While I don't particularly like this idea for prof. I feel the need to point out some problems with your analysis. The analysis is looking at things that are not real. ie a character can't be expert, master, or legendary at level one. When doing an analysis you need to consider only the things that can be true at that level.
After a few fixes (I'm looking at you bag of holding) that where done to resonnance, I also found that I was liking it, even with an alchemist (although I think the alchemist needed be something like Cha + int + level for resonnance. Other things like treat wounds also helped to offset the original perceived problems with the system.
Ok, I see what your saying and I mostly agree with you. The 0.8 x 3 makes your 3 a 2.4 and the 0.8 is a little high but close enough.
I'm not sure your giving bombs enough damage credit. An acid flask may do less upfront damage, but the target must spend actions or roll well otherwise they will continue to take damage every turn. I'm pretty sure that beats out cantrips.
I agree with ZmanO. I read, "For purposes of resistances and weaknesses, this is considered a single Strike", to account for the condition drop too (although it doesn't explicitly say it.)
I think the way healing works currently seems pretty good, but I will admit I think treat wounds should have a limit per day as does seem op. There are feats like toughness to increase rest healing if you want a boost to rest healing.
First off I'm not sure why anyene would want to stow a bow in the 'loaded' position. The rules already say that drawing an arrow can be done in the same action as firing the bow. Quick strick allows you to draw a weapon and strike with it. So you can use quick action on a melee or ranged weapon.
Rogue tent to die because they are fairly low hp, light armor, they have to move into (usually bad) positions to do any serious damage and when they do good damage they become targets of monsters designed to attack tanks with lots oh how and heavy armor.
The difference between athletics (strength based skills) and acrobatics (dexterity based skills) seem pretty decent to me. The athletics skills are more strength based skills and the acrobatics skills are more finesse or dexterity based skills. Most everything people do requires some strength and some dexterity. Even walking, requires muscles to move and balance to stay up right.
Agile makes your to hit close the gap, and some abilities like backstab, deadly and etc close the gap even more. This time there several good reasons to double wield...
This point about the 1 handed weapons traits making the weapons better doesn't seem to add up. It appears to me that 2 handed weapons default to the damage of a d12 (weapons below a d12 get more traits to boost them to damage closed to a d12) while 1 handed weapons default to the damage of a d8 (only weapons with less than a d8 have any of the traits you mentioned).I'll grant you that there are class specific abilities that can make 2 weapon fighting a little closer to 2 handed fighting, but its only for rangers and while it gives 4 attacks in 1 round MAP is still increased after every action for the number of attacks. Other class are still at the 27% disadvantage when using 2 weapons instead of 2 handing a weapon. It's like offering 2 spells that are identical except one does 8d12 and one does 10d12 and then wondering why people don't use the 8d12 spell.
We all know that 2 handed weapons greatly out damage 1 handed weapons at higher levels. So I was wondering how this small change would affect the balance.
While I agree you don't get alot from the archetype feat, they seem fair to each other.
I like the idea of variable action cantrips, but there just needs to be a few and if you want it you can take it.
shield spell says "After you use Shield Block, the spell is dismissed and you can’t cast it again for 10 minutes." So it appears you can't recast it in round 2. Maybe if there was a way to get shield without casting it, like from an item.
The more I think about this the more I like both ideas. Skill feats should become better as you become better (like cats fall) and increasing an average roll for a skill one is more proficient in makes since (adding more d20s and taking the higher). I'm not sure I would go so harsh on untrained though, maybe -4 or lower of 2d20. As lower of 2d20 really affects crits and crit fails.
When you look at the numbers you are suggesting there is a huge difference in an average untrained and legendary. Because of the way roll 2d20 take lower works an average roll would be level +5 -2 or level +3, with double the normal chance of rolling a critical fail. Trained get an average roll of 10 so that works out ok. Expert is just 11 so not a big change. Master becomes an average roll of 17. And legendary becomes something close to 21 average roll. That is a very large spread. It's hard to mix +modifiers with a +mod system and keep the numbers close.
Yes and no. The GM would have to ask for the players proficiency, but only when the player was making a roll. So the simple solution would be when a player makes a roll they announce thier proficiency. The biggest problem is when the GM needs to make a secret check for a NPC or mob, but I don't see a difference between that and the secret rolls GMs already make for traps and the such.
While I like the way this sounds the porblems come out in combat. For example an master in athletics grapples an untrained in athletics and acrobatics wizard. The wizard can never escape.
I agree this sounds interesting, but you probably already know the math given won't work, especially at higher levels with a 15 point difference between trained and legendary. But the increase rate could be reduced to allow the math to work out. Using Snowblind's modified method:
Untrained: same as yours
For a level 8 that is -1/1/2/3/4 and at 20 it's -1/3/5/8/10.
I guess the reason I still like the dice roll over the "take a 10' is because, as a player, I never know if taking a 10 will pass the DC. If skill assurance let the GM auto pass you on a check if the DC was below the assurance number or told you to roll if the DC wasn't, I think I would be much more satisfied with it. But I cant stand the idea of a character needing to cross a plank and taking assurance and falling to there doom, when they at least had a 50/50 change to roll a 10+.
I like where you are going with this. A quick glance at the DC table and combining your idea and the idea of 'roll a d20 if you roll X or lower, treat roll as X', you end up something like:
Trained: x = 3
This only assures up to master at legendary (assuming you have invested in the stats/items to increase your mod) but it prevents most critical failures and will help to assure that easy DCs are passed.
Relying on the d20 for success is unsafe and unsatisfying. I prefer to let the modifiers so the work for me. Failing checks more than 5-15% of the time feels crappy.
I agree that a d20 does cause issuses with consistant successes, without high modifiers. The problem with letting the modifiers do the work is that it makes some task impossible for people that haven't specialized in it and that makes them feel crappy. The is the main reason the modifier spread can't be too large and the reason I like the idea of giving legendary a 'roll 2d20 take the higher' even if you add a +1 or +2 too the roll. Doing so greatly adds to success rate without spreading the DCs out too much.