A proposal for boosting success rates on skills for Master and Legendary characters without unbalancing the proficiency system as a whole

Proposal:
(Conceptualized in other threads, formalized here)

Master proficiency in a skill would allow you to roll 2 D20 dice and choose the highest when making checks for that skill.

Legendary proficiency in a skill would allow you to roll 3 D20 dice and choose the highest when making checks for that skill.

The Math:

need an 11
1 die - Critical fail - 5% / Fail - 45% / Success - 45% / 5%
2 dice - Critical fail - .25% / Fail - 24.75% / Success - 65.25% / 9.75%
3 dice - Critical fail - .01% / Fail - 12.49% / Success - 73.24% / 14.26%

Need a 16
1 die - Critical fail - 25% / Fail - 50% / Success - 20% / 5%
2 dice - Critical fail - 6.25% / Fail - 50.00% / Success - 34.00% / 9.75%
3 dice - Critical fail - 1.56% / Fail - 40.63% / Success - 43.55% / 14.26%

Need a 19
1 die - Critical fail - 40% / Fail - 50% / Success - 5% / 5%
2 dice - Critical fail - 16% / Fail - 65% / Success - 9.25% / 9.75%
3 dice - Critical fail - 6.4% / Fail -66.5% / Success - 12.84% / 14.26%

Need a 20
1 die - Critical fail - 45% / Fail - 50% / Success - 0% / 5%
2 dice - Critical fail - 20.25% / Fail - 70.00% / Success - 0% / 9.75%
3 dice - Critical fail - 9.11% / Fail - 76.63% / Success - 0% / 14.26%

This drastically reduces the odds of critical failure on difficult skill checks and significantly increases the chances of Critical success without changing the actual bonuses. The advantage here is that Master characters can hit really close to that 75% success rate that people seem to want for an optimized character on a check that would be around a 50/50 split for an unoptimized character.

In many ways this is integrating Assurance into the feat, but keeping the die roll to make checks fun and still giving Master and Legendary characters the equivalent of a +2 to +4 bonus to checks (in terms of increasing the odds of success and critical success), without stretching the math or breaking the consistency of having skills utilize the same proficiency system as everything else.

This sounds like a decent proposal to me.

The main “hitch” is that it presumably wouldn’t apply to things like attack rolls or saves. So that monkeys with the desire for a uniform treatment of all of these things. (With the goal of making the system easier to learn, and to provide a uniform way of constructing numbers for challenges of all of these kinds.)

But given that these checks are already asymmetric in a number of ways — e.g., attack rolls are something you get to make repeatedly to overcome a challenge, while skill checks (and saves) are things you often only get to try once — maybe introducing *some* asymmetry into how these different kinds of checks are treated is not a bad thing.

A remarkable idea. Let me add some more lines, please...

* Untrained: Level/2 + other mods
* Trained: Level*3/4 + other mods
* Expert (and beyond): Level + other mods

* Master & Legendary: as proposed above

A remarkable idea. Let me add some more lines, please...

* Untrained: Level/2 + other mods
* Trained: Level*3/4 + other mods
* Expert (and beyond): Level + other mods

* Master & Legendary: as proposed above

I appreciate the support and thoughtful suggestion, but I am pretty strongly opposed to having level bonuses to proficiencies calculated differently. I think it adds a lot of unnecessary complexity that can be more easily accomplished through other means. (Which was the impetus behind this idea in the first place, let the math reflect what people want without stretching the numbers or complicating the process of determining bonuses). Untrained skills are already gated out of a lot of actions and activities, and with a -2 penalty as well, I think untrained is probably alright as is, or maybe some better descriptions of what should be gated out by being untrained would take care of any of those problems.

This really only leaves the separation between expert and trained looking pretty lackluster. I thought about starting the die progression at expert, but that makes a huge jump in success probability right from level 1.

Because keeping the proficiency system universal (as far as basic numbers go) is an obvious design goal of the system, I am not sure I see as elegant a solution for separating Trained from expert beyond making that bonus also a +2, (although that is a massive boost to fighters at low levels and will affect the AC numbers significantly as well.

This sounds like a decent proposal to me.

The main “hitch” is that it presumably wouldn’t apply to things like attack rolls or saves. So that monkeys with the desire for a uniform treatment of all of these things. (With the goal of making the system easier to learn, and to provide a uniform way of constructing numbers for challenges of all of these kinds.)

But given that these checks are already asymmetric in a number of ways — e.g., attack rolls are something you get to make repeatedly to overcome a challenge, while skill checks (and saves) are things you often only get to try once — maybe introducing *some* asymmetry into how these different kinds of checks are treated is not a bad thing.

I'm perfectly fine with some asymmetry. I never had a problem teaching skills to my players in PF1. Giving skills "advantage" dice seems like a workable way to make higher skill proficiencies more impactful.

This suggestion doesn't really add in any more asymmetry than is already present in different parts of the proficiency system. The bonuses are not changing at all, so it would just need to be listed in the description of skills, similar to the way that classes that get boosts to save proficiencies beyond expert usually get an additional rider as well. It also isn't something that needs to be learned until level 7, so it really doesn't impact the lower level game at all, especially with the removal of signature skills. Thus it is not something that really needs to be taught to new players, as you don't teach new players everything they may be able to do with each skill at some future point in the game.

The biggest hitch is that it makes the developer's math a little more complicated for figuring out appropriate DCs for high level challenges, since it skews the average roll for optimized characters significantly, even though it doesn't touch the range. But as long as this math is reflected correctly in DC charts, it should be pretty simple to run for GMs.

It probably also needs to not be usable in combat so it doesn't mess with combat maneuvers.

So we're ignoring the existence of untrained and expert proficiency I see.

Really it should look like this

• Untrained - Roll Twice and take the worst
  
• Trained - Roll Once
  
• Expert - Roll two times and take the best
  
• Master - Roll three times and take the best
  
• Legendary - Roll four times and take the best

Proposal:

Master proficiency in a skill would allow you to roll 2 D20 dice and choose the highest when making checks for that skill.

Legendary proficiency in a skill would allow you to roll 3 D20 dice and choose the highest when making checks for that skill.

That could be fun and I like the way it smoothes out some of the chance.

I am still in favor of some way to eliminate the die roll out of combat.
Personally, I would prefer something like below:
Untrained: Level -4, any checks you are allowed to make, you must roll.
Trained: Level+0, Out of combat you can choose to Take 5, rather than roll.
Expert: Level+1, Out of Combat you an choose to Take 10, rather than roll. In combat you must roll, but your minimum roll is a 2 (no autocrat fails due to a 1, although if your roll+your skill modifier is a crit fail then you crit fail).
Master: Level+2, Out of Combat you can take 10 rather than roll. In combat you must roll, but your minimum roll is a 5.
Legendary: Level+3, Out of Combat you can take 15 rather than roll. In combat you must roll, but your minimum roll is an 8.

So we're ignoring the existence of untrained and expert proficiency I see.

Really it should look like this

• Untrained - Roll Twice and take the worst
  
• Trained - Roll Once
  
• Expert - Roll two times and take the best
  
• Master - Roll three times and take the best
  
• Legendary - Roll four times and take the best

Lets break this down by math. Lets say this is a skill check with a DC of x+11, where the x factors in all of these characters bonuses by level, equipment, attribute and any other conditional bonus so that only proficiency exists as a modifier between them.

Long story short, being untrained in a skill is a terrible idea if you are ever going to have to roll that check for something that is a life or death situation. Especially because your numbers are probably going to slip even more from not having the equipment or the attributes to keep it at optimized levels. (This is true in the game without either of our suggestions, but yours definitely makes it even more so - instead of having a 40 % chance of success, the untrained drops to 12.25%)

Looking at a more difficult task DC 17+x, the numbers do exactly what you expect:

Notice the untrained goes from a 10% chance of success or better to a 1% chance of success or better (ouch! lets really hope nobody leaves athletics or acrobatics untrained, and you definitely do not want an untrained person trying any kind of administering first aid because their odds of killing you are going to be better than the odds of them helping you up to level8 and then they will retain almost a 10% chance of critical failure until their bonus is high enough that they can't critically fail on a 1 anymore.

I think that leaving Untrained as is and having better gates on what can't be done untrained is probably enough to separate Untrained and Trained.

Pushing the 3 levels above trained is not shattering, but pushing Legendary up to 4 dice really trivializes even difficult skill checks and even makes Master proficiency checks pretty easy against them (easier than trivial DCs currently). It might work, but probably DCs are going to have to go up across the board starting at level 3, and the passive DCs of skills like stealth might have to be adjusted based on proficiency as well, because 10+ modifier is going to be a joke, likely starting at expert proficiency.

willuwontu wrote:

So we're ignoring the existence of untrained and expert proficiency I see.

Really it should look like this

• Untrained - Roll Twice and take the worst
  
• Trained - Roll Once
  
• Expert - Roll two times and take the best
  
• Master - Roll three times and take the best
  
• Legendary - Roll four times and take the best

Lets break this down by math. Lets say this is a skill check with a DC of x+11, where the x factors in all of these characters bonuses by level, equipment, attribute and any other conditional bonus so that only proficiency exists as a modifier between them.

Here are the numbers:

** spoiler omitted **

Long story short, being untrained in a skill is a terrible idea if you are ever going to have to roll that check for something that is a life or death situation. Especially because your numbers are probably going to slip even more from not having the equipment or the attributes to keep it at optimized levels. (This is true in the game without either of our suggestions, but yours definitely makes it even more so - instead of having a 40 % chance of success, the untrained drops to 12.25%)

Looking at a more difficult task DC 17+x, the numbers do exactly what you expect:

** spoiler omitted **...

I like math.

How about another idea:

untrained rolls 1d20 (average: 10.5, 1-20)
trained rolls 2d10 (average: 11, 2-20)
expert rolls 2d8+1d4 (average: 11.5, 3-20)
master rolls 3d6+1d2 (average: 12, 4-20)
legendary rolls 5d4 (average: 12.5, 5-20)

It changes the distribution (less extremes, the higher the proficiency), but it also isn't a true d20 system anymore (much like the roll 3d20 take highest isn't a true d20 system anymore).

The 4 degrees of success would be less pronounced here, with the bell-curve distribution on the higher proficiencys though.

I like the general idea in the OP, but indeed it kinda leaves Expert hanging. I think Untrained vs. Trained is fine as-is; because in practice an untrained skill is probably going to have a worse ability bonus too, so they're probably 1-2 apart already.

I like the idea that Expert and up can Take 10 out of combat. That degree of certainty in your skill certainly matches what I expect from the word "expert". He's someone that under stable circumstances can output reliable results. But when faced with opposition things can still go wrong. A Master on the other hand distinguishes himself by being awesome even under pressure.

So that would be:

Untrained: the usual -2 penalty
Trained: no penalties or advantages
Expert: +1 and can Take 10
Master: +2, can Take 10 or roll twice and take the best
Legendary: +3, can Take 10 roll roll thrice and take the best

Assurance feat: put down by the vet.

What if you got 2D20 on any roll with a lower proficiency requirement than your current proficiency level?

Trained gets 2d20 on untrained, expert gets 2D20 on trained and untrained, and so on up to legendary.

Untrained: -2 penalty; you can attempt a trained task (or a task against an opponent's trained skill's DC), but in that case you can't crit succeed and a failure is considered as a critical failure. Just let the professionals do their job.

Expert: treat a critical failure as a failure; experts don't botch. If an opponent makes a roll against your skill's DC, he treats a failure as a critical failure (or his expertise is negated if he's expert as well).

Master: treat a success as a critical success; masters' accomplishments are always amazing. If an opponent makes a roll against your skill's DC, he treats a critical success as a normal success (or his mastery is negated if he's a master as well).

Badass: treat a failure as a success; badasses don't fail. If an opponent makes a roll against your skill's DC, he treats a success as a failure (or his badassery is negated if he's badass as well).

I like the general idea in the OP, but indeed it kinda leaves Expert hanging. I think Untrained vs. Trained is fine as-is; because in practice an untrained skill is probably going to have a worse ability bonus too, so they're probably 1-2 apart already.

I like the idea that Expert and up can Take 10 out of combat. That degree of certainty in your skill certainly matches what I expect from the word "expert". He's someone that under stable circumstances can output reliable results. But when faced with opposition things can still go wrong. A Master on the other hand distinguishes himself by being awesome even under pressure.

So that would be:

Untrained: the usual -2 penalty
Trained: no penalties or advantages
Expert: +1 and can Take 10
Master: +2, can Take 10 or roll twice and take the best
Legendary: +3, can Take 10 roll roll thrice and take the best

Assurance feat: put down by the vet.

That is pretty elegant. I like it.

I think making the whole way you do your dice rolls change depending on the relative required proficiency compared to yours, is making things more complicated than is really desirable.

Roll twice is fine, roll thrice is kind of "wrong", and unnecessary especially since you're already really good at the rolls.

How about (effects are additive):

Untrained: the usual -2 penalty
Trained: no penalties or advantages
Expert: +1, can now Take 10
Master: +2, can now roll twice at the cost of 1 RP
Legendary: +3, can now Take 10 in threatening situations

It feels like what the OP suggests would be a good option for a skill feat which requires mastery. As it stands in the playtest the master level skill feats are not especially impressive.

I like assurance, just give people with assurance and mastery in a skill the ability to choose like greater assurance improving the static results.

I would much rather this be saved for "epic or mythic" level abilities.

@OP - your suggestion is nearly identical to one I made nearly a month ago. Only I start adding extra dice on Trained.

Similar idea

Personally I would add 'roll twice' features only to Masters and Legends. That way, it sits right on the other side of the Signature Skill divide.

As far as the Assurance feat goes, welp, it could be the difference between 'taking 10' only out of combat, or even under pressure in combat. That's how assured the PC is about that skill...

Ascalaphus wrote:

I like the general idea in the OP, but indeed it kinda leaves Expert hanging. I think Untrained vs. Trained is fine as-is; because in practice an untrained skill is probably going to have a worse ability bonus too, so they're probably 1-2 apart already.

I like the idea that Expert and up can Take 10 out of combat. That degree of certainty in your skill certainly matches what I expect from the word "expert". He's someone that under stable circumstances can output reliable results. But when faced with opposition things can still go wrong. A Master on the other hand distinguishes himself by being awesome even under pressure.

So that would be:

Untrained: the usual -2 penalty
Trained: no penalties or advantages
Expert: +1 and can Take 10
Master: +2, can Take 10 or roll twice and take the best
Legendary: +3, can Take 10 roll roll thrice and take the best

Assurance feat: put down by the vet.

That is pretty elegant. I like it.

I feel like having expert/master/legendary be different for skills than it is for combat is not particularly elegant. Also it could cause problems where combat and skills overlap like grapple, shove, trip.

A more elegant solution unifies these problems.

I wish I had one :)

I feel like having expert/master/legendary be different for skills than it is for combat is not particularly elegant.

Despite having already made the same suggestion the OP has made, rolling multiple dice is not an "elegant" solution. But it is a solution that allows a greater range of skills to succeed at a task while allowing someone who has higher skill to actually feel as if they have more skill

What is "elegant" about the solution is that it allows a person to roll more dice as a reward. Don't over look that on a psychological level.

Snickersnax wrote:

I feel like having expert/master/legendary be different for skills than it is for combat is not particularly elegant.

Despite having already made the same suggestion the OP has made, rolling multiple dice is not an "elegant" solution. But it is a solution that allows a greater range of skills to succeed at a task while allowing someone who has higher skill to actually feel as if they have more skill

What is "elegant" about the solution is that it allows a person to roll more dice as a reward. Don't over look that on a psychological level.

The downside is that it slows down high-level play which is already slow.

Roll twice is fine, roll thrice is kind of "wrong", and unnecessary especially since you're already really good at the rolls.

How about (effects are additive):

Untrained: the usual -2 penalty
Trained: no penalties or advantages
Expert: +1, can now Take 10
Master: +2, can now roll twice at the cost of 1 RP
Legendary: +3, can now Take 10 in threatening situations

Ignoring for a moment that I still really loathe everything about Resonance - I like the core ideas behind the progression you laid out. The bonus D20 at Master seems like it would be rewarding without being bloated, and I like the idea that as you advance you can more freely accomplish basic tasks in more dire circumstances. Even taking a 10 out of combat at Expert feels like a reward for dedicating yourself to that skill, rather than how it is now with the Assurance feat tax.

I like it for thematic reasons, too. "Out of combat" doesn't mean "stress free", thus allowing the perks at Expert proficiency to still feel heroic without (IMO) being overpowered. And I like the idea that as you hit Master and Legendary the mechanics shift to mitigating the chance of pointless, ridiculous failure rather than ensuring blinding success every time.

I'm not sure I would go quite as far as that, since Rolling twice is really powerful, but it might make sense for legendary. Though that does make me wonder what to do for the other levels. Maybe something like this:

Untrained: -2, and cannot attempt many tasks which require trained proficiency
Trained: +0, baseline standard
Expert: +1, and you may reroll any critical failures; anything but a second critical failure counts as a regular failure
Master: +2, and you may reroll any failures; anything higher than failure counts as a success
Legendary: +3 and roll 2d20, take highest

The downside is that it slows down high-level play which is already slow.

I don't see how this is true. Maybe if you only have one d20 and have to roll it three times and remember the rolls and then pick one. If this became a thing, folk would know to bring 3d20s to a game.

At any given game table, I generally have six or more d20s.

I like the idea of the mechanic. It also fits nicely with the more dice are better feel of PF2e—where a +1 weapon means roll 2 damage dice, a +2 weapon means roll 3 damage dice, etc.

andrew

The downside is that it slows down high-level play which is already slow.

That's not a given. First off, how many skill checks are there in an average game? Ten, fifteen? How long would it take to roll a d20 a few more times? Every character is not going to have every skill check at Legendary levels, maybe one or two at most for non-Rogues.

Second, how long is a level 15 game? PFS sessions last 4-6 hours. So how much time do you think that extra rolling will add up to? 2-5 minutes, tops.

Third, getting more critical successes and fewer critical failures will probably speed up the game because you're probably avoiding combat with successful skill checks.

So no, I don't agree that it will slow down the game, it will probably shorten the game.

I find that re-rolling and multi-rolling adds time when people forget they got the opportunity. They roll once, they fail. The GM starts to tell them what happened. Another player pipes up, "don't you get to roll twice?" The player quickly stops the GM and rolls another die.

I find the same thing happens with too many buffs that players forget about.

It's not a huge amount of time, but I'd say more than 2-5 minutes and it can kill the flow of story telling. I'm not saying that it's a universal thing, but it would definitely slow things down at my table.