In PF, how would you optimize for soloing Rappan Athuk?

[upho]

To be fair, there are a couple of abilities that Orcus as written has in all his forms that can be seen as pumping his effective CR a pretty fair bit, depending on the DM’s proclivity for doing so to the hilt, especially the second:

First, I think you may have missed a very important sidebar found on three different pages, because Orcus doesn't have those abilities in all forms, and in a real game I think you'll agree it would be highly unlikely he'll have any of them when facing PCs who have managed to find his lair:

Spoiler: What the sidebar says

Show

For each shrine the PCs manage to cleanse, the Master gains 2 effective negative levels. In addition, if one shrine is cleansed, the Master loses the use of his 9th level spells. If 2 shrines are cleansed, the Master loses all his summoning spell-like abilities. If all 3 shrines are brought low, the Master is reduced to the following: DR 15/cold iron and good, SR 36, all At will Spell-Like Abilities are usable 1/day, and he loses the use of all Domain and Class Spell-Like Abilities.

(My emphasis.)

It's extremely improbable PCs who start at a reasonably early character level will find a path back and forth through the dungeon to Orcus which never passes through the shrine on level 4, or one which makes them somehow miss the almost equally obvious second shrine on level 9. And I suspect it's even more improbable the PCs won't destroy the shrines, as even though they most likely won't realize it directly affects Orcus, they'll still at least want to render the shrines' useless for any surviving cult members or undead. Not to mention that finding and destroying the shrines (and the cult) may be the very reason why many PCs or parties want to explore RA in the first place.

Second, in the hypothetical scenario of a lone high-op PC facing Orcus at full power, I doubt even the most powerful of these abilities will necessarily have enough impact to actually alter the outcome of the fight:

Spoiler

Show

1). Summoning demons — this shouldn’t necessarily be unexpected, but adding up to 3 balors or around 7 mariliths (specifically atop the mariliths already present) isn’t nothing, but more to the point

Practically any PC build I can think of which could achieve the combat values required to stand a chance against Orcus himself would barely register 3 balors, as their combat stats simply aren't high enough to matter. And 7 mariliths would probably have about as much impact on the outcome as 7 goblins.

2) Summoning undead — Orcus can summon 100HD of any undead (including, specifically, templated undead). If Orcus has any idea what the party is or does he can tune this ability — depending on the party’s composition, mobbing them with shadows might not actually be terrible, but book-diving and designing Orcus’ summons, yikes — “even” 4 semi-optimized 20th level liches should be, to put it mildly, a real problem. (Also, recall that all undead in the Boss Fight room, including the prestocked rabble, are fiat unturnable, and presumably could share Orcus’s and each other’s buffs.)

Yes, if this works as you say, it could definitely make the fight incredibly difficult, if not flat-out impossible if the GM knows how Orcus can get the most out of it. However, per RAW I believe summon (Sp) abilities can only summon generic creatures, never unique individuals (just as the SM line of spells the ability is based on), which of course limits the customization possibilities of the summoned undead significantly and in turn their usefulness. And while the most powerful generic undead - like say a great wyrm red dragon ravener (underrated CR 24) - are certainly dangerous foes to most parties, as in the case of the balors, their actually relevant numbers likely wouldn't be sufficiently high to really matter against most high-op PCs able to take out Orcus.

And it's probably worth noting these summoning SLAs can only be cast during Orcus' own turn and the summoned creatures remain for 1 hour. Which is a long time for summon spell, but may still turn out to be too short for casting them before combat since Orcus doesn't initiate the fight as written. So he may need/want to save especially his far more powerful 1/day summon undead for his first turn to ensure he at least has a chance of benefiting, which of course also means the summons won't be able to help protect him against one-shot tactics as neither Orcus himself or any of his minions are likely to win initiative.

That said, if the summon undead ability can be properly timed, if nothing else it might at least provide a real advantage simply due to the summoned creatures' potentially great size and many hp, acting like a mobile wall of flesh which may protect Orcus against for example charge shenanigans.

(As an example, the mentioned Master Summoner build would likely first buff herself and her 15-20 or so minions to the high heavens. As a comparison, that's most likely more than 300 hd of augmented flying Gargantuan CR 14+ minions, plus maybe two CR 22 psychopomp "lieutenants" with cleric levels. Then she'd win initiative with her +35 or greater bonus and start by having her minions storm the room before she herself dives in on her falcon at up to 440 ft. speed (enough to ignore a large majority of BFC effects/abilities) and full attack Orcus with her 10 pistols from within 60 ft. Each of her 14 attacks would hit on a roll of 2+ vs touch AC, have at least a 50% chance to crit - even should Orcus somehow not be flat-footed - and ignore all cover, concealment and wind spell defenses, dealing enough total average damage to kill the poor demon lord almost twice over.)

Given that summoned monsters in principle don’t add to the EL of an encounter because they are supposed to be accounted for in the CR of the summoner... Orcus’s summons alone could be 3x CR 20, 4xCR 22, or around EL 27 not counting environmental buffs for the level or room. It’s debatable whether that pushes Orcus all the way to CR 35, but it presumably establishes a bit of a floor as to his “actual” CR, and means this is one BBEG fight where action economy is not necessarily on the PCs’ side — Orcus summons can each have actions that are meaningful threats vs PCs, never mind that PCs looking to even the odds with battlefield control may find Orcus-liches, never mind Orcus himself, battlefield controlling right back at them.

Well, at least in theory and assuming a more "average" party I agree this is a valid point. The problem is that in practice such a party has basically been turned into obedient little wights the second they roll initiative if they face Orcus as written at full power, completely regardless of whether he uses one or more of his summon SLAs. And again, a PC actually capable of posing a real threat to Orcus at full power will most likely also not be significantly hindered by any of his comparatively very weak summoned minions, so the PC might very well still have the actual action economy advantage even should Team Orcus be able to perform ten times as many actions per round.

In short, these abilities provide lots of hd, but most likely no or very few actually relevant numbers. And again, since there are fewer builds able to stand a chance against say CR 30 Cthulhu...

But it may of course be that my way of looking at high CR ratings is actually off, and the theoretical challenge a creature poses to "average" low-op parties is the only relevant measure.