Ensign

If you play the character focused on offensive spells I'd prioritize PER over DEX, especially if you are dual wielding. The recovery bonuses from dual wielding dilute the DPS bonus of DEX down from 3% to ~2%, putting it closer to par with the other damage increasing stats. Also when it comes to cast speed, the big advantage it gives is the tempo of finishing casts before your foes do. From a damage perspective, unless you are chain casting the entire fight the extra attack speed is ultimately giving you more basic attacks, not more effective spells. When you are casting important heavy hitting spells making sure those land is more important than having them finish a little faster...because if they finish faster you get to auto attack more. Woot? If you aren't dual wielding it's a trickier trade off because the extra weapon DPS from DEX matters more. Then it really comes down to your identity as an attacker or as a spellcaster.

Pick your favorite damage dealer and roll. If you take Aloth as a Wizard I would take something more sustained, like a Cipher or Monk. With Ydwin I would take a Wizard or Monk.

Now that he has an Ascendant I have to imagine a priest is next, permanent ascension and BDD via Salvation of Time is exactly the sort of thing he'd love.

I'd argue it's one of the better modals in the game. Consider the alternatives of swapping to another weapon or turning on an alternative modal. If you swap to a higher penetration weapon, you take a base damage loss; if you turn on an alternative's penetration modal you lose DPS through higher recovery. Taking a deflection penalty via the sword modal instead is very similar to dumping resolve to boost might or dexterity, conceptually. You'd prefer DEX or MGT on a DPS character to resolve, right? Sword modal is preferable to sabre modal for the same reasons. It's one of the few really compelling reasons to use a sword (besides good uniques, of course).

That's a good point, melee + ranged dual wielding should generally outperform single weapon unless you need those extra crits to penetrate armor.

I would emphasize that a Mental Binding opening to mess up enemy engagement into Secret Horrors is very strong, especially early in the game. The main issues with Beguiler are more meta than mechanics. It is great with AoE deceptions that refund or even generate focus later in the game. In practice that means Secret Horrors and Mind Plague; Honorable mention to Ringleader, though I have been less impressed with the skill now that charm breaks easily. Outside of those it falls off pretty fast. Mental Binding is a great opener in a lot of fights to mess up enemy formation and engagement, but it loses value past the initial seconds of the fight and drops in priority as you get access to more impactful powers. Eyestrike is fantastic on paper but is badly outperformed by Chill Fog in most groups; Phantom Foes rarely seems worthwhile now that all perception afflictions flank. The other major problem it runs into is boss fights. While the aforementioned tricks are great early game, most bosses are STR/DEX/RES resistant INT immune, which is a fancy way of saying 'Immune to Beguilers'. Here your lower base focus generation and lower damage (which also lowers focus generation) makes you notably weaker than a core Cipher. I would argue the biggest benefit of the subclass is how nice it makes Mind Plague, as with the bounces it will consistently return a big chunk of its cast cost, making it a powerful spamable debuff. It's definitely a reasonable subclass design and does some things really well. I just find it narrow, and the 'much weaker in the hardest fights in the game' is a big enough drawback that I can't ever recommend the subclass to a new player.

It chooses between the two procs randomly, and only deals 4 damage per 3s tick. It's pretty aggressively bad.

It's not really possible to solve your deflection issues on a melee Witch. Best is to just accept it, dump resolve, and stack armor to try and make the hits and crits manageable. Devil of Caroc breastplate is ideal in general. It's a great armor and the INT affliction resistance deals with the confusion and allows you to function. You can use food as well, but it is expensive to do so and until you have the game nailed down I would hesitate to commit to it as plan A. Two handers are definitely ideal. Dual wielding is only really sexy on Rogues now, with a few additional interactions making it nice (Kind Wayfarer). Witch doesn't have much of that (just Interrupting Blows really), so I would eagerly use a two hander. Be prepared to pay close attention to the micro of the character. It can dish out pretty impressive damage, but it does need careful timing and positioning on top of the usual class babysitting. A longer range support would be advised.

Resolve is not checked all that often in conversations, and most of them are pretty minor (and I mean really minor - stuff that checks resolve 5, some literal resolve 0 checks, etc). The handful that make a difference are in the resolve 15-16 range. Note that these are not at all like the persuasion checks in PoE1, but are more often about resisting mind control, shaking off illusions, and stuff like that. As far as conversation stats perception and intelligence matter way more. Persuasion is governed by diplomacy, with alternatives of bluff, intimidate, and streetwise. If you want a tally character focus on those skills.

A similar process, but I suspect your model involves a lot more cases of severe under-penetration. In a vacuum, hand waiving, the 15% hit to crit is worth ~2 accuracy at full pen, closer to 3 at -1 pen, 4+ at -2 pen and 6+ at -3 pen, assuming crits get full damage. So if you have a lot of serious under-penetration in your model the hit to crit looks a lot better. But in practice I spend very little time swinging into -2 pen unless it's garbage time mop up. At -1 pen you just shrug and bear it, but at -2 pen you are inevitably better off swapping to a higher pen weapon; and if you still are underpenetrating you turn on the +pen modal, or you debuff armor or something. I used to model it pretty heavily, but I was basically modeling 'playing like crap' and what's the point? So I swap in the -3 base damage for underpen situations now, which prices penetration more reasonably (about what you pay for a high pen weapon in game) but de-values precision and hit to crit since they aren't dominating due to all your damage coming from crits. As for the relative merits of the extra crits vs faster recovery - it can go either way and it's pretty marginal. If that was all there is to it I'd think it's a spectacularly boring subclass and ripe for a redesign, but...eh, not everything is a radical game changer. Except that isn't all there is to it - you're paying a -10 deflection penalty for the privilege of jiggering your damage from marginally faster hits to marginally bigger hits. While I know there is a sizable contingent of 'backline defense doesn't matter at all, Guilty Conscience and a bathtowel #yolo' around here...that penalty is not nothing. It's not a huge penalty, and it's certainly something you can play around and mitigate with proper party positioning and engagement blah blah blah...but you are taking that penalty for the privilege of fiddling around with 2% DPS differences and damage compositions. Subclasses like Arcane Archer and Assassin at least get some oomph for their defensive penalties - I'd argue that Sharpshooter's penalty is bigger than either of those, and for what? I love the strength of Barring Death's Door on an ultra squishy murder machine going ham in a towel, or the really spectacular numbers from the cannon hidden in the back...but the opportunity cost is needing to have someone there to put the BDD up, needing to defend the cannon against divers / shades / archers that get a little frisky / AoE in cramped dungeons / etc. You need the payload. Where's the payload?

Anyway, to cut through all the crap there are two overriding philosophies, across games and genres, that apply to backline characters in this game: 1 - Backline characters only gets hit because you are bad. Dump RES dump CON, bath towels for everyone! Get better nubsauce or you're never gonna get peak performance! 2 - Backline characters get hit when you play bad. Fortunately they're wearing mail and have shields so it's no big deal. Even a total nubsauce can win with this. There is inevitably a lot of social pressure towards 1. It becomes easier with experience and provides more opportunity for skill expression and really spectacular plays. Most players should pick 2. Deadfire is a robust enough system to support both well.

I'm more curious to know why you think the 15% hit to crit can't be used as a worthwhile trade-off to the 10% recovery time penalty? (the two upsides can also be used in tandem with the same weapon depending on the situation) It's a question of what the use case for it needs to be. To minimize the drawback of the speed penalty you need a *lot* of speed modifiers; the hit to crit conversion is most valuable with low accuracy. So ideally you'd want to use it with a hunting bow or single/dual pistols with the modal on - but not pistols because blunted criticals nullifies the advantage of the hit to crit; with the modal off it's pretty hefty damage loss. So call it a hunting bow Specialization. Under ideal conditions (hard to hit target) you can squeeze about 1%, 1.5% more damage out of a sharpshooter. But you can get a bigger damage boost by turning the modal off and taking the free accuracy. With the modal off...base ranger does appreciably more damage. So a good start but not enough on its own. The other part is if you have a low hit rate and consistently under-penetrate. -1 pen doesn't cut it, but at -2 it is interesting. If your crit rate is low enough sharpshooter starts to look good. Until you're down to -4 underpen stepping into close range is preferable, but that isn't always possible. You're also better off swapping to a war bow for the better penetration, but there base ranger will again out-damage a sharpshooter. So on paper the best use case is a sharpshooter / devoted with a hunting bow; I think in that case you can probably squeeze 2, maybe 3% more damage out of it than a base ranger. An easier way to think about it is that 15% hit to crit is generally about as good as 2.5 accuracy. You can twiddle with the other components that make recovery and accuracy more or less valuable, but that is generally not a very exciting trade.

You're right, that was a mistake. It's not a strict downgrade. So for example I honestly have no idea how one could claim that the sharpshooter is "strictly worse" than a vanilla ranger. It is a pair of upsides (bonus PEN or bonus hit to crit) traded off with a downside (+10% recovery time and a barely-relevant deflection penalty). That's basically how all subclasses work, as you say, and both the upsides can be easily taken advantage of in exchange for the +10% recovery time. The former upside can be taken advantage of in exchange for the 10% recovery time. I'd be very curious to know how you'd take advantage of the latter. What constitutes a trap is the ease of building some nonviable or extremely suboptimal. I mean I'm not talking about equipping a wand on a shattered pillar, but requiring an extreme amount of metagame knowledge just to not suck. Under this definition I agree with you entirely. It is very hard to make choices that leave a character totally non-viable. That reflects the strength of the design of the game and is a big part of why I think it's worth playing past story mode.

Strictly speaking every subclass in the game has both unique benefits and downsides, so with sufficient system mastery you can build a character around those differences and make something that isn't strictly worse than the vanilla class. I think that is an impossible standard for what makes something a trap. The players posting here generally have a high level of system mastery. We're going to be able to figure out how to take advantage of corner cases. Much more important to the game's design is how choices appear, and how they play without a high level of system mastery. Those are the players that are going to be 'trapped'. Thus the real questions for a design are, 'does this design offer a clear value proposition?', and 'does its choice deliver on that value proposition?' Strictly speaking, Sharpshooter has a niche - the blunderbuss build, while it gives up a lot for that point of pen, is a build that makes sense. If you where thinking you wanted to play a long range pew pew Ranger and clearly Sharpshooter is for you, well...dealing *lower* DPS at the cost of lower defenses is probably not what you had in mind. Oops. Control Cipher was a very popular character in the first game, and Beguiler is clearly there to double down on that role for those players, right? Won't they be disappointed to find that Ascendant is better at pounding foes with a steady stream of debuffs in almost every way? That there is a corner case where you're better at spamming low level debuffs is little to offset the disappointment of Ascendant was the right choice if you wanted to spam high level debuffs instead. Priest of Eothas has a valuable corner case as a multi-class core priest in a build that wants to focus its points on the other half of its multi-class. It's the priest version of Stormspeaker, in a sense. Cool story. How many players thought to play a single class priest and picked Eothas because he's both the canonical 'good' deity and central to the plot of the game, and accidentally stumbled into a poor choice? I stand by all of the listed subclasses as either deserving buffs (because they fail to deliver on their value proposition) or reworks (because their value proposition lacks clarity or is too much of a corner case).

There are tons of items in the game that provide +12 accuracy. There are skills that provide +100% damage; passives that give 10 accuracy or 20% damage are common. None of that has anything to do with how effective their respective attributes are. All that matters are the marginal effects and opportunity costs of the different choices available. <cut a lot because reading this argument just makes everyone involved stupider, I apologize for indulging it>

Not to devolve this into yet another Resolve thread, but while the headline of 'dump RES for maximum effect' is generally correct, the argument of it being a completely useless attribute is not. When you are allocating attribute points you are selecting between the options on a 1:1 point basis, so only their relative values matter. In most cases, a point of some other attribute will have higher value than a point of RES, so the 'best' allocation dumps RES. But that does not mean RES is worth nothing, just less than the next best option. If, for sake of illustration, Resolve gave two points of deflection per point instead of 1, it would be ideal to pump RES to the max on virtually every character to maximize their power. So RES is clearly not worthless. A good faith estimate would put a point of RES as equal to about 0.7 points of DEX/PER/MGT on average, with specific builds of course valuing it more or less highly. But that would mean the difference between an optimally dumped RES and a 'baseline' RES isn't 7 'wasted' points, but closer to 2 in end effectiveness. All in I would price it at about a 4% power loss in the abstract. That isn't optimal! It also certainly isn't unplayable, and is honestly pretty minor in the grand scope of sub-optimal build decisions for aesthetic reasons.

Not necessarily utter trash, but generally poor choices (excepting the brand new specializations that haven't seen a pass yet): Corpse Eater: penalties too severe and hard to use for the benefits. Mage Slayer: severe penalties make it often a trap choice. Stormspeaker: not usually mentioned because it's multiclassed, but single class chanter Tekehu is ready bad. Beguiler: worse at its described role than Ascendant with no real upsides. Wild Mind: lololol. Black Jacket: I don't know that it is underpowered given a relatively minor drawback, but the role that this was envisioned to fill was nerfed hard enough that it isn't really attractive anymore. Shattered Pillar: honorable mention. It does have a niche, but has so many drawbacks that even in places where you might otherwise see it vanilla monk is better. Sister of the Reaping Moon: not too bad, but almost a strict downgrade from vanilla monk. Darcozzi Paladini: while every other Paladin order has found a mechanical niche, Darcozzi basically exist as a no frills roleplaying order. Nothing really to advise it here. Brotherhood of the Five Suns: another NPC class - actually worse than a theoretical vanilla paladin with no special benefit. Priest of Eothas: ultra generic spell list makes it's only niche an ultra second class focused multiclass, and I don't think any of those are particularly good. Priest of Berath: similar to the above. Priest of Gaun: a bad version of Eothas. At least Blessed Harvest is good for Maje Isle? Sharpshooter: almost a strict downgrade on vanilla Ranger. Excessive drawbacks for minimal benefits. All wizard specializations except Evoker - penalties are excessive for very minor bonuses. Many of these don't scale appreciably with power level. In addition, rogue and chanter have powerful, generic enough subclasses in trickster and troubadour that their vanilla versions are almost always a poor choice.

The hit to crit is actually the main reason it isn't competitive with dual wield or two handers, on paper - it is a much weaker trait than two handed and two weapon styles. All else equal, single sabre is around 10% less DPS than a two handed sword or dual sabres. The extra accuracy does make it more reliable, however, and the use of a strong one hander for all your attacks instead of switching back and forth between two weapons can definitely make up that 10% difference. I haven't run the numbers explicitly but I would suspect a single high damage sabre (say, Tarn's Respite) would out-perform any unique greatsword in a broad spectrum simulation. Dual wield devoted falls behind single weapon devoted for similar reasons, I bet - it's really hard to find a weapon type where both weapons are close enough in power to outweigh just using the better one more often.

Stats don't make that big of a difference. I would want high INT for better buff AoEs and durations but the rest is fungible. No point in maxing RES but not dumping it isn't some huge handicap. Moderate CON, pick MGT, DEX, PER to taste. Early game medium armor is a sweet spot of effectiveness and is reasonably common. Can move to light with the added defenses of a few levels. Armor choice is going to depend a lot on the uniques and how you want to distribute them to your party.

Devoted / Barbarian or Devoted / Monk are the only combos I would consider a greatsword on. Don't worry about the modal, it is only a damage boost against trash on low difficulties. The incoming damage is an issue for a melee Helwalker. I personally think it is a better combo with a spellcaster (which benefits more from the additional might) or a Paladin which is tanky enough to not care. For a Devoted / Monk i would use Nalpazca or vanilla.

Stalker does the most to emphasize the pet, yes. Ghost Heart works really well with Cipher, as you can cast the pet behind your enemy to flank / suicide onto the backline while giving you a body to anchor spells on - Ectopsychic Echo and Amplifying Wave chiefly.

Single class Eothas is pretty simple, you take the good spells and a bunch of passives. Restore, Dire Blessing, Devotions of the Faithful, Barring Death's Door, Salvation of Time, Storm of Holy Fire. Those plus the free spells on level up are 95% of what you'll need. I think Holy Meditation is good and maybe you'll cast Shining Beacon now and again - and all of the PL8 stuff is good except Crowns. But for the most part take all the passives. Monk + Priest I would do the scepter/helwalker thing. Self damage from scepters to build wounds, which give might and int which buff your spells. Cast spells as above. Priest / Cipher is just so bad. No way I can suggest that. I'd single priest from the choices listed.

Pets scale with character level, not power level, so they are equally good on a multi-class and single-class Ranger. Most of their important traits are in the lower tiers as well, so you don't get any notable advantages from gaining Ranger levels faster. My feeling is that the party build matters more than the Ranger build when it comes to pets. Pets generally need to be supported to be effective, but Ranger is not particularly good at that. Herald is the cheap answer, but it's very true - Chanter phrases and Paladin auras buff pets as well as party members, and they go a long way towards making it feel more effective. Having a healer that can fix up, or revive, a pet is also valuable to have so you aren't stuck with Bonded Grief at the first mistake. Otherwise the key is to micro your pet. They are kind of caught in a place where not microing them is suicide, but microing them isn't super satisfying when you could be microing a flanking Rogue or Barbarian instead - but if you enjoy the micro game they're pretty effective. Use them as off-tanks or flankers, where they won't take focus fire, and they'll be pretty effective - especially if getting the aformentioned passive buffs.

It has incredible gameplay for the genre. Really top notch, fluid, fun to play. Those games live and die on reward loops and economy though. It *needed* an auction house to be the 1000+ hour, huge active player base game they wanted it to be, but their reward loops simply did not do enough to support that. Without the auction house they were able to re-balance around a less integrated economy, which made for a better core game experience but that loses engagement faster. They had no hope of that long term value prop that Path makes so much money off of. I know Blizzard had economists looking at their economy model because on a high level it was a good design (for the time, flaws are obvious in retrospect), but they didn't have any game design / balancers who know a damn thing about the game economy because it didn't reflect their model at all. ...but I digress into my own real world rabbit holes and get far off topic. ...ok I can't help but nerd on it a bit. You also have to keep in mind that D&D and its descendants are a ~40 year old product from an era without a lot of competition. It was a revolutionary product! It was also really poorly optimized and a lot of the rules were bad and it didn't matter because it was the only thing of its kind, and it got enough right that radical departures didn't really stack up. It has been 40 years, however, and we know a ridiculous amount more about game system design than we did then. At the same time D&D and its lineage has a ton of design debt, locked in because they want to create new versions for players that are locked in from a combination of nostalgia and familiarity. If you want to make a game for those players they *can't* shake things up radically. Everything on that system is the modern equivalent of put putting around in your Model T because it was great god dammit, why would you ever change anything? It works well enough for tabletop gaming because the market for radical system design there is so limited, it's all about the modules and fitting them in seamlessly. On the flip side video games have at a minimum 8 figure and frequently 9 figure development budgets, with a need to hit sales that justify that, and a crappy nostalgia system isn't going to get it done there when we know how to build race cars now. I totally get why BG and friends grabbed the D&D system along with the IP to tap into that tabletop market - wise at the time - but doing that today is just a bad decision, betting big to try and squeeze small returns. Expectations are too high, you need to do better.

Yes, elegant. In a system with compounding effects our standard approach of arithmetic and geometric percentage changes is really poor. I like the way they handle the math under the hood a lot. It splits the difference between a more natural Platonic scaling system and a more cleanly understood geometric system; unfortunately it gets counterintuitive when you start stacking positive and negative modifiers. You can reproduce some of the same effect with a combination of multiplicative and additive bonuses and maluses, for sure - but then you have an additional layer of complexity to communicate, and you don't really solve the problem of how to handle maluses (your system only handles multiplicative maluses). That they didn't finish squaring the circle didn't help either - their power level scaling outside that system is a clear systems mistake. But their double inversion approach is really clever. One of the best solutions I have seen. Cleanly handling all modifiers with the same sign additively, but of opposite signs effectively multiplicatively, gives you a system that won't blow out but also maintains high impact of individual modifiers.