Pathfinder 2E PF2: Second Attempt Post Mortem

[GrahamWills]

I'm primarily talking about complex traps, which are generally the ones encounters are built around. Looking at the guidelines for what a trap for a level 9 standard encounter should be (level 11 complex trap), we see that it will have a stealth and disable DC of 33 (or possibly 36 – a hazard is supposed to have at least one stat at "extreme", usually either Stealth, Disable, or save DC/attack bonus, but let's ignore that for now). Let us for the moment ignore the skill gating of requiring Master proficiency to deal with a level 11 trap, because proficiency gating is naughty word of the highest order.

So, Step 1: Trap Detection
Assume we don't expect a trap and are in explore mode. Half the party are probably in search mode (actually in my group it's all of them except one, but let's be more general). So the two high perception characters are searching.

I'm happy with your suggestion of a +17 for a good but not focused character. +23 is pretty high even for a super-focused character. I'll take a +20 for a 'focused' character -- because that's the stat for my level 9 investigator who hasn't put feats into it, but is naturally good by class.

So, doing the math, between the two of them they have better than a 50% chance of noticing the trap without triggering it. And since you trigger the trap only if you crit fail, then they will nearly always disable it at no cost. So over half the time a full encounter will consume zero resources and provide full XP.

Step 2: Being Damaged by the Hazard
So, now for the case when they are unlucky and trigger the hazard.
At level 11, the complex hazard attacks at +24 doing 28 points damage. Let's assume it can attack 3 people every turn at full attack (or cause them to make saves). A rough look and AC and saves says they'll hit 75% of the time, so it'll do about 21 points a round, assuming no-one raises shield, has DR or anything useful (and at level 9 ... they will!) So it'll take 4 rounds before anyone needs to start any healing.

Step 3a: Disabling the Hazard
"A good complex hazard often requires disabling multiple components or otherwise interacting with the encounter in some way" so although Thievery is the prime skill, we can assume another skill is also useful. Let's say 2 thievery and one other check.

A character good at, but not specialized in Thievery might have +18 or so. It's hard to think of a reasonable party where another party member doesn't give them a +1 at least to a skill check, and hitting a +20 to assist is not hard, so with help from 2 other people (say), +20 on their roll is what I'd expect from a party that has paid minimal attention to dealing with traps. To make those 2 thievery checks will take 4-5 rounds.

That's the worst case -- a part with only minimal attention to traps against a trap requiring thievery. For them they will take maybe 5 rounds of activity to to get through the thievery checks and will take around 300 hits of damage.No-one will be in danger, but that will require a fair amount of resting afterwards. If you have an hour you can wait, no issues, but if you need to move on fast that will be very nasty

Step 3b: Beating the Hazard to a Pulp
Hazard has AC 31, hardness 20 and 80 hits. It becomes broken at half hits, so we only need to do 40 points damage to it.
I'll assume a trap can't be dazed, stunned, flanked or anything else -- this is actually uncommon, but if it can be affected by player powers, it's going to drop very fast as it's one one creature facing 8-12 stunning / dazing / etc. attacks every turn.

So my level 9 fighter has +22 attack doing 2d12+10 on their first attack, and I'd probably do a defensive move on my second, so just +17 doing 2d8+8 (yeah, not very optimal ...) -- let's ignore the second attack. So his attack has a 10% chance of a critical for 46-20 = 26 damage and 40% chance of averaging 3.8 damage so averaging about 4 damage a round. If everyone has does about the same, that means 16 damage a round and it takes 3 rounds to break the trap.

Again, note that this is the zero-resources strategy with a party not combat optimized. Best case is probably having a double-strike pick fighter (like my son has at 9). Tricky to calculate the average math, but about 25% of the time he's hitting twice regularly for 30-20 = 10 damage, =10% of the time one attack crits and one misses for 46-20 = 26 damage. 10% of the time one attack hits and one cries and he one-shots the trap. About 10 damage a round.

So assuming you have one actually competent fighter (or monk -- basically anyone who can combine attacks to defeat hardness) your party will do effectively about 10 + 4 + 4 +4 damage a round and defeat it in 2 rounds.

Again, this assumes you don't want to expend resources and that the trap is immune to pretty much everything. So in this case you'll take 2 rounds of damage and so about 120 points of damage across the party, which is easily handleable.

Summary

• About half the time you'll detect the trap and then can leisurely defeat it. If you have a hero point to negate a critical fail, it's 99% safe, otherwise there's about a 10% chance you may enter combat
• If you only have one thievery person, and they are only "ok", not great, and the trap requires multiple Thievery checks, then you are best just beating it to a pulp, otherwise disable
• You'll never be in any danger, but it will deal about 120 points of healing on average

Is this actually true?
The above is all theoretical, but I've played quite a lot of PF2 and it matches my experience. A trap or hazard that was solo and not part of a scene was only once an issue (Age of Ashes, one of the towers in the forest) and I'm pretty sure that was not built with regular rules. We generally run a high-perception party so maybe ⅓ of the time we triggered a trap. Of those some of them we simply withdrew and slowly killed from range (this works surprisingly often -- traps are programmed so really easy to exploit) and others we usually disabled (we've always had 2+ good thievery people between rogues, monks, investigators, swashbucklers and dex fighters). We did beat a couple up, which worked fine, but was pretty boring. Once we started a second encounter on purpose and dragged those people into the trap.

Find the fun!
Solo traps are only fun if they have lasting effects that go beyond the encounter. If they just do damage, they can't do enough to threaten a party who can rest 20 minutes afterwards, and even for that situation, it just consumes a few healing resources.

However, as party of an encounter, they are way, way more fun (as @JmanTheDM says -- I totally agree) Playing an investigator, a common turn was "make a plan -- ah, a 4. I guess attacking is not great, let me disable the arrow trap!" and in general trap increase options and decisions, making combat more fun.

Even at high levels, they are still fun. Last night we decided we hated the range 1000' True Sight + See Invisible Archery hazards so much we used a wish to teleport our 16 adamantine golems off to attack them so we didn't have to deal with them. Our first attempt to sneak up and disable them went poorly and we didn't want to fight the ghost army, the flesh-warped artillery and the death knight regiment while taking 50 hit point damage arrows each round ...

 

[CapnZapp]

No @GrahamWills that's not how math works - you can't use averages to create your worst case scenario!

First off, you simply accept that two heroes must be actively working together to even reach above a 50% detection rate. What you don't appear to be cognizant of is how problematic and gamey the circumstances leading to this outcome are in every general role-playing situation outside of Paizo's very narrow-minded point of view. Basically, in Paizo's world, it's completely acceptable to call somebody an expert when that only means they can pair up with another "expert" and together achieve perhaps a 60% success rate. In every other perspective, a 40% failure rate is abysmally low, and something you wouldn't expect even of the clumsiest neophyte... that you still trust to take the assignment!

Second, the fact that you might detect a trap half the time does not excuse or justify your experiences the other half. Not to speak of how detecting it does not mean you're automatically in the clear.

Third, where your math really leads you astray. Just because three people on average only needs healing after four rounds does not mean this outcome is especially likely. And even if it is, just because some mild scenario happens to beat out other harsher scenarios by a few percentage points, and thus is the most likely one, does not mean you should then focus on that scenario over others.

In all balancing, you need to calibrate for the worst cases. So why aren't you discussing worse cases?

To be very clear, even if 9 traps out of ten are "fine", if the tenth one kills a character, then no, traps aren't "fine". (And before you focus on that scenario, let's not forget how this just obscures the real charge against the PF2 trap implementation: "much ado about nothing")

I will just skip the section where you discuss how you beat a trap to pulp, since we all know that if it bleeds you can kill it, and that by far the most incongruous hazards are the ones you can't damage (basically since the game never bothers to tell you how you explain the difference to your players).

 

[GrahamWills]

No @GrahamWills that's not how math works - you can't use averages to create your worst case scenario!

I’m pretty confident that I know how math works, actually. Let me help you out here. The two factors that I am considering affecting the outcome of an encounter with a trap are (a) how effective the party is at dealing with traps, and (b) how lucky they are at rolling. The other obvious factor is the effectiveness of the trap at dealing with the party, but we have established that we want to deal with a standard trap as defined by the book.

Now, trying to analyze over two combinations of effectiveness x luck is tricky. It’s actually the sort of thing I do every day because I’m a professional research statistician, but it’s tedious and would take forever to write out, so I elected a simpler plan — use a slightly less competent than average party for (a) and average out (b). My comment on “worst case” was within a subsection of my argument; if that confused you, let me state it more clearly — where not otherwise stated, I am modeling an average or slightly less good than average party dealing with a standard trap , estimating average effect.

First off, you simply accept that two heroes must be actively working together to even reach above a 50% detection rate.

I’m pretty sure that it’s not contentious to assume that heroes work together. Especially when dealing with a situation that is defined to be scaled for 4 heroes working together. As several others have pointed out, a single hero working on a trap scaled for a singled hero is a completely different aanalysis.

Although if it is standard for players in your games not to work together, then maybe that might explain why you find a lot of PF2 hard? Or maybe all your groups are full of fighters warily raising their shield and falling into pit traps? I can see the latter especially if you don’t have anyone with any form of perception In the party.

But again, I’m dealing with a normal or slightly sub-par party, assuming that players cooperate with each other.

What you don't appear to be cognizant of is how problematic and gamey the circumstances leading to this outcome are in every general role-playing situation outside of Paizo's very narrow-minded point of view. Basically, in Paizo's world, it's completely acceptable to call somebody an expert when that only means they can pair up with another "expert" and together achieve perhaps a 60% success rate. In every other perspective, a 40% failure rate is abysmally low, and something you wouldn't expect even of the clumsiest neophyte... that you still trust to take the assignment!

The above appears just to be a general ”I hate Paizo” rant. It’s also inaccurate, as 40% failure for a single action is quite common in the game; in fact in most game systems 50% failure is probably the norm for a single action. It’s no different from other actions, like convincing the guard to let you go through a gate, or sneaking past a sentry, or whatever,.

I am curious though: what success rate do you expect to get for a level 9 character who’s is trained but has no other relevant feats when attempting a skill check against level 11 opposition?

Second, the fact that you might detect a trap half the time does not excuse or justify your experiences the other half. Not to speak of how detecting it does not mean you're automatically in the clear.

I’m not making any form of moral argument here! I’m just responding to the “traps are deadly” comment. So I think it’s completely relevant to say “half the time they are trivial”, because, you know ’trivial’ != ‘deadly’

If you are really interested in understanding this situation, I have a suggestion — why don’t you do the analysis for a trained, high dex level 9 character getting a +1 or +2 bonus disabling a level 11 trap that they have discovered and not triggered. It’s a pretty easy analysis — all you have to do is, for the N successes needed, work out the chances they will roll a crit fail before the N successes. For a real-world analysis, assume they have a hero point that they’d really like not to spend, but will if necessary.

Third, where your math really leads you astray. Just because three people on average only needs healing after four rounds does not mean this outcome is especially likely.

Ummm .. what? Honestly, I find it hard to believe that you don’t know how averages work. I don’t really want to have to define them here, but since we are dealing with many rolls, and those rolls are unimodal and symmetric, the law of large numbers applies very well, and the average will also so be very close to the mode and the median, so under pretty much any English way of stating it, yes, if the average is 4 rounds, it is definitely ”the most likely“ outcome under the three most usual ways of defining most likely: the long run frequency, the most likely point outcome, and the point at which half the results are lower and half are above.

And even if it is, just because some mild scenario happens to beat out other harsher scenarios by a few percentage points, and thus is the most likely one, does not mean you should then focus on that scenario over others.

In all balancing, you need to calibrate for the worst cases. So why aren't you discussing worse cases?

Worst case is trivial; everyone rolls 1s all the time for everything and they all die in about 15 rounds after expending all healing.

That’s one reason. The other is that the premise of this argument is not that “in the worst possible case, traps are deadly”. It is that traps are very often deadly. So the main reason is thag I want to address the situation under discussio, not make up a different one. I am happy to say that the worst case for pretty much any encounter in any system is ”everyone dies”.
Less facetiously, I see that in your next argument you consider using not the worst case, but the 90% worst case — this is a much better suggestion, but again, it’s much more tricky to math out, so when I address it below, I’ll use common sense rather than statistics.

To be very clear, even if 9 traps out of ten are "fine", if the tenth one kills a character, then no, traps aren't "fine".

Honestly, it depends on your group. I’m personally not a fan of “every 10th serious encounter kills someone” levels of difficulty, but some groups like that level of difficultly. But let’s assume you want a game where no-one ever dies to traps.
Our trap does 4d8+10 damage. The 90% most extreme result is something like a crit with average rolls, so about 56 points damage. That just won’t kill anyone. Like, ever.

It’s honestly really hard for me to envisage a scenario where a damage-dealing trap is going to kill anyone — certainly not as often as a combat encounter will. The damage numbers are so low for complex traps that unless your party has made it to 9th level with no in-combat healing and has had all their magical healing stolen (although that would make an interesting hazard … my players might just kill me though) I just can’t see it.

Help me out here, @CapnZapp — point me to the level 11 damage-dealing trap (or build one) that you think will kill a PC. I’ve never seen one, never had it happen, and this thread isn’t exactly full of people reminiscing over when the trap TPKed their group.

You have the really easy task here — find a single counter example. so I think, that to progress this examination, we need you to stop just making general “I don’t like Paizo” and “you don’t know how math works” statements and come up with a single, concrete example and show how it can kill a standard party.

I will just skip the section where you discuss how you beat a trap to pulp,

Actually, don’t. we are discussing non-haunt traps here and so beating them to a pulp is a completely reasonable thing to do. If you are going to ignore the fact that half the time a trap is trivial, and ignore the fact that you can defeat them with physical damage, then please do me the favor of ignoring my whole post and moving on.

 

[Teemu]

I don’t think the PF2 rules are particularly well presented but I don’t quite understand why haunts are so gamey when the GM can dole out information with Recall Knowledge checks like with any other situation where the characters don’t know what they’re dealing with.

 

[Uni-the-Unicorn!]

I don’t think the PF2 rules are particularly well presented but I don’t quite understand why haunts are so gamey when the GM can dole out information with Recall Knowledge checks like with any other situation where the characters don’t know what they’re dealing with.

I can see it both ways. It, the many things in RPGs, comes down to the GM and the players.

 

[Retreater]

I don’t think the PF2 rules are particularly well presented but I don’t quite understand why haunts are so gamey when the GM can dole out information with Recall Knowledge checks like with any other situation where the characters don’t know what they’re dealing with.

Recall Knowledge, RAW, is actually not very effective. You're looking at a pretty high DC to get any useful information - and even then it's like one piece. A critical success gives you information of like an extra attack and weakness.
There's not enough there to be able to win a mini-game encounter that essentially does not fit in with the fiction of the world.

 

[payn]

I dont understand why haunts are anymore gamey than anything else in PF? Swarms are also a different type of encounter where typical methods are not always effective. The only difference is with Haunts there is a complex method of interacting with them if you want to permanently disable them. Otherwise, you can temporarily disable them and move on like any other trap or monster encounter.

 

[Teemu]

Recall Knowledge, RAW, is actually not very effective. You're looking at a pretty high DC to get any useful information - and even then it's like one piece. A critical success gives you information of like an extra attack and weakness.
There's not enough there to be able to win a mini-game encounter that essentially does not fit in with the fiction of the world.

So one piece of info is enough, isn't it? How to calm down the spiritual haunt?

 

[Retreater]

So one piece of info is enough, isn't it? How to calm down the spiritual haunt?

If it's the one weakness that someone in the party can access. But if it's Occultism checks - and only one person in the party is trained in that skill - the rest of the party is supposed to do nothing?

 

[Teemu]

If it's the one weakness that someone in the party can access. But if it's Occultism checks - and only one person in the party is trained in that skill - the rest of the party is supposed to do nothing?

I think it’d be Religion or Occultism? Even if only one PC is good at those, how’s that any different from figuring out a golem or some other specialty monster? Or disabling a mechanical trap?