| 1 person marked this as a favorite. |
This is less of a specific thread and more a general rambling about encounter design, and came mainly from a couple of other threads, mainly this Reddit thread about incorporating monster templates from D&D 4e. One of the general assumptions held about Pathfinder 2e is that its math, by and large, just works, and that this applies particularly well to encounter design: rather than apply a mixture of convoluted rules and individual adjustments to create encounters of a certain challenge levels, it's easy to craft a balanced encounter using monster levels and a couple of simple tables.
By and large, this system works, and thanks to PF2e's level-based scaling, this has the added benefit of making solo boss monsters live up to their stated difficulty level, making up for differences in action economy with greater defenses and more powerful attacks. However, there are also several other general understandings players end up having of those rules with play experience:
As a result, I think the statement that Paizo's encounter-building "just works" is largely true, but doesn't tell the whole story: while it is true that an encounter against a lone PL+3 boss will be balanced, and that the rules also allow building a PL+3 encounter with a dozen PL-4 enemies, in practice these rules still don't make every encounter fun or practical to run right out of the box. This is why Paizo had to change from their early AP's encounter design and implement fewer solo boss encounters, and why many players often suggest a PL+1 or PL+2 boss with a few minions or hazards for an ideal PL+3 encounter.
Thus, with the encounter-building rules we have, the most common type of challenging encounter tends to be fewer but more powerful monsters, which I think has a few knock-on effects: because crit-fishing depends in part on monster defenses, routinely dealing with monsters above party level can often reduce chances of critical hits, or critical failures on spells, down to 5%. If your AP is mostly made up of solo boss encounters, your Gunslinger or AoE caster are going to have a bad time, as is any character who wants to make use of powerful incapacitation effects, and so it is no surprise that criticism of those classes' performance happens often, and particularly in relation to older APs that are still among the most frequently-played.
---
The heart of this problem, in my opinion, is that current encounter building rules are creature-focused: to make up an encounter, you generally add creatures of certain levels until your XP budget is met. While there are rules for supplementing encounters with hazards, those are secondary, as are the details of the environment in which your encounter happens. This is definitely intuitive in a way, as most players tend to think of encounters, especially challenging encounters, in terms of big powerful monsters, but in practice it doesn't factor in a bunch of other considerations that affect the fun factor of an encounter, like player agency, speed of play, and additional choices and complicating factors via environmental design. In the worst of cases, your encounter turns into a white room with a random monster in it, and while PF2e's monsters are generally interesting enough to offer unique gameplay, that's not always enough to guarantee a fun and meaningful experience.
In my opinion, an alternative way to approach things, and the point to this ramble, is to approach the encounter itself as if it were a creature. Rather than think in terms of individual monsters with their own actions and HP, think of the encounter as its own entity, with a larger number of actions and HP distributed among component creatures, hazards, you name it. This I think could have several interesting implications:
While perhaps slightly less intuitive, I feel the above methodology could carry several benefits:
---
The big question to all of this is how to apply this methodology to PF2e, and I don't feel I have a full answer to that. The Reddit thread I linked mentions using templates from D&D 4e, which I think can work, and I do think there is room for applying traits to monsters that either strengthen them by increasing their HP, level, and number of turns per round, with extra rules for handling persistent conditions, or weaken them by reducing their HP and actions. A possible extension to this, which I haven't tested, would be to also implement proficiency without level to open up a greater pool of monsters to include in encounters in this way, as monster difficulty wouldn't depend as much on level-based increases to defenses and success chances. All of this, however, is still a bit fuzzy, and my gut feeling is that no matter the implementation, it's always going to feel like a bit of a kludge, as Pathfinder isn't inherently built with this methodology in mind. I'm sure others have thoughts on this too, including different takes on encounter design, and I'd be keen to hear what everyone else has in mind on the matter.
|
| 2 people marked this as a favorite. |
The encounter design tables work decently, but they try to capture a very complex mathematical phenomenon inside a simple model (those tables) that people can use even if math wasn't their favorite subject. But that does mean that around the edges, it's not gonna work so well. In particular I've noticed:
* Severe/Extreme encounters with single enemies tend to be Not Fun. Their really high defenses means most of peoples' "that was my big spell for today" will just bounce off. Even attempts to debuff them so someone else can try to land something big, will fail. Eventually the party might get through just by sheer amount of trying. But failing so many times doesn't feel like a heroic victory.
* Moderate and easier encounters against three or more enemies on the other hand, dilute the enemy oomph so much that they feel kinda lame.
* Simple hazards have to hit and damage that's higher than extreme for creatures, because they deliver it all in one activation. But that does mean that if you use a simple hazard that's higher than the party level, there's a real chance it'll suddenly out of nowhere kill a PC through massive damage. Which doesn't feel clever at all, just unlucky dice and you're dead. So, don't use these as standalone encounters.
* Mooks if they're not too low level are credible threats. Monsters by design have high to hit so even if they're a level below the PCs, they can still hit. And they can help the boss flank, which increases boss crits.
* Facing a crowd of mooks with area damage abilities can actually be really dangerous, those stack up quite fast. Facing these seems to be worse than facing single-target mooks of the same level.
| 2 people marked this as a favorite. |
Based on my experience with 2E, I very much agree with the big single enemy encounters don't feel all that fun to play. Far too often it just feels like rather than these difficult encounters being difficult due to some capability unique to the foe, they're difficult solely because you're being ground down by the implacable weight of the creature's numbers. On the flip side, low level foes are squished again solely through the weight of your better numbers. The consistency in numerical capabilities between creatures and PCs of similar level means that any fight against a single foe feels surprisingly similar - getting pasted by enemy crits while you eke out small advantages (debuffs, buffs, etc) until you can even things out.
Despite how PF2E's encounter building promises balance and easy GMing, there are plenty of traps. Many monsters are just boring when solo. Others are too complicated to be a minion. Others have cool auras that introduce challenges for the players to tackle when a higher level than the party, but which only challenge the GM with relentless busywork and bookkeeping while having minimal impact when used as a minion. Basically, I feel like the encounter building guidance we have right now leads to fights that are technically "balanced" but are rarely a good time. The more I GM 2E, the more I find myself leaning back on the encounter building principles I learned back in PF1E.
|
|
| 3 people marked this as a favorite. |
I don't think we need to throw out everything, because that fundamental balancing of the numbers is super useful. But we should augment it with easy to understand guidelines to help design the most enjoyable encounters. Such as:
- How to see if a monster is a good boss monster?
- How to see if a monster is a good mook or lieutenant?
- How to see if a monster would be bad for those roles?
- What kind of changes can you make to a monster to make it better for a role?
- How can you achieve the feel of a solo bossfight, without needing those sky high numbers on the boss?
- How can you achieve the feel of a mass brawl, without the administrative tedium?
- How can you use the terrain to the advantage of the monsters?
- How can you use the terrain against a powerful monster, to weaken it, so the PCs can fight a normally out of band enemy?
- How to use foreshadowing to build up to a big boss
- How to ensure the PCs can actually come prepared with the right preparations?
Based on my experience with 2E, I very much agree with the big single enemy encounters don't feel all that fun to play….Basically, I feel like the encounter building guidance we have right now leads to fights that are technically "balanced" but are rarely a good time. The more I GM 2E, the more I find myself leaning back on the encounter building principles I learned back in PF1E.
[snipped for clarity] Any chance you could list or comment on those PF1 encounter building principles?
| 3 people marked this as a favorite. |
I don't think we need to throw out everything, because that fundamental balancing of the numbers is super useful. But we should augment it with easy to understand guidelines to help design the most enjoyable encounters. Such as:
- How to see if a monster is a good boss monster?
- How to see if a monster is a good mook or lieutenant?
- How to see if a monster would be bad for those roles?
- What kind of changes can you make to a monster to make it better for a role?
- How can you achieve the feel of a solo bossfight, without needing those sky high numbers on the boss?
- How can you achieve the feel of a mass brawl, without the administrative tedium?
- How can you use the terrain to the advantage of the monsters?
- How can you use the terrain against a powerful monster, to weaken it, so the PCs can fight a normally out of band enemy?
- How to use foreshadowing to build up to a big boss
- How to ensure the PCs can actually come prepared with the right preparations?
I think all the guidelines in the world won’t help me as much as tacit examples. I know it might be pie in the sky and incredibly niche, but I would completely devour a book of step by step thoughts on building various encounters and populating them with creatures, traps, hazards, environments, even if it were filled with tidbits and Golariona plot paraphernalia that I usually hate.
For example, I know reading Mathmuse’ posts always bring me great joy because they:
A) posit it within an actual ongoing campaign that
B) explains both the relevance to the metaplot and current situation (“In my PF2e Ironfang Invasion campaign”) as well as
C) the party composition so that
D) the tweaks to the encounter Mathmuse is making is really brought to life AND further explained by how a,b and c inform d, all while feeling like a vicarious playthrough of someone else’s campaign.
I have more trouble with “guidelines” than I do with representations of said guidelines. Even if it is along the lines of “we could do x but that might mean y” occasionally,
As for the initial post, I can totally get on board with the concept. Of course I would treat the encounter as more of an “ecosystem” than a “creature” but semantics aside, I get the point of adding HP and actions to the “creature” as a whole rather than focusing theoretically on the participle creatures. I mostly build encounters around terrain, environments to enhance the narrative dynamic, but also pay attnetion to the “tone” of the current trajectory of the play session. If things are lagging a little or the session is nearing a flat end I have someone “kick in a door with a gunne”. Worked for Raymond Chandler. It makes for a little seat of the pants, but it does create a lot of experience with winging it. But sadly, having played so little of PF2, nor GM’ed any, I am yet to avail myself of the vaunted ease of encounter design in PF2.
| 1 person marked this as a favorite. |
I don't think we need to throw out everything, because that fundamental balancing of the numbers is super useful. But we should augment it with easy to understand guidelines to help design the most enjoyable encounters.
I agree with the general principle: it's not that PF2e's encounter building rules are bad at all; they're generally excellent and in my opinion give some of the clearest and most consistent instructions for building encounters that I've seen. I also agree with OceanshieldwolPF 2.5 that, past a certain point, no rulebook is going to be able to tell you exactly how to elevate a specific encounter to be the absolute best it can be, because the most memorable encounters tend to be rooted in the context and narrative of the adventure, and that's where the GM needs to use personal judgment and experience and not rely entirely on general-purpose guidelines.
With that said, I do think the methodology outlined in the OP can help address several of the questions you ask, all of which I very much agree are pertinent to good encounter building and go beyond the remit of current encounter-building rules. For instance, creatures tend to be able to do better as boss monsters when they get more of the spotlight, which I think can generally be described as more actions to do their thing and more HP to stick around long enough to stuff. Conversely, mooks and individuals in a horde work best when they don't take too much of the spotlight at all, which I think can be described in the same broad terms: fewer actions to take up less time per round, and less HP for less sticking power.
Factoring all of this on a per-monster basis and adding environmental features on top I think would be fairly complicated, especially when trying to design a horde-style encounter, which is why I think the overarching methodology of giving the encounter itself a HP, action, etc. budget that you spend on monsters, hazards, and other stuff could help simplify things greatly. Even if you're applying templates or traits to individual entities to make them fit your encounter, having a set idea for how much HP and how many actions your encounter is meant to have in total I think would help prevent designing encounters that drag on for too long or end too quickly. I do think there's this fundamental clash where PF2e relies on its level-based scaling to scale difficulty, whereas this kind of system would work better if nothing scaled that way at all, but even so, variant rules like proficiency without level can help bridge that gap still.
Based on my experience with 2E, I very much agree with the big single enemy encounters don't feel all that fun to play….Basically, I feel like the encounter building guidance we have right now leads to fights that are technically "balanced" but are rarely a good time. The more I GM 2E, the more I find myself leaning back on the encounter building principles I learned back in PF1E.[snipped for clarity] Any chance you could list or comment on those PF1 encounter building principles?
I don't really have a good source to point at (though the GM's Guide to Challenging Encounters from Alexander Augunas is always good reading), or an organized list of everything, but broadly and in no particular order:
- CR is a starting point. After a few encounters you know better how much your PCs can handle, as well as what kinds of encounters are easy and hard for them. If a particular kind of threat is harder for your PCs, then consider it higher CR. The same is true in the inverse (for example, you can throw any number of mundane archers at the PCs if they're always packing wind wall).
- Players are always more stressed about the difficulty than it seems on your side of the table.
- Action economy is king. A foe with fewer actions only wins if it can control the PCs action economy. (Generally less true in 2E, though still valuable to keep an eye on)
- Complicated monsters are best used as bosses, mini-bosses or solo threats. Especially if they have auras, things that trigger very frequently, and action economy advantages.
- Simple monsters serve as better minions. Feel free to cut abilities off a monster to make it more minion-like.
- Variety in creatures, objective and terrain is mandatory.
- You can throw any amount of very weak foes into an encounter without adjusting the encounter difficulty. PCs will find a way to AOE them down, bypass them, or ignore them. They'll still provide "texture" and serve as body blockers. You can even treat them as having a damage threshold for KO rather than HP, to reduce bookkeeping.
- Unfair encounters are A-OK and strongly contribute to both narrative and game feel, so long as they are A) telegraphed both ahead of time and within the encounter itself and B) able to be escaped from in a way that is clearly shown.
Difficult encounters made up of hordes of low-level enemies, while technically possible, are a nightmare to run in practice, because having to manage lots of creatures and potentially dozens of actions a round really slows down play. This is likely why most moderate-to-severe encounters in Paizo's APs tend to feature smaller amounts of higher-level enemies, and why most people think of PL+3 encounters as solo bosses, or minibosses with a small number of minions or hazards. Swarms and troops technically exist to simulate lots of enemies with less micromanagement, but no rules exist as of yet for applying those traits to any given enemy.
Tangential, but I believe this is a large part of why so many players don't care to play spellcasters. It's often said that in 2E, martials are more suited to dealing with single high-threat targets, while spellcasters are more suited to quickly dispatching large groups of weaker enemies. Which, mathematically, is accurate. But, while I can only speak for myself, as a DM, I absolutely cannot be asked to run a fight with more than four, maybe five enemies, especially at mid-high level (for the record, I've only DM'd 1e, not 2e, but I've played enough 2e to understand it). As you say, it's a nightmare to run.
So while casters are suited to dealing with horde fights, realistically, practically, there's a very real chance that a given DM just isn't going to ever include such a fight in their game.
Just a thought.
| 1 person marked this as a favorite. |
One of the general assumptions held about Pathfinder 2e is that its math, by and large, just works,...
Heh heh, when I find math, I take it apart to see what makes it tick. The PF2 math (for example, Encounter Balance: The Math and the Monsters is simpler than the PF2 math (for example, The Mind-Boggling Math of Exponential Leveling, but both are built around hidden mathematical limits that define level-appropriate creatures. And my players discover and exploit those limits.
Looking at the guidelines in the Building Creatures section of the Gamemastery Guide and the actual creatures in the Bestiaries, we see that creatures defend well and hit hard for their level. For example, a 4th-level Tiger has AC 21 and a +13 to hit with its jaws. For a 4th-level human with Dex +3 like the tiger to have AC 21, assuming trained in unarmed defense and all armors, he would need a +2 item bonus from armor. But the tiger has no armor, not even thick skin nor dense hair. The tiger must have expert proficiency in unarmored defense, like a monk. For a 4th-level human with Str +5 to have a +13 to an attack (note that 4th-level humans can't ordinarily have Str +5), he would need a +2 item bonus from an enchanted weapon or expert proficiency in the attack. The tiger dodges like a monk, hits like a fighter, damages like a barbarian, and sneak attacks like a rogue.
So why is the tiger only 4th level? Because it is missing a lot of other features. Its only skills are Acrobatics, Athletics, and Stealth. It has no ranged attack. It cannot Treat Wounds. It's only feat-like abilities are Grab, Pounce, and Wrestle.
Imagine that a tiger is hiding in the tall grass of the savanna when a well-rested 1st-level party of cleric, fighter, rogue, and wizard walk by. The tiger would roll initiative with Stealth, and the +13 means it would have a good chance of going first and avoiding detection. The GM randomly rolls its target: 1 means the cleric. On its first action the Tiger pounces on the cleric, dealing 16 damage due to its sneak attack. Its second action is Grab and its third action is Wrestle, which includes a claw strike. The tiger's streak of good luck runs out and the claw strike with a multiple attack penalty of -4 misses. The cleric is now down to 1 hit point and grappled, but not prone.
This is a Severe-Threat encounter, so the party goes all out. The warpriest cleric goes first, casts a 2-action heal for 12 points and raises his shield for defense. The fighter strides to the tiger and Strikes at it twice. The rogue strides to flank the tiger and also Strikes twice. The wizard does Recall Knowledge about the tiger, learns about its relatively weak Will save, and casts her most damaging spell.
On the second round, a lot of the tiger's special abilities are irrelevant. Pouncing is pointless. It is not hidden, so it cannot sneak attack the non-grabbed party members. Its jaws are grappling the cleric so it cannot exploit its powerful jaws attack. This is by design. The tiger gets a frighteningly effective first round and then cannot follow up as effectively. It can Wrestle, so it does so. This time the claw strike from the Wrestle hits. The cleric Shield Blocks most of the damage, but is knocked prone. Then the tiger claws, hits again despite the raised shield and multiple attack penalty because it has a +13 to hit with a claw Strike, and adds in sneak attack damage because the cleric is prone. The cleric is unconscious at dying 1. The tiger would need its 3rd action for a Grapple check to keep the cleric grabbed, but since the cleric is unconscious, it skips that. If I were playing it as a real animal, it would try to drag the cleric's body away. But to keep the combat tense, I declare that it claws at the fighter. Due to a -8 multiple attack penalty and the fighter's good AC, it misses.
Then the myriad of abilities of a well-build party come into play. They have effective actions on their second round. The fighter and wizard continue their attacks against the tiger. The rogue is the back-up healer and knows Battle Medicine. In addition to Striking at the tiger twice, he heals the cleric. The cleric had moved to last in initiative, so he has time to rise, Stride away from the tiger, and raise his shield again.
On the third round, the tiger has its +13 jaws attack available again, but has lost its prey. None of its possible prey are flat-footed or prone. If this were just a random encounter, I would have the tiger run away. If this encounter was designed to show that the savanna was unusually deadly, then it would continue to fight. But all it has going for it is its high AC, high attack bonus, and high damage. The party has a 3-to-1 action economy in its favor even though they are succeeding in their Strikes half as often and each one could be taken down to dying by a single critical hit. Instead of Striking, the cleric is devoted to keeping his teammates healed up so that regular hits do not knock them to dying. The party could still lose, but they probably won't.
And it is by design.
Imagine that the GM wants a similar Severe-Threat battle of an advanced 5th-level tiger versus a 2nd-level party. The local village has asked the heroes to eliminate a dangerous tiger stalking the area, and the wizard made a Recall Knowledge check and said, "Tigers are 4th level, a moderate threat. We can handle it." Instead the GM wants to surprise them with the Severe Threat of an advanced 5th-level tiger.
The GM has to consider that makes a tiger battle act like a tiger battle and preserve those features.
We have the 6th-level Smilodon, also known as the saber-toothed tiger as an example of an advanced cat that resembles a tiger. But details demonstrate that it is not meant to be an advanced tiger. The smilodon has AC 23, compared to the tiger's AC 21, a +16 fangs attack compared to the tiger's +13 jaws attack, and gained Grab on its claw attack. Those feel a tiger at higher level. On the other hand, it replaced the tiger's Wrestle ability with a Pierce Armor ability, but that is probably to emphasize its saber fangs. The smilodon's Stealth is only +12 compared to the tiger's Stealth +13. The smilodon is not as much of a stealth stalker as the tiger.
The key elements of a tiger are the Stealth and the Grab. Everything else is routine. Table 2–5: Armor Class in Building Creatures says that AC 21 at 4th level goes up to AC 22 at 5th level. Table 2–9: Strike Attack Bonus lacks a 4th-level entry of +13, but all the entries go up by +1 at 5th level, so the +13 jaws and claws Strikes go up to +14. Saving throws go up by +1, too. The 4th-level tiger's 60 hit points is in the middle of the Moderate column in Table 2–7: Hit Points, and the middle of the Moderate column at 5th level is 75 hit points, but I am vetoing that. The fun in a tiger fight does not come from a slugfest. More hit points would lead to a tedious extra turn of combat simply to finish off the tiger. The advanced tiger will have only 70 hit points. To compensate, the tiger gets its Stealth raised by +2, rather than the +1 suggested by Table 2–3: Skills, so that the advanced tiger has Stealth +15. I really want it to ambush the party, bwahaha.
The 4th-level tiger's jaws 1d10+7 piercing damage (average 12.5) and claws 1d8+7 slashing damage (average 11.5) are Moderate in Table 2–10: Strike Damage, so they would go up by changing the +7 to a +8. I considered leaving them at +7 to make critical hits less likely to drop a PC, but doubling an extra +1 is not all that significant, and the extra damage would make it feel higher level.
For the adventure, as the 2nd-level party tracks the tiger in the savanna, it ambushes them and hopefully the fight goes a lot like the 1st-level fight. But the tiger runs away when the cleric gets back up. The party returns to the village and plans a better attack against the tiger the following day. Maybe they summon an eagle to spot the tiger from the sky, bypassing the concealment of tall grass so it cannot hide from the eagle. Or it could simply be that the blood trail of the tiger, still injured from the first battle while the party healed up, lets them track it better and go on high alert when it is close. This adventure sounds like it could be fun.
| 1 person marked this as a favorite. |
For example, I know reading Mathmuse’ posts always bring me great joy because they:
A) posit it within an actual ongoing campaign that
B) explains both the relevance to the metaplot and current situation (“In my PF2e Ironfang Invasion campaign”) as well as
C) the party composition so that
D) the tweaks to the encounter Mathmuse is making is really brought to life AND further explained by how a,b and c inform d, all while feeling like a vicarious playthrough of someone else’s campaign.
I am honored by the compliment. I have learned that applied mathematical algorithms need testing against authentic data, game design needs playtesting, and theorycrafting needs campaign examples. Besides, I love the adventures my players create.
The PF2-converted Ironfang Invasion campaign ended Friday, August 25, 2023. Then I started up a mini-campaign Playtesting in A Fistful of Flowers with 7 Leshies., That ended two weeks ago, and on Tuesday we will run the first session of a Starfinder mini-campaign.
The A Few Flowers More game session on Tuesday, October 3, (which I did not chronicle) had the 6-member 3rd-level party of leshies fighting against a 5th-level boss. That was frustrating for the swashbuckler. He had to succeed in an Acrobatic check against the boss's Reflex or Will DC to gain panache, and he failed often. The module had been written for four 4th-level characters, but I had dropped the party's level due to the additional players.
I continued the mini-campaign for two more chapters, one in Galt at 4th level based on the Scarlet Pimpernel Broadway play (of course, the Scarlet Pimpernel was secretly a pimpernel leshy) and one in a border town between Galt and Taldor at 5th level based on the Seven Samurai and The Magnificent Seven movies. I had learned from the boss battle in A Few Flowers More to not throw a level+2 opponent at the swashbuckler. They outfoxed 2nd-level guards in the Scarlet Pimpernel and fought 2nd-level mounted bandits in The Magnificent Seven.
Difficult encounters made up of hordes of low-level enemies, while technically possible, are a nightmare to run in practice, because having to manage lots of creatures and potentially dozens of actions a round really slows down play.
My players and I learned how to handle that, because they fought plenty of armies in Ironfang Invasion.
In The Magnificent Seven chapter the leshy heroes had a chance to scout out the bandit camp, which was a tiny village they had taken over. The scouting party consisted of the six 5th-level leshies and two 2nd-level halfling townsfolk, the deputy and the mayor of the village that the bandits planned to raid after the harvest. That is 270 xp in the heroic party. The plot required that they realize that the bandits are stronger than them and retreat to fortify the to-be-raided village. Thus, the PCs counted 24 2nd-level bandits, 360 xp. The bandits were 33% stronger than the party, a Beyond-Extreme-Threat encounter.
The players decided to attack the bandits at their home base, because they figured out how to even the odds. They waited until dusk when the dim light would conceal them from the bandits and their low-light vision would leave the bandits unconcealed from them. They planned to chase off the bandit's corralled horses so that most bandits could not use their mounted abilities against the party. Then they got a lucky natural 20 on my roll for weather conditions. It was a clear sky with a full moon. They could have dim light at midnight and attack while most of the bandits were sleeping. That shifted the overall combat from a single Beyond-Extreme-Threat encounter to a unbroken stream of around three Moderate-Threat encounters followed by an Severe-Threat encounter when the surviving bandits finally managed to group together mounted. But the bandits woken from sleep had no time to don their armor, so their AC was reduced by 2.
The party won. And the swashbuckler had a great time gaining panache easily against the bandits and sometimes taking one out with a single two-action finisher Strike. The barbarian leshy had half her Strikes against unarmored bandits be critical hits, killing them in one blow.
Likewise, in a battle against a high-level boss, the party needs to even the odds. A good example comes from my Ironfang Invasion campaign. At 12th level after a Great Victory in Assault on Longshadow they decided to free the enslaved villagers in Phaendar, the village where the adventure path had begun. The adventure path had planned for the retaking of Phaendar, but it was supposed to occur at 16th level in Vault of the Onyx Citadel. Commander Scabvistin of the Ironfang Legion was a bugbear rogue (knife master) 14, CR 14. I converted him to a 15th-level rogue NPC (fewer abilites and better numbers than a rogue PC) with two +2 greater striking tri-blade katars that he could use in Twin Feint and Disarming Strike. He was deadly in melee combat.
The party planned on avoiding combat. They dug a tunnel with a Dig-Widget to the slave quarters and would sneak the captured villagers out the tunnel to boats in the river. They did not avoid combat. The catfolk monk Ren'zar-jo was holding off an enemy troop unit from entering the slave quarters when Scabvistin showed up. He was no match against Scabvistin, but four other party members who had been fighting troops by the river rushed over to aid Ren'zar-jo.
Scabvistin was unmatched in melee combat compared to the party's level, so the party shifted the odds by not engaging in melee combat. That is the key to defeating high-level bosses. Do not face against their strength. Force them to battle with their weaker skills.
The sorcerer Gold-Flame Honeysuckle was in Dragon Form for the river battle. She offered the rogue Binny a ride. Binny was the party's debuffer, because she had the 10th-level rogue feat Precise Debilitations that let her sneak attacks render her target flat-footed to everybody. Honey flew in, landed out of melee range, and engulfed Scabvistin in a breath-weapon line. Next turn she took to the air again. Scabvistin pulled out his +2 striking composite shortbow, +29 to hit, and shot at Honey. Then Binny popped out out of hiding on Honey's back and successfully shot Scabvistin and debiliated him to flat-footed, reducing his AC 36 to AC 34. That mattered, because the ranger Zinfandel was shooting his longbow at Scabvistin from 100 feet away and that doubled the number of hits he got. Likewise, the stormborn druid Stormdancer zapped Scabvistin with Ray of Frost from the air. In contrast, Ren'zar-jo moved away to resume battling the troop unit.
Scabvistin was stuck in ranged combat. Honey was stuck, too, flying while waiting for her breath weapon to recharge. But Binny, Zinfandel, and Stormdancer were whittling away at Scabvistin's hit points faster than Scabvistin was whittling away at Honey's hit points. So four party members defeated Scabvistin by forcing him to battle with his backup weapon for which he had no feats and by reducing his AC with Precise Debilitation.