 1.6 to-hit/CR, and about 0.96 DC/CR for specials (based solely on HD and Cha).

Full BAB classes gain 1.125 to-hit/CR without magic or about 1.625 to-hit/CR with magic. 3/4 BAB classes focusing on combat gain 0.825 to-hit/CR without magic or 1.125 to-hit/CR with magic. Casters get about 0.8 DC/CR, and a Cleric (so a Wis-based character with good Will) gets 0.8 Will/CR.

(These numbers assume level-based stat increases go to the relevant stat. It's assumed that you get +2 to a stat from magic about every 6 levels and +1 on weapons every 3 levels. And for ease of math, this ignores everything that isn't level, stat-boosting items, and magic weapons)

Armor Bonus + Max Dex tends to hover between +8 and +10. And as an interesting statistic, related to the AC/CR number, I scraped data on natural armor, dexterity, and BAB from Nethys, ran a regression on it, and found that among monsters with a natural armor bonus that don't wear manufactured armor, Natural Armor ~= 2.5 + BAB - Dex. In other words, natural armor follows the same law of Dex+Armor staying roughly constant, but increases with BAB to stay a challenge for PCs. (Although given the other numbers, full BAB classes will readily outpace it)

[Zman0]

RazarTuk wrote:

Some bonus statistics from 1e, for comparison:

[spoiler]Monsters gain 1.2 HD/CR, 1.15 AC/CR, 0.9 Will/CR, 1.6 to-hit/CR, and about 0.96 DC/CR for specials (based solely on HD and Cha).

Full BAB classes gain 1.125 to-hit/CR without magic or about 1.625 to-hit/CR with magic. 3/4 BAB classes focusing on combat gain 0.825 to-hit/CR without magic or 1.125 to-hit/CR with magic. Casters get about 0.8 DC/CR, and a Cleric (so a Wis-based character with good Will) gets 0.8 Will/CR.

(These numbers assume level-based stat increases go to the relevant stat. It's assumed that you get +2 to a stat from magic about every 6 levels and +1 on weapons every 3 levels. And for ease of math, this ignores everything that isn't level, stat-boosting items, and magic weapons)

Armor Bonus + Max Dex tends to hover between +8 and +10. And as an interesting statistic, related to the AC/CR number, I scraped data on natural armor, dexterity, and BAB from Nethys, ran a regression on it, and found that among monsters with a natural armor bonus that don't wear manufactured armor, Natural Armor ~= 2.5 + BAB - Dex. In other words, natural armor follows the same law of Dex+Armor staying roughly constant, but increases with BAB to stay a challenge for PCs. (Although given the other numbers, full BAB classes will readily outpace it)[/spoiler/

Ummm, how is this relevant to my post? Literally nothing you've listed pertains to the topic of this thread. Nothing in the post deconstructs to hit, or AC, or Saves, etc. This is specifically about the back-end math the skill system specifically Monster's vs PCs and PCs vs the Skill DC table.

You listed them for comparison, but comparison to what? I just don't see the relevance.

[Laik]

Thanks for the analysis!
I really hope all that gets addressed in official errata, playtesting currently goes wrong in many aspects exactly because of these bestiary faults.

[ChibiNyan]

Have you thought about simply not factoring item bonuses at all in table 10-2 and, as a result, monster skills? As you said, some of those items are pretty rare and all skills are not equally represented. It is also unlikely that a character will be equipping many of them at the same time.

Doing this would mean that a character with skill items would have a sizeable advantage in the 1 or 2 skills he has item bonuses on, representing them going "above and beyond" their expected competency. This still allows characters without the item (Common even at high level) to succeed at some of the tougher tasks.

EDIT: Or maybe just consider them for Extreme difficulty tier, to represent that only the most optimized character has a shot, but still not accounted for in the lower tiers.

[Dante Doom]

Nice analysis!

I hope this get strated as soon as possible

[Zman0]

Laik wrote:

Thanks for the analysis!

I really hope all that gets addressed in official errata, playtesting currently goes wrong in many aspects exactly because of these bestiary faults.

You are welcome!

Not sure they'll tackle a lot of this in the errata. I'd like them to implement Suggestion 1 before the end of the playtest requiring a new 10-2 Skill DC Table and a whole new version of the Bestiary to be released. Not sure they will do that though, but I have some confidence they will address these problems in some before by final release. The real question will be the magnitude of those changes.

[Zman0]

ChibiNyan wrote:

Have you thought about simply not factoring item bonuses at all in table 10-2 and, as a result, monster skills? As you said, some of those items are pretty rare and all skills are not equally represented. It is also unlikely that a character will be equipping many of them at the same time.

Doing this would mean that a character with skill items would have a sizeable advantage in the 1 or 2 skills he has item bonuses on, representing them going "above and beyond" their expected competency. This still allows characters without the item (Common even at high level) to succeed at some of the tougher tasks.

EDIT: Or maybe just consider them for Extreme difficulty tier, to represent that only the most optimized character has a shot, but still not accounted for in the lower tiers.

Sure! What you are suggesting is fairly close to my Suggestion 1, though it basically only applies to Low, High has a little bit factored in. This approach lies closer to my desires, but I tried to create these suggestions with what I felt was the design goal, and that obviously assumes some level of general item bonus. Basically what you are looking for would be the equivalent of a DC 29/33/37/42/47 which I am more than ok with. This actually wouldn't quite meet what you want, but is the closes I can do by equally splitting that 18 point gap. You might want more a 29/32/36/40/45.

Also, not factoring item bonus in at all for Low/High checks really widens the gulf between the Low/High DCs and the Severe/Extreme DCs. This isn't the worst thing, but it could result in High DCs becoming fairly automatic for semi-specialized and specialized characters.

The problem comes in when the character doesn't have one or two skills they have items in, but spend what becomes relatively small amounts of GP to pick up a plethora of lower tier items. Ie the level 15 character who picks up a dozen or so of the +2-3 items. Those item hoarders can really shine compared to PCs that don't. Resonance may or may not fix this when a check comes up outside of combat they can just invest in it.

My Suggestion 1 really kind of tries to ride the line here and honestly errs on falling on the Dev's original design side. Essentially, its a conservative suggestion of change.

[Zman0]

Dante Doom wrote:

Nice analysis!

I hope this get strated as soon as possible

Thanks! I tried to sort all of the thoughts rolling around into something digestible. A lot of this I've been saying here and there around the forums. I even made a suggested Monster skill thread where I tried to keep Ability scores etc, but it was too big of a change from the original design. So, I re-geared, accepted the design direction and tried to work with them than against them.

In my ideal world, we'd get the Suggestion 3 changes. Proficiency based items would really just fix all of these problems. And an arbitrary +2 bonus doesn't mean as much as your Untrained barbarian investing in the Master Lute and suddenly displaying high end competence and playing masterful tunes. This gets even better if the work on making more skill proficiency gates, or better defining them. IMO its a way better way of representing the skill granting benefits of an item than a simple arbitrary static modifier. And did I mention how elegant it works with the back-end math. It just works and lots of problems just go away.

[Dasrak]

I think the problem is most succinctly illustrated by deconstructing the monsters. In PF2 monsters don't have to strictly account for where all their bonuses are coming from; the author can just set "appropriate" numbers (scare quotes intentional, because I don't think the numbers being chosen are appropriate at all). I'm not here to argue for or against that, only to say that we can still deconstruct these monsters and show how they differ from a straight calculation based on the monster's level and ability score modifiers. This will tell us how beefed up from that baseline these monsters really are.

For this purpose, I'll randomly pick a monster from the bestiary. I flip through the book and choose the Ettin, a 6th level creature:

[format is calculated / actual (difference)]
Perception: 6 / 14 (8)
Acrobatics: 5 / 9 (4)
Athletics: 11 / 13 (2)
AC: 15 / 20 (5)
TAC: 15 / 18 (3)
Fort: 7 / 14 (7)
Ref: 4 / 9 (5)
Will: 5 / 11 (6)
Melee: 12 / 14 (2)
Damage: 1d6+5 / 2d6+5 (1d6)

Since the Ettin is 6th level it could be expert in many of these categories, but obviously an additional +1 doesn't come close to closing the distance in most of these cases. This Ettin is getting huge racial bonuses to literally everything. You can repeat this on pretty much every monster and get the same result, and these discrepancies show just how beefed up monsters really are. While I understand and appreciate the "quick creation" approach PF2 is taking for monsters, I do think that it masked the problem of how excessive monster skills, saves, and abilities really are.

[Zman0]

Dasrak wrote:

I think the problem is most succinctly illustrated by deconstructing the monsters. In PF2 monsters don't have to strictly account for where all their bonuses are coming from; the author can just set "appropriate" numbers (scare quotes intentional, because I don't think the numbers being chosen are appropriate at all). I'm not here to argue for or against that, only to say that we can still deconstruct these monsters and show how they differ from a straight calculation based on the monster's level and ability score modifiers. This will tell us how beefed up from that baseline these monsters really are.

For this purpose, I'll randomly pick a monster from the bestiary. I flip through the book and choose the Ettin, a 6th level creature:

[format is calculated / actual (difference)]
Perception: 6 / 14 (8)
Acrobatics: 5 / 9 (4)
Athletics: 11 / 13 (2)
AC: 15 / 20 (5)
TAC: 15 / 18 (3)
Fort: 7 / 14 (7)
Ref: 4 / 9 (5)
Will: 5 / 11 (6)
Melee: 12 / 14 (2)
Damage: 1d6+5 / 2d6+5 (1d6)

Since the Ettin is 6th level it could be expert in many of these categories, but obviously an additional +1 doesn't come close to closing the distance in most of these cases. This Ettin is getting huge racial bonuses to literally everything. You can repeat this on pretty much every monster and get the same result, and these discrepancies show just how beefed up monsters really are. While I understand and appreciate the "quick creation" approach PF2 is taking for monsters, I do think that it masked the problem of how excessive monster skills, saves, and abilities really are.

And if we build Monsters your way that Ettin no longer is a CR6, it’s much much weaker. Maybe it’s only a CR3, so you drop it and all it’s values drop again, etc. So that doesn’t work. Those racial bonuses you’re referencing are actually approximations of the host of things PCs get. Or at lvl 6 they have +1 weapons accounting for your damage due and partial to hit. They’ve got magic armor for AC and saves. Would you rather we have to give the Ettin +1 Hide? From a feel perspective I like the veneer of understanding how the monster was created, of playing by the same rules. But, when it doesn’t result in usable monsters and if the other set of rules makes better monsters faster that work better, I’m going that route. As long as the PC rules and Monster rules play together in the way we want, do we need the veneer of making them look the same when they haven’t been since even 3.5.

To build them your way we create tables for every thing, then we add ability mods to them, then we tweak them etc. What ends up happening is we highs that are much higher, and lows that are much lower and overal balance and matching cr expectations becomes much tougher. And you’d likely need padding here or there, kind of like natural armor in 3.P. Or worse, we have to build them just like PCs and we get no real benefit and expend an inordinate amount of time doing so.

The way monsters are being built now is meant to ensure that their abilities fall within the appropriate ranges for their challenge rating. The fine balancing act of trying to get monster abilities exactly right it gone, as is the need to really fudge the numbers. Building a cr appropriate monster this way will be much faster and easier and yield a result that is much more usable. Can you argue the 3.5 or P1 method resulted in cr appropriate monsters? Or did PCs often hit way above their weight class? How many weird natural armors have you seen etc?

This thread was about addressing the Monster rules and asssumotuons that happen behind the curtain. For now, do we need to peak behind the curtain, or do we need monsters that function. For once I’d like monsters that were “useable” at their cr guidelines and IMO this is going to give us the best result even if the veneer of parity peels off.

This thread wasn’t about challenging the base monster creation methodology, but improving th one decided upon.

[Deadmanwalking]

I don't agree with all your solutions (I think something on the order of moderate but with somewhat bigger number changes and making skill items actually available is probably the best solution), but your analysis of what the problem is, is pretty much spot on.

I will note that the 'known problem' is actually as bad as a 3 point swing in monster skills (it's +3 to all their skills at 13th level), though it's -1 or -2 most other levels. So a bit worse than you state.

I will also note that any solution that actually uses items needs to make such items actually available. Which they aren't. Most skills lack mundane skill items entirely, and at most levels maxed out magic skill items are also only available for a handful of skills. This combines to make the problems you note with skill items notably worse.

[ChibiNyan]

Zman0 wrote:

Dasrak wrote:

I think the problem is most succinctly illustrated by deconstructing the monsters. In PF2 monsters don't have to strictly account for where all their bonuses are coming from; the author can just set "appropriate" numbers (scare quotes intentional, because I don't think the numbers being chosen are appropriate at all). I'm not here to argue for or against that, only to say that we can still deconstruct these monsters and show how they differ from a straight calculation based on the monster's level and ability score modifiers. This will tell us how beefed up from that baseline these monsters really are.

For this purpose, I'll randomly pick a monster from the bestiary. I flip through the book and choose the Ettin, a 6th level creature:

[format is calculated / actual (difference)]
Perception: 6 / 14 (8)
Acrobatics: 5 / 9 (4)
Athletics: 11 / 13 (2)
AC: 15 / 20 (5)
TAC: 15 / 18 (3)
Fort: 7 / 14 (7)
Ref: 4 / 9 (5)
Will: 5 / 11 (6)
Melee: 12 / 14 (2)
Damage: 1d6+5 / 2d6+5 (1d6)

Since the Ettin is 6th level it could be expert in many of these categories, but obviously an additional +1 doesn't come close to closing the distance in most of these cases. This Ettin is getting huge racial bonuses to literally everything. You can repeat this on pretty much every monster and get the same result, and these discrepancies show just how beefed up monsters really are. While I understand and appreciate the "quick creation" approach PF2 is taking for monsters, I do think that it masked the problem of how excessive monster skills, saves, and abilities really are.

And if we build Monsters your way that Ettin no longer is a CR6, it’s much much weaker. Maybe it’s only a CR3, so you drop it and all it’s values drop again, etc. So that doesn’t work. Those racial bonuses you’re referencing are actually approximations of the host of things PCs get. Or at lvl 6 they have +1 weapons accounting for your damage due and partial to hit. They’ve got magic armor for AC and saves. Would you rather we...

What you say is true for some values such as AC, Attack and Damage, which are rpetty important for monsters to stay relevant. But their skills? Those are the ones with the most ridiculous inflation and it's not like the players can boost many of them very reliably. Also they have little impact when you're just trading blows with the monster, but a lot of (unfair) impact when you're trying other stuff, like sneaking or negotiating. Maybe the perception is form it having 2 heads (though it's kinda the same as most level 6 mons anyways), but it could maybe have a lower Will Save, Acrobatics or TAC for sure.

[Zman0]

Deadmanwalking wrote:

I don't agree with all your solutions (I think something on the order of moderate but with somewhat bigger number changes and making skill items actually available is probably the best solution), but your analysis of what the problem is, is pretty much spot on.

I will note that the 'known problem' is actually as bad as a 3 point swing in monster skills (it's +3 to all their skills at 13th level), though it's -1 or -2 most other levels. So a bit worse than you state.

I will also note that any solution that actually uses items needs to make such items actually available. Which they aren't. Most skills lack mundane skill items entirely, and at most levels maxed out magic skill items are also only available for a handful of skills. This combines to make the problems you note with skill items notably worse.

See, I'm of the mind that flat out static modifier skill items are the real problem that needs to go away. I personally like Suggestion 3 the best with the proficiency based skill items. I think with proficiency gating skills etc it offers the best way to represent a skill item in the game.

3pt swing. I'm seeing +26 at the high skill for monsters at that level. We need to account for a +13 outside of level. That is +5 Ability, +2 Proficiency, and +4 Item. I get max PC at +11. And that +2 comes from Proficiency being one too high at that level and from ability being smoothed out incorrectly. Ability Mod assumes the Ability Item at level 14 which pulls the smoothing upwards and it part of their baseline problem.

So, I see a 2pt swing, I just can't seem to account for that third point. But it is very possible that I am wrong there and it is as bad as +3 on that one level, but the most part the 0-2 should stand.

I absolutely agree, items need to be available. That was on reason I created the item templates. If those could be codified they allow us ways to create skill items at each level using a logical pattern. Hell, we could even use them to create a generic Skill item to fill in the gaps, you just plug in the particular skill when selecting the item and flavor it to match.

[Zman0]

ChibiNyan wrote:

What you say is true for some values such as AC, Attack and Damage, which are rpetty important for monsters to stay relevant. But their skills? Those are the ones with the most ridiculous inflation and it's not like the players can boost many of them very reliably. Also they have little impact when you're just trading blows with the monster, but a lot of (unfair) impact when you're trying other stuff, like sneaking or negotiating. Maybe the perception is form it having 2 heads (though it's kinda the same as most level 6 mons anyways), but it could maybe have a lower Will Save, Acrobatics or TAC for sure.

Ohh, if you poke around you'll find my skills fix suggestion that uses Ability Mod for Monsters named skills. This thread is addressing the system in place and conforming with the dev's design strategy.

Skills are important, just not quite as important as attack modifier or AC. There are things that target skills, and since PCs have essentially an action economy advantage and a skill diversity advantage you can create some imbalances that make a monster really suffer as an adequate CR threat.

Remember, that the listed Perception value for an Ettin at level 6 is way way too high right now for all of the reasons in the original post. When fixed, it will be brought back down to earth. I wouldn't be surprised if it ended up being more like +10 instead of +14. They gave it named Acrobatics, which made it bad for a named skill. It definitely would have been worse had it not been named. Again, it'd be a bit lower after changes etc. For saves you can see it has one good, one medium, and one poor save. If you used Ability Mod it's Fort save would go through the roof and it be extremely weak to Ref and Wis, possibly to the point it starts to strain the ability to balance the monster. Saves would probably be +13/+7/+7 if it had E/E/T and the +1 level bonus for Magic.

I'm not here to assess the validity of how the devs chose to design monsters, just take the design strategy they have and work the kinks out. Though, I wouldn't mind another thread to discuss the merits of the different monster design strategies. The biggest problem I have is that devs were always tweaking monsters when needed to adhere to CR, and when they didn't the CR system became a joke. How do you balance the veneer of parity with functionality on their CR.

[Dasrak]

Zman0 wrote:

And if we build Monsters your way that Ettin no longer is a CR6, it’s much much weaker. Maybe it’s only a CR3, so you drop it and all it’s values drop again, etc. So that doesn’t work.

Alternately, maybe the standards for what constitutes a level 6 monster have been set unreasonable high? The PF1 Ettin was also a CR 6 monster with an almost identical ability score spread, and didn't have anywhere near as many racial adjustments as this (just +4 perception and +8 natural armor). The sheer number of adjustments necessary to get the baseline creature up to where it "should" be is the very thing I'm drawing to attention here. These aren't small +1 patches here and there, nor are they limited to a targeted +5 to something critical like AC. It's big bonuses everywhere. And having that visible in front of you makes the stat inflation very obvious.

Zman0 wrote:

Those racial bonuses you’re referencing are actually approximations of the host of things PCs get. Or at lvl 6 they have +1 weapons accounting for your damage due and partial to hit. They’ve got magic armor for AC and saves.

The Ettin explicitly is wielding expert-quality weapons; I took that into account when reverse-engineering his attack. Magical armor is only +1 at these levels, so that doesn't even come close to accounting for the saves.

[Zman0]

Dasrak wrote:

Zman0 wrote:

And if we build Monsters your way that Ettin no longer is a CR6, it’s much much weaker. Maybe it’s only a CR3, so you drop it and all it’s values drop again, etc. So that doesn’t work.

Alternately, maybe the standards for what constitutes a level 6 monster have been set unreasonable high? The PF1 Ettin was also a CR 6 monster with an almost identical ability score spread, and didn't have anywhere near as many racial adjustments as this (just +4 perception and +8 natural armor). The sheer number of adjustments necessary to get the baseline creature up to where it "should" be is the very thing I'm drawing to attention here. These aren't small +1 patches here and there, nor are they limited to a targeted +5 to something critical like AC. It's big bonuses everywhere. And having that visible in front of you makes the stat inflation very obvious.

Zman0 wrote:

Those racial bonuses you’re referencing are actually approximations of the host of things PCs get. Or at lvl 6 they have +1 weapons accounting for your damage due and partial to hit. They’ve got magic armor for AC and saves.

The Ettin explicitly is wielding expert-quality weapons; I took that into account when reverse-engineering his attack. Magical armor is only +1 at these levels, so that doesn't even come close to accounting for the saves.

Well, yes, they have been. As I've explained in the entire first post of this thread that Monster Skills and Perception have been miscalculated. The actual numbers once fixe will be lower, and the situation gets better.

Are you holding up P1's CR system as an adequate gauge of threat? This is kind of comparing apples to oranges here. In P2, a CR4 monster is supposed to be equivalent to a lvl 4 character. Fighting one on one, they should be an Extreme threat, which means its pretty close to a 50:50 fight. PCs still have certain advantages like consumables etc.

Also, P1's CR system was notoriously insufficient at gauging appropriate threat. Monster's in P1 were often padded ie natural armor to make them competitive at their supposed cr.

If you calculate the Ettin using your methodology, it is no longer a CR6 enemy according to P2's definition of what CR is supposed to mean. Yes, you could change what a CR actually means by definition, but then you have to change so much more and you're going to be back to 3.P's padded monsters trying to keep up to whatever your definition means. This method doesn't really work out and nicely as I think you mean it too.

This is not the thread to be hashing out how monsters are constructed in P2. That deserves its own thread. This one is about taking the design decision the devs have gone with and making it work as intended.

[swordchucks]

The monster stats really are a mess. I don't have much trouble out of AC Attack and Damage as I like hitting things as the GM (and I like PCs being able to hit things, too), but the skills are nuts. Part of the problem is that they're clearly designed around an optimized PC, but even then they're brutally high. (The DD manticore encounter is a good example where the optimized stealth character had a 35% chance of success.)

Other skill checks are also weirdly difficult for specialists, which makes them nearly impossible for everyone else. Because the swing between "untrained" and "best possible" is so short a road, I'd rather see the tuning be for the middle of that range instead of the best possible.

[Mark Seifter]

Interesting analysis, thanks!

[Dasrak]

Zman0 wrote:

Are you holding up P1's CR system as an adequate gauge of threat? This is kind of comparing apples to oranges here. In P2, a CR4 monster is supposed to be equivalent to a lvl 4 character. Fighting one on one, they should be an Extreme threat, which means its pretty close to a 50:50 fight. PCs still have certain advantages like consumables etc.

That's roughly true in PF1 as well. An NPC build with full PC rules (20 point buy and PC WBL) had a CR equal to his level.

Zman0 wrote:

Also, P1's CR system was notoriously insufficient at gauging appropriate threat. Monster's in P1 were often padded ie natural armor to make them competitive at their supposed cr.

It was close enough. There were over-CR'd and under-CR'd exceptions, but "it's just a big brute who hits things" monsters were usually pretty consistent. Ettin is no exception, and in fact he's almost perfectly on par with guidelines for a CR 6 monster. The only stat he has that isn't on point is his reflex save and ranged attacks, which were 3 points below guidelines.

Zman0 wrote:

If you calculate the Ettin using your methodology, it is no longer a CR6 enemy according to P2's definition of what CR is supposed to mean. Yes, you could change what a CR actually means by definition, but then you have to change so much more and you're going to be back to 3.P's padded monsters trying to keep up to whatever your definition means.

As I said up front: "I'm not here to argue for or against that, only to say that we can still deconstruct these monsters and show how they differ from a straight calculation based on the monster's level and ability score modifiers." I'd be happy to discuss the finer points of the monster creation system in another thread were it would pull this off topic (long story short: I have a mixed opinion), but that's not what I'm getting at here.

The point is that we can see where those adjustments are, and how big they are. Both are alarmingly large values, and greater in magnitude than can be explained as equalizing to PC gear and proficiency bonuses. Being aware that the Ettin is getting a +6 bonus to will saves on top of its level and ability score modifier tells us a lot about the monsters - that what they get for free as a matter of course in literally every attribute exceeds what the PC's are getting with full optimization.

[Zman0]

Dasrak wrote:

Zman0 wrote:

Are you holding up P1's CR system as an adequate gauge of threat? This is kind of comparing apples to oranges here. In P2, a CR4 monster is supposed to be equivalent to a lvl 4 character. Fighting one on one, they should be an Extreme threat, which means its pretty close to a 50:50 fight. PCs still have certain advantages like consumables etc.

That's roughly true in PF1 as well. An NPC build with full PC rules (20 point buy and PC WBL) had a CR equal to his level.

Zman0 wrote:

Also, P1's CR system was notoriously insufficient at gauging appropriate threat. Monster's in P1 were often padded ie natural armor to make them competitive at their supposed cr.

It was close enough. There were over-CR'd and under-CR'd exceptions, but "it's just a big brute who hits things" monsters were usually pretty consistent. Ettin is no exception, and in fact he's almost perfectly on par with guidelines for a CR 6 monster. The only stat he has that isn't on point is his reflex save and ranged attacks, which were 3 points below guidelines.

Zman0 wrote:

If you calculate the Ettin using your methodology, it is no longer a CR6 enemy according to P2's definition of what CR is supposed to mean. Yes, you could change what a CR actually means by definition, but then you have to change so much more and you're going to be back to 3.P's padded monsters trying to keep up to whatever your definition means.

As I said up front: "I'm not here to argue for or against that, only to say that we can still deconstruct these monsters and show how they differ from a straight calculation based on the monster's level and ability score modifiers." I'd be happy to discuss the finer points of the monster creation system in another thread were it would pull this off topic (long story short: I have a mixed opinion), but that's not what I'm getting at here.

The point is that we can see where those adjustments are, and how big they are. Both are alarmingly large values, and greater in magnitude than can be...

But in P1 it largely wasn't the case. Monsters often fell behind their cr expectations. Sure, building a NPC using full PC rules would, but did Monsters routinely live up to those recommendations, especially using moderate to higher optimization characters against them?

So the Ettin was below recommendations, just not by that much. In a different system. That was known for not meeting CR expectations in practice. And required padding on statistics for many monsters. See, take that Ettin in P1, it had +8 Natural armor? Why? It was padded to reach an adequate AC for its HP and CR.

P2 is a different animal. If that Ettin needs to play at a CR 6 level it needs to be equivalent of say a Dwarven Fighter with a two handed bastard sword. We end up with the Ettin having similar saves, +1 attack mod, more attacks at less damage, and lower AC. Overall, they are pretty darn comparable from a combat standpoint. Right now the Ettin's skills as a whole are much better, but once the systemic issues I've pointed out above are addressed it'll be better. If we start building it like you want we need to make a lot more changes to the Ettin. For instance, a PC will have +12 or +13 in ability mods, but the Ettin has only a +5. Without padding the Ettin just doesn't compete.

If your goal is to figure out much this way of calculating monsters, then you need to first account for the differences in the base systems. The Monsters would need significantly more ability mods over levels, and they would need the benefits feats etc. If you want to make those comparisons, the go ahead. Make a thread, and dig into the issue and the pros and cons of constructing monsters in that way. If we start looking hard we can find all the places that P1 hid their bonuses ie extra hit dice for monsters resulting in additional BAB, all to keep their attack modifiers close to level appropriate and HP near level appropriate. All that natural armor tacked onto whatever veneer of armor they gave them just to get their AC to level appropriate. Etc Etc.

But, this really isn't the thread for that analysis. Looking at the Ettin example, the designer could easily have chosen to not name a Dex skill as named, and skills would have gotten much closer to making sense. It boils down to the wild variety in monster abilities makes getting their relevant statistics into level appropriate ranges very hard. The highs are higher and the lows are lower. The P2 method tightens that down. Maybe too much, but this really isn't the thread for that. This one is about making sure the numbers they are getting are viable.

[don't wake baby]

Zman0 wrote:

Warning: This is a long and in depth thread digging into the foundational math of skill DCs, their Scaling, Monster Skills, Monster Perception, and consequently Skill Items. .

Thank you, this is an exceptionally well written post.

Based on experience,

With #1, I had a second level character attempting a DC 17 dying save. Death was pretty much automatic with this.

#2 made stealth a useless skill - unless I was trying to hide from party members

Lastly #3 means that tumble through doesn't work and that Cat Fall becomes a required pick because you are going to fall when climbing (up or down).

Whatever the intended outcome, the result is that the game is less enjoyable than it could be. The character that could sneak into the Dragon's lair to tell the party what was inside, no longer exists. The acrobat who would tumble through combats in order to gain advantage on the enemy, no longer exists. The hero who would climb castle towers to rescue the princess, fell to his death a long time ago.

It does need to be fixed and we should focus on what the outcome should be - characters that can do more than smash, cast spells, and die.

[Zman0]

Thanks! I'm happy you found it useful.

Actually, the Dying save probably isn't really related to these numbers. The monster DC etc should be right. A DC17, that should be what a level 3-4 monster?

Yep, stealth is the biggest loser in the paradigm. Monster Perception is too high, and Stealth is awful tough to keep up.

At mid to late level, for sure. And with ACP at early levels. ACP really is too steep and my "Bonus Fix" goes a long way to make things better.

Though, I know the devs are fully aware of these problems and will address them. Can't say if it will go far enough for some of our preferences, but they are on it. Not sure if we'll get a new copy of the bestiary at any point in time before the end though. At least it doesn't impact too too negatively the low level stuff. Mid to high level and those extra +1-3 are going to be killers.

[Zman0]

Spreadsheet with Suggestion 1, Suggestion 2, and Suggestion 3 calculated with appropriate 10-2 Skill DC Table Replacement, and Monster Stats.

Fixed PC Skill and Monster Math by Zman

I found a little time between clients today to toss together a spreadsheet outlining my essential recommendations in the first post. Unfortunately I can't edit it in.(Paizo, please let us edit posts). It has all the new Skill DCs and what my suggestions would essentially do to Monster skill mods and perception.

I did end up deciding that Monster Skill bonuses for varying tiers could just be an reflection of the PCs 10-2 skill table. It resulted in very usable values.

[Dante Doom]

Really nice spreadsheet! Hope the Devs get to look it!

[Zman0]

Dante Doom wrote:

Really nice spreadsheet! Hope the Devs get to look it!

Thanks! I know that it isn't telling the Devs anything they don't already know. We might disagree on some of the finer points, but what we should see from them coming down the pike should fix most of these problems. They might not be perfect and make mistakes, but they're competent. I've got confidence in their abilities.

[Starfox]

This is the elephant in the room, the one big thing that is bad with PF2. This needs to change. I wonder if a blanket modifier will be introduced as we get into the later playtest scenarios?