Pathfinder 2e

[Dean]

LotR and Final Fantasy don't go together. Like, they REALLY don't. It's actually the reason for me NOT to play 4e, 5e or PF2e - they're a bit too heavy on LotR and low on Final Fantasy.

Yo not to threadjack but does anyone have the Final Fantasy writeup Frank made a long time ago? This made me want to look through that and I realized I've probably gone through 3 computer since I downloaded it god knows how long ago.

I know Lago had it cause he made a thread about it a hundred years ago, I assume Frank has it somewhere. Anyone got a copy?

[Thaluikhain]

But the 4e team was never going to do it better because they honestly weren't creative enough or mathy enough to even understand why the first draft didn't work. And they weren't interested or dedicated enough to playtest the fucking things twenty times and take honest assessment of how it worked to even realize that their first draft didn't work.

How does something like that happen (in a serious company with a product they want to get real money from)? I get maths being hard sometimes, but surely you'd want to test the thing as much as possible.

Way back in the late 20th century, GW started putting rules for what became Mordheim in WD and asking people's opinions of them long before making the finished product. Back when people had to write letters and mail them in.

[Username17]

"adequate foundation" for what though?

Collective action montages in a non-combat situation. So imagine for the moment that you're 'preparing the town for the invasion' in Red Hand of Doom. You make some speeches, you train some local troops, you build barricades, you make nice-nice between the Elves and Humans, whatever. And a simple challenge framework would be sufficient to output a defenses number that indicated how well the town does in the siege.

Or let's say you need to do a 'prepare the ship/caravan/whatever for the voyage' montage. And then your output after three rounds (or whatever) gives you an preparedness number for your journey. Or you need to 'canvass the town for clues' and your output tells you how many clues you get.

Multi-character montages are totally a thing, and 3e D&D legitimately doesn't handle them very well. Making a simple framework for group montage contribution is both valuable and extremely easy. That the 4e designers didn't even manage to create such a system that granted non-binary outputs after fifteen die rolls remains mind-boggling.

How does something like that happen (in a serious company with a product they want to get real money from)? I get maths being hard sometimes, but surely you'd want to test the thing as much as possible.

Designing new systems is hard. Designing content for existing systems is easy. If you hand someone the assignment to write a new system from scratch they might get back to you in five years and they might not. On the flip side, if you ask someone to write in the lines and make content that is within the system you've already got you can expect 20-30 thousand words of primary writing per week.

So basically what you have is a fundamental failure to understand what parts of the game are 'like writing new potions or elf subtypes' and what parts are 'like writing a new system of dicepool interactions.' And so you have people assigned to write the entire skill challenges system in a week because if you asked someone to write up rules for randomly determining town demographics or some fucking thing, you could get that much usable text in that amount of time.

4th edition was a content scaffold. The concept at every level was to create empty boxes that could be filled with new text as quickly as the writers could shovel it out. But there was a very real failure to understand which parts were load bearing and which could be filled by shovelware. Also the math was really weak and the scaffold wasn't nearly strong enough to support that amount of shit they intended to throw at it. The thing where they frickin boasted of having taken out 'Heinsoo Craziness like 6d12 damage attacks' is really front and center. These are people who straight up did not understand what mathematical progressions implied and thus balked and did ad hoc revisions of damage outputs that 'looked scary' rather than ever crunching numbers and figuring out how many rounds or die rolls they were calling for.

But the bottom line is that the Skill Challenges went to print without being playtested at all, and this happened because people in power thought they were a part of the game they could hack out in crunch time alongside flavor text about dragonboobs. And they were wrong.

-Username17

[Thaluikhain]

Ah, ok, thanks, I can kinda see how people could make that mistake.

[Lago PARANOIA]

How does something like that happen (in a serious company with a product they want to get real money from)? I get maths being hard sometimes, but surely you'd want to test the thing as much as possible.

The biggest problem is that the game designers were lazy. The failure of skill challenges is the most obvious one. They straight-up didn't test their revisions. The second-biggest problem was that they didn't understand their own game. These fuckers were proud of their Ancient Red Dragon combat, even though anyone would be absolutely embarrassed at that failure state. Even if the 4E D&D team did test their skill challenge revisions, they probably would've thought that the high failure rate was just the DM being a dick and would've stared at you like pithed frogs if you noted how additional players never help with skill challenges.

But those are kind of boring observations. The 4E D&D devs were lazy and stupid, so what? So lets talk about a common systematics problem that might apply to non-lazy/non-stupid people. That of inputs and degrees of complexity between elements. Turns out, these are MUCH harder problems that stymie systems simpler than a TTRPG.

The biggest reason why the game designers failed is because they didn't realize what to re-test after they did tweaks. Especially if you're not mathematically inclined, it's hard to see how boosting monster defenses +2 across the board affects your entire system besides the combat encounter. But something like that has huge ramifications for things like combat length, how frustrated people get by blowing their daily powers, how much people value raw pluses, soforth.

So going back to the Skill Challenges example. There was a time in 2008-2009 when people cared about min-maxing their skills. That shit was hard and the published APs tied quest progression and encounter difficulty to how well you did in them. People would legit spend their resources to get themselves an additional +6 to +8 on a couple of skills. Then they decreased the DCs across the board. Those minor mathematical tweaks completely destroyed the system. Now you're considered a goddamn idiot if you put any real resources into juicing your skills. Skill challenges before the Big Easy Errata were actually pretty tense and if you managed to win them in APs you'd see a big effect on the game. But because winning skill challenges became so easy, DMs were disinclined to tie their outcome to the rest of the adventure (for obvious reasons) and thus they became disinclined to use them at all. Because they became disinclined to use them at all, no one cared about the other failure modes of skill challenge. Changing the DCs to 'you always win' instead of 'you almost always lose' destroyed any interest not only in fixing the system, but tying it to other parts of the system.

And THAT is an outcome that could not have necessarily been caught with playtesting. It would've required thinking about the interplay between skill challenges and broader adventure design.

[rasmuswagner]

"adequate foundation" for what though.

Let's take an old D&D standby, a room in a tomb with a locked chest and things in it: someone can roll search to look for traps and hidden panels; someone can try to pick or disintegrate the lock on the chest; someone can roll knowledge on the tapestries to recognize some ancient myth or whatever; and a bunch of other random things. What would a good Skill Challenge mechanic add to your "interacting with the room" experience?

The situation lacks pressure, though.

"Get as much loot and/or info out of this room before the enemy breaks down the door/before the guards realize we're here/before our buff spells run out", that could be a challenge.

[Username17]

Skill challenges before the Big Easy Errata were actually pretty tense and if you managed to win them in APs you'd see a big effect on the game. But because winning skill challenges became so easy, DMs were disinclined to tie their outcome to the rest of the adventure (for obvious reasons) and thus they became disinclined to use them at all. Because they became disinclined to use them at all, no one cared about the other failure modes of skill challenge. Changing the DCs to 'you always win' instead of 'you almost always lose' destroyed any interest not only in fixing the system, but tying it to other parts of the system.

Makes a good story. Unfortunately, I don't think you can pin the public's ultimate disillusionment with Skill Challenges to a single event. Not because there isn't an event that would logically make people turn against Skill Challenges, but because there are too many that came too close together.

Mearls himself wrote more than twenty skill challenge revisions. The DMG2 had over two dozen alternate rules written by at least nine different authors. What event was the straw that broke the camel's back? What was the straw that broke the median camel's back? For fuck's sake, the DMG2 was year two of that fucking train wreck!

-Username17

[Pedantic]

Mearls himself wrote more than twenty skill challenge revisions. The DMG2 had over two dozen alternate rules written by at least nine different authors. What event was the straw that broke the camel's back? What was the straw that broke the median camel's back? For fuck's sake, the DMG2 was year two of that fucking train wreck!

-Username17

This situation has made it all but impossible to discuss skill challenges with their apologists. Everyone claiming to really like skill challenges has completely mind-caulked together their own system, and views the words "skill challenge" as a generic catch all for "group of skill checks made by the party collectively." No one actually has a system for them, just a series of house rules, and whatever random set of actual rules they've decided to throw into the pot.

I'm with Foxwarrior, in that I have yet to see a utility case for skill challenges that's better than a standard skill system, though that is rather complicated by the fact I can't seem to get a hard enough definition of what a skill challenge entails mechanically to actually compare the two.

The elements that people seem to like are "everyone is involved" and "each roll creates incremental progress." Fundamentally, I think these aren't problems that require a new and unique subsystem, so much as more careful class design (to give characters enough tools to contribute to most challenges) and more thoughtful challenge/scenario design (to ensure some of your problems are complicated enough not to be subject to 1 roll resolution).

Montages are an interesting argument. I can see a case for some kind of downtime, or "off-screen" rules, that allow for more abstracted resolution or use of less detailed, longer-term abilities when there isn't moment-moment focus on the character's actions, but I'm not really sure skill challenges in any conception I've seen, are the best way to do that.

[Lago PARANOIA]

Fundamentally, I think these aren't problems that require a new and unique subsystem, so much as more careful class design (to give characters enough tools to contribute to most challenges) and more thoughtful challenge/scenario design (to ensure some of your problems are complicated enough not to be subject to 1 roll resolution).

I'm just going to quote FrankTrollman in this thread.

A skill challenge could easily do that. Having nonbinary outputs for the skill challenge is an especially important design goal, because having a nonbinary outcomes on dependent on the outcome of a particular d20 roll (or several particular rolls) is fucking nuts mathematically. But even if you do nail that, a frequent challenge I see in D&D adventure design is how non-binary adventures really aren't. They're set up so that you have to succeed almost all of the substeps.

For example, say you want to do a heist of the castle vaults without getting detected. There are plenty of failure points (not knowing the vault's location, not knowing the password, being seen by the guard patrols, activating the magic alarms, not having a way to get the treasure out, etc.) and while in base d20 it's possible to make that outcome nonbinary, you and I know that most DMs will run a failure at any of those points a hard failure of the skill challenge.

Skill challenges, if properly designed, could force the adventure forward by refusing to let the players have a hard fail/pass state. I know 'fail forward' is a dirty phrase on this board after Bear World, but it's a defensible design goal.

[Foxwarrior]

If you represent the heist as a skill challenge/montage where some of the actions fail (various players contribute skill checks etc), and they successfully make the "knowing the password", "hiding from the guards", "avoiding the alarms", and "getting the treasure out" checks, but fail the "knowing the vault's location" check and don't do anything else to determine the vault's location, is the nonbinary outcome that they get 80% of the treasure? In-fiction, how do they steal from the vault without finding the vault?

I can see how it would be okay for the town-training montage, because there are a bunch of independently helpful things you can do, and maybe your game doesn't have an elaborate mass battles simulator so you can't represent the difference between having walls but no spears versus having spears but no walls, so you ultimately just want to boil things down to a single number. But if you did have an elaborate mass battles simulator, you could just have a time limit and skills and abilities that take time to use, and then you do a battle with whatever defenses and training your characters actually specifically provided.

So basically, I see it as working when what you want is the aesthetic of solving problems by using your character abilities, but don't actually care about the details so much as you care about the visuals of your characters contributing in various thematic ways.

[Lago PARANOIA]

If you represent the heist as a skill challenge/montage where some of the actions fail (various players contribute skill checks etc), and they successfully make the "knowing the password", "hiding from the guards", "avoiding the alarms", and "getting the treasure out" checks, but fail the "knowing the vault's location" check and don't do anything else to determine the vault's location, is the nonbinary outcome that they get 80% of the treasure?

C'mon, are you even trying? I can think of five ways to move that adventure forward from your setback without doing a boring percentile reward reduction or just doing a hard stop.

[*] The evil corrupt court jester leaks the location of the vault to one of your contacts in a way that lets the party know they're now being blackmailed.
[*] Your clumsy investigation caused the 'suspicion' meter to go up and now you have to deal with more patrols surrounding the vault.
[*] From one of your contacts, the captain of the guard got wind of your heist but has a plan to settle your hash; he threw in a few cursed coins with serial numbers he knows about. The contacts can tell you the serial numbers for some additional money.
[*] One of your NPC contacts got arrested while researching the location for you and now you have to figure out out to deal with it.
[*] You missed the window of opportunity before the seasonal blizzards start and now you have to deal with the complication of getting the loot carts out in driving snow.

[Foxwarrior]

I guess the "suspicion" meter one fits with the skill challenge design, but the other ones would work fine in a non-skill challenge system that just happened to have suggestions for various events that could happen to move things along when the players failed.

You might be able to put most of those onto your "failed roll" section of the Gather Information skill in 3.5e, with only some tweaking to make them seem a bit more generically applicable.

[Lago PARANOIA]

I guess the "suspicion" meter one fits with the skill challenge design, but the other ones would work fine in a non-skill challenge system that just happened to have suggestions for various events that could happen to move things along when the players failed.

That's the thing -- they DON'T move along. When you present challenges as a series of binary and sequential gates, people treat the gates or complications that arise from them as obstacles to be taken care of sequentially. The complication becomes blocking or a narrative tumor rather than something you have to deal with post-hoc. Because that's what happens when you present challenges and the resolution thereof sequentially.

What's more, having non-binary outputs for single d20 rolls is a fucking nightmare, especially if you want to balance the thing. It's the nature of the sequential graph it creates. Instead of having a state graph that just has a
single node connecting to N outcomes each with some arbitrary probability of transitioning to a state, you instead have a Markov chain where you have literally quadratically more nodes (states) than a skill challenge.

That's not just a nightmare to mathematically balance, it's a nightmare to narratively plot. So instead of having non-binary outcomes based off of total successes, you have non-binary outcomes connecting to non-binary outcomes connecting to more non-binary outcomes etc. until you decide that the players have done enough legwork for training the peaceful villagers/running a heist/helping in the local election they can have the final outcome.

Have fun planning that rigmarole with your method.

[Foxwarrior]

Ideally I think the designer and the DM shouldn't be planning any rigamaroles, only putting in things that could become rigamaroles when the players touch them. Antics and tactics dude, not adherence to scripts. And if things can't balloon into a bizarre mess every once in a while, I'm not sure your game is going to have many amusing antics in it.

But how do "Your clumsy investigation caused the 'suspicion' meter to go up and now you have to deal with more patrols surrounding the vault." or "You missed the window of opportunity before the seasonal blizzards start and now you have to deal with the complication of getting the loot carts out in driving snow." work in a non-sequential system? Surely those things make later checks about sneaking past the guards and carting out the treasure harder, and thus remain causal?

[Kaelik]

I'm still confused about how not knowing the vault is not a binary pass fail because if you don't know it then your failed investigation increases security around the vault you still don't know the location of so you still can't take any treasure from.

Still seems like "finding the vault" is pretty binary.

[Omegonthesane]

I'm still confused about how not knowing the vault is not a binary pass fail because if you don't know it then your failed investigation increases security around the vault you still don't know the location of so you still can't take any treasure from.

Still seems like "finding the vault" is pretty binary.

Only if you insist on interpreting "You fail your Know Where The Vault Is roll" as "You don't know where the vault is and can't find out where the vault is, also you can't take more time with all the added consequences of that to reroll, gg qq".

If hypothetically you were making a driving roll to drive to McDonalds, you would not interpret a failure of that driving roll as "you fail to drive to McDonalds and can't try again". You'd interpret it as "you are stuck in traffic and late to the McDonalds" or "you are caught on a speeding camera" or maybe on a critfail if it was important "your car doesn't start, your challenge is now to jumpstart the battery".

[MGuy]

I don't think having time or the ability to try again at finding the vault makes the failure to not find the vault not a binary stopper on the adventure. Your adventure is still stopped up until vault is found. You can fall to know the security details and still get to the vault. You can fail to know the password and have to find a workaround when you get to the vault. You cannot do anything dealing with the vault until you know where it is.

[Chamomile]

"adequate foundation" for what though.

The problem with this question is that any specific answer is going to be better served by a single-purpose sub-system dedicated to that specific thing. When skill challenges are brought up, the baking contest is usually brought up as the situation you'd use them in, and the point there is not that skill challenges are particularly good at modeling baking, it's that this is such a bizarre situation that you wouldn't ever write specific rules for it. In order to justify their existence, skill challenge rules need to be better than MTP at a very broad variety of situations.

[Username17]

"adequate foundation" for what though.

The problem with this question is that any specific answer is going to be better served by a single-purpose sub-system dedicated to that specific thing. When skill challenges are brought up, the baking contest is usually brought up as the situation you'd use them in, and the point there is not that skill challenges are particularly good at modeling baking, it's that this is such a bizarre situation that you wouldn't ever write specific rules for it. In order to justify their existence, skill challenge rules need to be better than MTP at a very broad variety of situations.

That's the core of it. Obviously if you decide that your game is about ocean voyages, you should probably have a core ocean voyage mini-game. If you decide that your game is about gathering clues in libraries, it should probably have a robust library research mini-game. If you decide that your game is about preparing fancy feasts, you should probably have a detailed cooking mini-game. But probably your game isn't about any of those things, and yet any of those things are things that could happen.

What your game does definitely need is a system for handling teamwork. There are between four and six player characters and often less than that number of meaningfully operational goals at any given time. And so it's not really surprising that you will very frequently be in a situation where one player wants to contribute towards solving a problem, and at least one other player wants to contribute to solving the same problem. And if your characters are as... distinct as RPG characters frequently are, their preferred (or available) means of contribution won't be identical or even necessarily obviously compatible.

So what you want is a collaborative montage system where different player characters do different things and this nebulously adds to the overall success of the mission until the montage is complete and you total up the score. But this also necessarily means that you need to have this system assign contribution numbers to things like 'I'll cast my divination spell' or 'I'll go smash things with my giant hammer' that aren't otherwise defined with the same kinds of outputs as 'I'll roll my Diplomacy check.'

4e's skill challenge system was an unmitigated disaster. Every part of it was a failure. It achieved zero design goals. But the determination that such a system would be a good thing to have in a cooperative storytelling game is dead on. Yes, your cooperative storytelling game would be better if it had a functional default mechanic for cooperative approaches to problem solving in the story. I mean, frickin obviously.

-Username17.

[Kaelik]

I'm still confused about how not knowing the vault is not a binary pass fail because if you don't know it then your failed investigation increases security around the vault you still don't know the location of so you still can't take any treasure from.

Still seems like "finding the vault" is pretty binary.

Only if you insist on interpreting "You fail your Know Where The Vault Is roll" as "You don't know where the vault is and can't find out where the vault is, also you can't take more time with all the added consequences of that to reroll, gg qq".

If hypothetically you were making a driving roll to drive to McDonalds, you would not interpret a failure of that driving roll as "you fail to drive to McDonalds and can't try again". You'd interpret it as "you are stuck in traffic and late to the McDonalds" or "you are caught on a speeding camera" or maybe on a critfail if it was important "your car doesn't start, your challenge is now to jumpstart the battery".

It sounds like in every single one of those cases finding the vault is a binary thing, and you have to find the vault before you can continue on with the mission.

Saving throws in D&D are still binary pass fail even if the enemy can cast Stun Ray again next round and even if someone can dispel stun ray. Still seems like the save on Stun Ray is a binary pass fail.

If you fail to find the vault that can cause problems during the mission if you got the wrong location, but it doesn't mean you steal half the money in the vault, it means you try to find the vault again until you succeed.

[deaddmwalking]

Knowing that a task gets progressively more difficult as you fail to make progress creates tension.

Even in this case it would still be possible to find success without EVER finding the vault. It might be possible to create a situation where the treasure must be moved; now your vault heist is a train robbery.

[Lago PARANOIA]

It sounds like in every single one of those cases finding the vault is a binary thing, and you have to find the vault before you can continue on with the mission.

Yes, narratively, finding the vault is binary. You either find it or you don't.

In terms of game mechanics, though, why should the die roll (or whatever) to find the vault be a binary, self-contained output? I know D&D does it like that, but there are plenty of systems that output results of 'yes and', 'yes but', 'no but', and 'no and'.

Tying this back into our skill challenge example, say the SC is 'raid the vault and get out with the goods scot-free'. Your 5-person party has two rounds to come up with ways to lift the treasure and you get one additional roll you can distribute as you see fit. However, you also mandate that at least one of the actions taken has to explicitly for completing the following tasks: finding the vault, finding the password for the vault, evading the patrols, disabling the alarms, and getting away with the loot. Your doesn't have to succeed at a particular task but they do have to do something about it, otherwise you automatically fail at the skill challenge.

That said, as long as every mandatory task is attempted, you can still do things that will add to the overall success of the mission. For example, the party fighter convinces the shogun that the vault (rather than the shogun's personal bodyguard) is the best place to store the military relics while they're not using them. If you want to add additional complications, you can also put a limit on successes for a particular task. For example, completing the 'find the vault' task will only contribute one success. But you can get up to three successes by doing the 'disable the alarms/trap' task repeatedly.

The results of the skill challenge rest on how many successes you have.

[*] 5-6 successes, succeeded at every mandatory task: the heist goes off without a hitch.
[*] 7+ successes, succeeded at every mandatory task: you pull off the heist without a hitch and in fact you get some additional rolls on the treasure tables. Or a player's rival in the courts gets fired for incompetence. Or you impress the underworld and you get bonuses on checks to interact with them.
[*] 8-10 successes, failed some mandatory tasks: you pull off the heist without a hitch. Even if there's a task you failed to accomplish entirely, something happened to make it moot. For example, you weren't able to find the vault on your own but while researching the schematics the thief lifted the Psion uses their Psychometry power to determine the map was created in the mountains.
[*] 5-7 successes, failed some mandatory tasks: This is where we get into 'yes, but' territory. You win the skill challenge, but this is when some penalties start accruing. The court jester blackmails you. You too too long and now there's a surprise blizzard while you're transporting the loot back. There's not as much treasure as you think. The carts break while you're heading back into town. Ideally, the 'but' part relates to the parts of the challenge that you failed at.
[*] <5 successes: You fail the skill challenge. You get consequences depending on what mandatory tasks your party failed at. You can decide whether or not the party can recover outside this skill challenge or that's that. For example, if the party got successes on every mandatory task but disabling the alarms, you can run a Very Hard fight scene where the fiends get summoned and they're notified that the kingdom is after them. You can also throw in 'yes but/no but' as you see fit if the failure of a mandatory task would in a narrative sense logically block the plot progression of the skill challenge. So not only did you fail the skill challenge, but your lack of success on the mandatory task of 'find the vault' means that you got a 'yes but' from a rival guild who also got the information from your contact and they're going to pull off a successful raid with YOUR LOOT.
[*] Any number of successes, did not attempt or do some mandatory tasks: The skill challenge automatically fails. Depending on which mandatory task you didn't undertake, your adventure gets stuck there with the DM to decide.

[Yesterday's Hero]

I do something similar, with a twist inspired by MTG, for my skill challenges. Successes net you "Charge Counters" and you need a certain amount to complete the challenge (getting more gives you extra rewards), while failure (sometimes failure by 5 or more, depending on the amount of danger) give you Doom counters to determine bad stuff that happens along the way.

I list the checks and the DCs and PCs are encouraged to come up with alternative methods that grant big bonuses or auto success. "You have to climb this rather smooth wall with a DC of 25." "Can I use levitate and then lower a rope?" "Yes, you get an auto success and that gives you a Charge counter".

[MGuy]

So... Why not just not start the heist challenge until you find the vault? I'm really having a hard time understanding this desire to claim that finding the vault somehow isn't a binary thing by pointing to ways to find the vault anyway. The bottom line is the that you have to find the vault. No vault location no vault adventure.

This would probably even apply to anything you would tag as 'mandatory'. If a thing you need to get done is pretty much the only way forward at all, and I mean something as integral as knowing where the adventure even is, then I think you pretty much either have to give the information out for free (at which point why fuck around with rolling?) or you make getting that necessary thing the skill challenge/an adventure all on its own.

[Iduno]

If you represent the heist as a skill challenge/montage where some of the actions fail (various players contribute skill checks etc), and they successfully make the "knowing the password", "hiding from the guards", "avoiding the alarms", and "getting the treasure out" checks, but fail the "knowing the vault's location" check and don't do anything else to determine the vault's location, is the nonbinary outcome that they get 80% of the treasure? In-fiction, how do they steal from the vault without finding the vault?

Failing one skill challenge means you need to spend more time to work around them, which makes other things harder or more critical. You know generally that there is a vault in the building. If you've done well with everything else, nobody knows you're there and you should have some time to just go search for the vault. If you set off some alarms, had to run from the guards, the police are on their way, and your getaway vehicle with all of your tools was stolen, you're not going to have time to go looking around for anything other than a way out. Somewhere in between, you maybe need someone who knows architecture/structural engineering to narrow down your search, or pull some luck out of your ass.