WHAT THIS IS
Trying to do something different here. I'm going to discuss a character build that I had a lot of fun putting together and playing on 1.1 PotD (probably the most fun I've had in Deadfire so far), but I'm going to use it as an entrypoint to teach-by-doing various game mechanics and how to think about them and use them for your own builds. People already comfortable with min-maxing their way through Deadfire might not get much out of this, but hopefully for everyone else there's something useful here, because there's a lot of information scattered through this forum, reddit, and the in-game cyclopedia and I hope to consolidate some of it here and put them to actual use. In fact, I would say something like 90% of this is just going to be talking about game mechanics, because understanding the game mechanics thoroughly is key to understanding this character build. I hope you find this useful!
Oct 2018 - Updated stats, items, build for 3.0+, added alternatives, and god challenges notes
July 4 2017 - Updated stats for patch 1.2.
July 3 2017 - Partially updated for patch 1.2.
TABLE OF CONTENTS
- What does this build do
- The build itself
- TL;DR playstyle
- Detailed playstyle
- Details - Armor
- Details - Weapons
- Details - Surviving
- Details - Consumables
- Details - Miscellaneous spells and abilities
- Details - Stats
- Conclusion/Putting it all together
- Rejected approaches
- Notes for Magran's Fires
- Appendix: Linear returns
"What the hell is an 'Umezawa?'" Well, while I don't play it anymore, I still follow and am a big fan of Magic: The Gathering. And with one of the more recent sets out (Dominaria), I got to thinking about a couple of cards that struck me as particularly flavorful:
The latter card is actually from a set more than a decade old and is/was so stupidly powerful that it's banned in one of the super-powered tournament formats (banned in Modern). The former card is just a nostalgic callback to the latter that was in Dominaria (which itself was a set all about nostalgic callbacks). It's not particularly tournament-worthy.(*)
(*) If you're a M:TG lore purist out there, yes yes I know that the Umezawa on the left is not the Umezawa that the Jitte on the right is named for.
Even if you don't play Magic: The Gathering, the takeaway here is a fragile, weak, but elusive hero, and a stupidly powerful tool-kit of a weapon that the hero is lorewise linked to. I liked the idea of taking a very blue (crafty, subtle, evasive) approach to stupid levels of power, and decided to personify it in Deadfire as a Streetfighter/Wael multiclass that I'll just brand here as an Umezawa build.
WHAT DOES THIS BUILD DO
This character is very versatile and mobile, tanking on huge packs of enemies or jumping around behind enemy lines to assassinate troublesome foes. This character is not going to be great at doing burst or area of effect damage (at least until you have a steady supply of explosives), but it will do frankly a stupid amount of sustained damage, maximizing uptime for the Streetfighter's special (at least without resorting to just being a ranged blunderbuss dummy).
While this is probably not the most powerful Streetfighter build you can make (a Streetfighter/Monk is probably better for pure damage) and is fairly micromanagement-intensive, it can nonetheless be a very fun and engaging way to play.
THE BUILD ITSELF
Before diving into the mechanics of it all, let's just lay out the build order.
CLASS: zealot - streetfighter + priest of wael
RACE: wood elf
BACKGROUND: living lands + scientist
STATS: 10 (9+1 living lands) might, 10 constitution, 17 dexterity, 16 perception, 18 intellect, 7 resolve
SKILLS: roughly 2:1 ratio between Explosives to Alchemy. For your secondary skill, shove as much as you can into Religion, though you can respec out of this after a certain point. Be sure to pay 3000g each to train both Explosives and Alchemy.
STORY ABILITIES: be clever at the first summons to the gods so you get Wit of Death's Herald
ABILITIES: (active priest abilities with arabic numerals, active rogue abilities marked with roman numerals for their power levels, automatic priest spells in angle brackets, weapon choices in italics) Updated for 3.0: some skills reordered, dropped Searing Seal for Champion's Boon, weapon proficiency recs changed
1. Restore (1), Escape (I) <Arcane Veil (1)>; hatchet, blunderbuss
(**) Defenses offer increasing returns because of the way attack rolls work in Pillars/Deadfire. The higher the relative defenses to an attacker, the more significant each additional point of defense becomes. A pair of examples illustrating this: let's say you have 0 defense against an attacker with 25 accuracy. What would the impact of 5 additional defense be? Well, with Deadfire's rolls, you'll go from being graze/hit/crit by the attacker 100% of the time to 95% of the time. Not much of a change in your total survivability. Now, let's say that you have 95 defense against that same attacker. What would the impact of 5 additional defense be? You'll go from being grazed 5% of the time to never being touched ever again. You will have gone from having finite--if huge--survivability to literally infinite survivability: that enemy could attack you until the heat death of the universe and they will never so much as reduce your health by 1 point, where with a 5% chance to graze they could probably kill you within an hour if you do nothing but stand there. This might sound like an absurd example, but for many types of "The Ultimate" runs for Pillars of Eternity 1 (beat the entire game solo on the hardest difficulty in iron man mode) some fights could literally last for more than an hour and the difference between being grazed 5% of the time and never being touched could be the difference between a successful run and one that fails after many hours of playing. Plus, when you are surrounded by enemies in melee and targeted by enemies at range, even a 5% chance to be grazed can be significant; when there are ten enemies on the battlefield one will be expected to graze you every other attack, which will quickly add up to something fatal if you aren't killing things quickly).
You'll note that I list Escape as a source of deflection. And while it lasts an extremely short time (3s), with decent intellect and stuff like Meppu/Roe it can last almost 5s. Early on it can be a way to get some extra emergency protection without burning an Arcane Veil (especially since you don't have many other non-situational Guile-spenders you can potentially just chain together a bunch of Escapes to your current location since it has only a .5s base action time and no recovery), and at all points in the game can be used so that you can jump straight behind enemy lines and have a few seconds of unconditional +50 deflection protection to buff yourself or do something else. (If you want to be tricky, you can use Salvation of Time to extend the unconditional +50 defense of Escape by 20 seconds.)
Now, it's important to highlight that this character is not intended to be an immortal riposte build. We could just leave a large shield equipped and stack on all sorts of bonuses to be untouchable, but frankly I find that playstyle boring (I did that for my own The Ultimate achievement run and while it was certainly impregnable it was also tedious). What we really just shoot for is enough defenses and debuffs to not be squished into oblivion within a few seconds of being flanked, so that we can go on a killing rampage. This character picks up riposte not because we are going to rely on it for as a centerpiece for our damage, but as an accent of some additional damage in certain situations (which we maximize by dual-wielding since riposte does a full attack). In fact, in the late game, we may just want to deliberately get to near death ourselves and not worry so much about defenses, which leads me to another aspect of the survivability equation:
CAN NOT DIE EFFECTS
If you played Pillars of Eternity 1 with a priest, you'll be forgiven for ignoring Barring Death's Door and similar effects in Deadfire, since the effect in Pillars 1 was pretty lame. All they did was prevent you from dying, and dying in Pillars of Eternity 1 meant getting knocked down to 0 health (as opposed to 0 endurance), which would have been a permadeath instead of just a knockout. What Barring Death's Door and similar effects do in Deadfire is prevent anything from reducing your health below 1 (even instant kill effects from something like Death Ring). Fortunately for the Streetfighter, one way to trigger Heating Up is to get Bloodied or lower, and the only way to get On The Edge is a combination of being Bloodied or lower and being flanked. Both Barring Death's Door and Potion of the Final Stand give us a nearly foolproof way to trigger Heating Up and sustain On The Edge. But both Potion of the Final Stand and Barring Death's Door have low base duration. It's for this reason why we pick up something like Prayer for the Spirit (+5 intellect means an extra +25% of base duration for Barring Death's Door), invest in Alchemy (+5% duration per point in Alchemy to Potion of the Final Stand), and love food/drink like Meppu/Roe (+15% beneficial effect duration, additively stacks with intellect and power level scaling) and importantly why one of our most important late-game spells is Salvation of Time (+10 seconds to beneficial effects but as of 1.2 actually grants +20 seconds at least the first time you cast it per encounter). This stuff also helps Arcane Veil--which has a fairly short duration as well--but is more critical for these "can not die" effects because the moment they wear off with you at 1 health, you are probably going to be knocked out.
This leads to the second of this character's weaknesses, and it is Arcane Dampener. It's not too common for much of the game, but during the Paradise of the Mind quest and Nemnok the Devourer quest, literally every enemy wizard will try to hit you (and especially this character) with Arcane Dampener at least once; will Arcane Dampener temporarily suspend any current protections for a long time. You can try to hope that your will defense--which will be sizable thanks to a high intellect (buffed further by Prayer for the Spirit)--protects you, but this hope is dangerous because even a graze will dispel all your protections for a few seconds, which is more than enough time for you to be interrupt-locked to your death. You have two main approaches. First, you can try to use Smoke Veil to go invisible the moment you see the Arcane Dampener icon appear above wizards' heads (they tend to all cast it at the start of the fight, so if you see one you will probably see a lot) and let the enemy wizards re-target it to another member of your party who is less reliant on spell protection for survival. Second, you can try to use Smoke Grenade/Smoke Cloud or something like Grenades or Concussion Bombs to interrupt them while they are trying to cast it. This is very risky because if you miss you don't get another chance to interrupt them, so generally prefer the first approach. Either way, you should then eliminate the enemy wizards with extreme prejudice, because some of them will hang on to their third level spell cast to try again later.
DETAILS - CONSUMABLES
In addition to blunderbussing with Powder Burns or charging in to get flanked/brought to Bloodied quickly, an additional way you are going to get uptime with Heating Up and On The Edge is with explosives. Sparkcrackers is the way to do it for much of the game, since it will afflict you with Distraction and with a high explosives skill can last ~30s on a hit. The catch is that it needs to hit deflection first, so it will not work very well if you have already buffed your deflection or are in the middle of an Escape. A smaller catch is that your Intellect and possibly your Resolve are high, so your will defense will be high, making it hard for Sparkcrackers to hit, so only do this if you're desperate for a buff or are under the effects of Deadeye, Potion of Deftness, or something like Potion of Perfect Aim (all of which will give you a modest boost to accuracy). Do note that with a modest intellect, Sparkcrackers will attempt to hit you twice: one upon contact and once again a second later; its distraction effect triggers every second and Sparkcrackers actually has a base aoe duration of 1s.(*)
(*) Note that patch 1.2 significantly weakened the effects of Deadeye and Potion of Deftness (no additional accuracy from alchemy) and somewhat weakened Sparkcrackers (no extra duration scaling from intellect).
If all else fails, you can use Cinder Bombs (or rely on a friendly wizard to cast something like Chill Fog). Unlike Sparkcrackers, Cinder Bombs don't need to hit deflection first and instead of targeting will targets reflex which may not be as high if you have a perception affliction. Note that the blinded affliction is much worse for you than being distracted, because in addition to being flanked and losing 5 perception, you will also have an additional -10 accuracy penalty and a severe +50% recovery time penalty. However, even though the +50% recovery time penalty has the same magnitude as Heating Up's -50% recovery time bonus, the recovery time bonus is much more powerful than an equivalent magnitude penalty, and so you will still gain a significant speed up from being blinded(*). Cinder Bombs can also be used suicidally in a pinch if you want to lose some health to either trigger On The Edge or get into Heating Up in the first place, but do pay attention to that ongoing damage because it would be stupid if you ended up actually killing yourself. In higher-level fights, Cinder Bombs can also be a useful protection since if you're blinded you cannot be hit by Fampyr's Dominating or Charm Gaze. (Theoretically, blinding enemies also blocks them from using gaze attacks, but as of 1.1.1 this is bugged and doesn't work.)
(*) This point comes up again and seems confusing, and is a minor disagreement I have with MaxQuest's otherwise excellent work on action speed; in the pinned post he asserts that all maluses (through something he calls "double-inversion") are stronger than their equivalent magnitude bonuses. This may be true for damage, but is not quite the right way to think about this in terms of action/recovery because action/recovery has different "native units" depending on whether it's a bonus or a penalty. The true complexity is left for the appendix, but for here it serves us just to remind you that because of the different "native units" a -50% recovery time bonus is not countered by a +50% recovery time penalty. Instead, if you are at all familiar with investing or finance, it's related to why if your investment loses 10% of its value one year, you actually need more than a 10% gain the following year to make up for it, because you're starting from a smaller base.
In fact, let's stick with this investment analogy and swap in numbers from Deadfire here: imagine you had $1000 worth of stocks that lost 50% of its value one year. What return would you need the next year to make it back? If you lost 50% of your $1000, you are down to $500. So you would actually need a 100% return (doubling your money) the next year to get your money back. And in fact, this holds true for the blinded/Heating Up interaction. For the blinded affliction to cancel out the -50% Streetfighter recovery time bonus, it would need to be a +100% recovery time penalty. Sure enough, you can verify this in-game by looking at your weapon recovery time and then blinding yourself. Even though you have a +50% time penalty from being blinded your recovery time will go significantly down. To get the magnitude of the specific effect, we have to convert to our native units; -50% recovery time bonus is natively a +100% action speed during your recovery; a +50% recovery time penalty is already in its correct native unit; we take the +100% and subtract the +50% recovery time penalty; the answer is positive, so we treat it as an action speed adjustment of +50%, or a net recovery time bonus of -33% which is still significant. Yes, this math is weird.
Other than those explosives, you should load up on whatever else floats your boat. Remember that, like weapons, explosives have a short action time and a longer recovery, so the Streetfighter with their special will be able to spam explosives like nobody's business. because of this Grenade and Concussion Bombs are a little less useful than others because spamming bombs is a little harder to do when Grenade and Concussion bombs are knocking everyone around.
For alchemical uses, Deadeye and Potions of Deftness are mostly there as accuracy bonuses (though the action speed bonus from the potions of deftness is a nice plus), not just for landing Sparkcrackers but also because even with 17 perception there's a high likelihood that you have a perception affliction, so you are going to find that sometimes you need the accuracy help. Note that once you get Devotions for the Faithful, you can just use that for a powerful if slower accuracy boost (though it may be too slow for just trying to trigger Sparkcrackers early on since you'll have to go through at least one slow spell recovery from Devotions before attempting a Sparkcrackers).(*)
(*) As of patch 1.2, Devotions for the Faithful's +10 accuracy bonus is the best active common accuracy bonus that I can immediately think of. This is because patch 1.2 significantly nerfed all consumables but in particular Deadeye and Potion of Deftness by no longer letting their accuracy bonuses scale upwards with your alchemy skill, so Devotions's +10 bonus got significantly better as a result. In addition, Devotions is special because most other accuracy bonuses come from being perception inspirations (like Fighter's Disciplined Barrage), so the +10 accuracy will stack freely with other perception inspirations and importantly for this build won't counter perception afflictions.
Potions for the Final Stand are a good, if uncommon, supplement to Barring Death's Door (and your main option--aside from a friendly Shieldbearer Paladin--before you get Barring Death's Door). Potion of Impediment meshes extremely well with the Streetfighter's ultra-low-recovery rate for weapon attacks; as of 1.2 you no longer get effect scaling with alchemy but with a 30% interrupt chance and a fast attack rate you can still prevent a dangerous enemy from getting much done (interrupts, in addition to countering any active ability adds a little extra time to their current recovery). And lastly, a major weakness is that the Streetfighter/Wael lacks a lot of good ways to penetrate enemy armor (you won't be picking up any weapon modals that give you bonus penetration, and there are no armor penetration skills), so when you can, keep a Potion of Piercing Strikes ready to give you bonus penetration for big, hard fights.(*) For much of the game, you'll also want Potions of Minor Healing (nothing stronger) just to help you when your health gets too low.
(*) Due to the fact that damage maluses are significantly stronger than damage bonuses due to double-inversion, under-penetration can lead to severe loss of damage. 25% underpenetration cancels out upwards of +300% worth of damage bonuses (it's a similar situation to a -75% recovery time bonus requiring an equivalent +300% recovery time penalty to cancel out), which is well more than what even a Streetfighter that is On the Edge and critically sneak attacking can put out. Even just going up to a 50% underpenetration (which only cancels out +100% worth of damage bonuses) can actually more than double your damage output. E.G. a neutral might, level 20 streetfighter/wael on the edge, sneak attacking, critically hitting, at 25% underpenetration would have a net damage adjustment of almost -40% (-3 + .6 + .5 + 1 + .25 = -.65 => 1/1.65 = .61 => -39%) whereas just managing to go up to even 50% underpenetration would have a net damage adjustment of +35% (-1 + .6 + .5 + 1 + .25 = 1.35 => +35%) which is a 2.25x increase in damage. Shows you how important penetration can be.
DETAILS - MISCELLANEOUS SPELLS AND ABILITIES
As mentioned before, we pick up Pillar of Faith, Pillar of Holy Fire, and Cleansing Flame just to do some extra damage and cause some interrupts (for Pillar of Faith), and hopefully trigger a Marux Amanth double-cast. Remember that opportunity cost is very important; make sure with all of these that you are doing something that, on its own, is worth giving up the melee damage you are foregoing by not auto-attacking instead. For Pillar of Faith, that means trying to interrupt several enemies in one go (remember that the prone itself targets Fortitude whereas the damage targets Reflex). Pillar of Holy Fire against groups of enemies (and potentially yourself if you're trying to trigger Bloodied). Cleansing Flame to eliminate enemy buffs and when it looks likely you'll get some good jumps off of it.
Holy Meditation is just for some early resolve inspiration if needed, and later on as a way to cancel out a growing number of ways that enemies can Terrify your party. This character doesn't have many offensive abilities, so Frightened is not actually that big a deal, though the resolve hit can be a liability.
Searing Seal is there as an additional source of blind and a way to lure enemies and blind them for free. That's because if you cast a spell outside of combat and you stay out of combat for a few seconds, you get that spent resource back. For many abilities this is not too useful. Seals, however, have a long duration. So you can cast it, wait for your resource to come back, and either lure enemies into it or wait for them to walk over and be hit by it. In fact, so long as you cast it a decent distance from enemies away, but still close enough to alert them by its noise, they will walk over the seal and be affected just after you get your 5th level spell cast restored. Because your 5th power level of spells is going to be oversubscribed (Barring Death's Door is just so good in this build), an opportunity for a free blind is too good to ignore.
I no longer recommend Searing Seal and instead recommend Champion's Boon. Champion's Boon solves several problems for us. First, it gives us +2 penetration from its Tenacious inspiration; this won't stack with a sabre or stiletto weapon modal, but will stack with the mace effect (because the mace gets extra penetration by implementing it as a debuff on the enemy, not as a buff on yourself) and importantly is the only consistent way we have to boost the penetration of daggers. In fact, for daggers this is better than a weapon modal because most other bonus penetration weapon modals give you a significant recovery penalty, whereas Tenacious gives us the +2 penetration without any drawback.
The second and smaller perk is that Champion's Boon gives us +3 engagement. Combined with Persistent Distraction, this means up to four nearby enemies will be Distracted. Combined with the inherent Resolute inspiration Champion's Boon gives you, this is a net 10-point hit chance swing in your favor, in a way that stacks with accuracy penalties from Despondent Blows or Devotions of the Faithful or the hatchet weapon modal. It also means your Riposte attacks have an easier chance to hit on virtually anyone attacking you.
The downside to this choice is that we no longer have a "free" spell like we did with Searing Seal. Because Barring Death's Door is so important for survival, you really should evaluate your combat situation carefully and weigh whether or not you are going to need the bonus penetration or you're going to need to prevent death/trigger On The Edge.
We pick up Smoke Cloud solely as a prerequisite for Smoke Grenade. And we pick up Smoke Grenade for two reasons: as an additional fast interrupt, and as a way to help enable Deathblows. Persistent Distraction will take care of Deathblows for you automatically (the free flanked that perception afflictions bestow counts as a second affliction), but Smoke Grenade will be useful for triggering Deathblows on enemies you are not able to engage and for situations where the enemy is resistant to perception afflictions (you get the weakened effect from Smoke Grenade, and then rely on either manual flanking or some other explosive or party member to apply another affliction). If you have no problem with keeping near 100% Deathblows up-time without Smoke Grenade's help and don't mind the friendly fire, you can pick up Pernicious Cloud instead for the extra damage.
I added a story choice to the build--to be snarky at the gods when they first summon you after leaving Port Maje. This is because this unlocks the Wit of Death's Herald upgrade to your Death's Herald watcher ability (it also fits in with Wael's preferred disposition of being clever, coincidentally). Wit of Death's Herald adds on an intellect inspiration to the base effect, so is a "free" way to cast Prayer for the Spirit once/rest without using up a PL3 slot.
DETAILS - STATS
The last part of this is just the easiest: why are stats the way they are? Intellect is the only mandatory max-out, because you want to squeeze as much duration and area of effect out of everything you have. Resolve/deflection is an important stat, but after the consumable nerfs of patch 1.2, this build really needs the accuracy help due to perpetually being hit with a perception affliction (and no longer able to rely on Deadeye or Potions of Deftness with a high alchemy), so we actually skip resolve and invest in perception instead. We'll eventually get Shorewalker Sandals for its +1 Resolve to get up to 10, but we don't actually want more than that, because patch 1.2 also weakened Sparkcrackers' duration, so having too high of a resolve will negate many Explosives skill points. Basically we'll rely on deflection-boosting buffs, enemy-accuracy-penalizing debuffs, and Holy Meditation for our survivability instead of requiring a high inherent resolve. If things are too rough for you early game, you can put on a Cloak of +7 Deflection or some resolve boosting gear or some such, but as you get more tools you will generally want to dump some of that stuff just so that it can be easier to self-Sparkcrackers when you need to. (By contrast, we like Entonia Signet Ring because its defense bonus only occurs while you are actively engaged by enemies and you frequently are going to be doing Sparkcrackers while you aren't actively being engaged yet. Perfect for this build.)
A lower resolve means we're going to get hit more regardless of what we do, so we don't dump constitution anymore. We don't have the spare points to invest in it, but we also don't want to make ourselves squisher than we need to be. In fact we pick up Tough later on for extra buffer room against high-impact enemy spells/abilities, but it's really a double-edged sword. More health means you can stay at Bloodied or below for longer, and there's more room for mistakes (like not paying attention to the fact that your Arcane Veil just ran out or that your Arcane Veil isn't doing anything because the enemy is using guns even while you melee them), but it also means it takes a lot longer to get to Bloodied.
Whatever might adjustments you make here will be generally dwarfed by the huge amount of damage bonuses you'll get from a Streetfighter sneak attack (the base +30% plus power level scaling, plus the +50% from the Heating Up special). We opt for a neutral score to put the points elsewhere (and we want to avoid a penalty because damage penalties are stronger than bonuses).
We put points into dexterity because action speed stuff has linear returns (see appendix) so we'll be able to be even faster, and importantly our spell cast times aren't affected by the Streetfighter special (just their recovery), so dexterity is the main way we'll make our spell casts a little more nimbler. (Picking up Rapid Casting later on will help, too.)
Wood elf background is pretty important for its Dexterity affliction resistance. This character really wants mobility, and being able to shrug off hobbling, and convert a near-lethal Paralyze into a not-bad-at-all Immobilize is pretty valuable. You can find gear that provides Dexterity resistance, but this way you can reserve those inventory slots for more interesting stuff.
CONCLUSION/PUTTING IT ALL TOGETHER
Whew! That was a lot of text to read. If you made it this far, congratulations! And I hope you learned a thing or two. If this build still sounds really abstract to you, here are a couple typical examples of mid-to-high level encounters and how an Umezawa would handle it.
SCENARIO: mixed group of melee
0. already affected by deadeye.
1. main tank charges in, unstealthed, while Umezawa is still stealthed. Targets self with a Sparkcrackers with ~70% accuracy (which is almost a guarantee to graze/hit, and the will effect will be likely to at least graze), if the main tank pulled right maybe an enemy gets affected too.
2. Escape towards the bulk of the enemies. While you still have the +50 Escape bonus active, make sure hatchet modal is enabled (in case Riposte procs you can debuff enemy accuracy), and then self-buff with Prayer for the Spirit and then cast Despondent Blows before disabling your hatchet modal and then casting Arcane Veil.
3. DPS everything down. Throw some bombs if you feel like it. Re-cast Arcane Veil if necessary. If fight is still going on and you're out of Arcane Veil, self-empower, chain Escapes together, use Mirror Image, etc.
SCENARIO: high-level mixed group of melee and casters
1. Charge in. Notice that casters start casting Arcane Dampener.
2. Use Smoke Veil to go invisible. Continue to run Umezawa towards the back while the Arcane Dampener gets retargeted to your other party members.
3. Unstealth by throwing Sparkcrackers at point-blank range with one of the back casters, hopefully affecting yourself with Distraction. Cast Prayer For the Spirit, then Barring Death's Door as enemies reconverge on you.
4. Let them bring you down to 1 health and flank you while you DPS the caster down (you can also self-target with Pillar of Holy Fire to help bring your health down), and Escape to target the next caster. If you see another Arcane Dampener start getting fired, use Smoke Veil or Smoke Grenade.
5. When all the casters are gone, DPS everything else down. Use Salvation of Time for extra Barring Death's Door time. If you're in a situation where Riposte could be relevant and you're already at 1 health, go ahead and use Arcane Veil just for the added damage from your counterattacks. If you run out of Salvations of Time and Barring Death's Door, self-empower for another round of both.
SCENARIO: "deal with it" nemnok fight.
1. Let main tank unstealth and trigger everything.
2. While stealthed, Umezawa either Powder Burns or self-Sparkcrackers.
3. Carpet bomb the area with Lightning Bombs and whatever else you got (Immolator, Frost Bombs, Blister Bomb) and mix in a couple of Smoke Grenades to lower enemy fort/reflex and reduce their ability to heal back from this hellscape you're creating.
4. Watch the entire enemy fight disappear under a hail of bombs exploding every other second.
SCENARIO: single-enemy tough fight
0. Start off with deadeye already enabled.
1. Start off with your melee/blunderbuss weapon slot, the melee weapon being a fast (base 3s recovery) weapon. If you are able to land a non-graze self-sparkcrackers, switch to your dual-wield. Otherwise stay with this weapon slot and just find ways to keep Powder Burns uptime for the rest of this scenario.
2. Drink Potion of Impediment. (The interrupt chance from this and Deadeye are multiplicative with each other, so you'll have a net 40% [1 - .7 * .85] chance of interrupting with any given attack.)
3. Attack the enemy with a fast weapon. While as of 1.2 you won't completely interrupt-lock the enemy, attacking almost every second with a 40% chance to interrupt will turn down the danger level of any given tough enemy a lot.
4. Use Salvation of Time if necessary. Use Cinder Bomb, Pillar of Holy Fire, or other explosive to get your health down if you also want to be On The Edge while doing this.
After having dug up how monastic unarmed training works, I suggest two alternatives that rely on picking up Monastic Unarmed Training instead of Fast Runner. (You can pick up Fast Runner later in lieu of e.g. Holy Meditation or Pillar of Faith). TL;DR: Monastic Unarmed Training gives you potentially extremely fast, high-penetration weapons, but you need to find bonus sources of PL to really get mileage out of it (ideally at least +3, hopefully even +6).
Option 1: Nature Godlike
Until you pick up Champion's Boon you are going to a friendly party member who can buff you with a body inspiration, but you get +1 PL, and then you can get up to another +2 PL from either food (+1 PL) or a Potion of Ascension (+2 PL). +2 PL is enough to get you better-than-superb fists pretty early on, though not quite enough to get better-than-legendary, so you'd need to start crafting/buying Potions of Ascension on a regular basis to get to better-than-legendary scaling (at level 19). This is the "consistent" option, but is dependent on party composition.
Option 2: Death Godlike
For much of the game you won't get much bonus PL, but once you unlock Barring Death's Door, you can hover around Near Death for +3 PL, with the possibility of another +2 PL form either food (+1 PL) or a Potion of Ascension (+2 PL). +3 PL is enough to get you better-than-legendary fists by the end of the game (level 19), and conditionally +5 PL will get that for you by level 13. Unfortunately I don't think there exists another way to stack on another +1 PL to get better-than-mythic fists by the end of the game, at least without exploiting a bug(*). If there does exist another source, do let me know (it would have to be a non-class/keyword-specific PL bonus that comes from an item). This is the "spikey" option, but doesn't require you to have a body-inspiration-buffing-capable party member.
(*) The Heart-Chime Amulet (reward for Pallegina's quest) is supposed to give you a variable bonus as a godlike or watcher, but is currently bugged to always give you the bonus that a Moon Godlike should get, which is a +1 PL stackable bonus. You can use it here to get +6 PL, but be warned that Obsidian will fix this bug and it's only a matter of time before this loophole goes away.
In either option your main dual-wielding set of weapons will be your fists, which you can treat as fast blunt weapons with bonus inherent accuracy, damage, and penetration and a +30% damage lash. The weapon modal proficiency for fists is an additional +2 penetration, so you could theoretically re-drop Champion's Boon and bring back Searing Seal. You'll have a strong early game due to how the monastic unarmed training talent works, and the bonus PL will help ensure that you continue to scale at a reasonable pace. Though you need to be a Death Godlike comfortable with dancing with Barring Death's Door to really take advantage of it; fortunately for this build lots of good things happen at near death, so you'd be really leaning into that "glass cannon" philosophy. You're giving up a head slot, but this build was mostly using it for Fair Favor, which you don't need if you're busy punching everything in the face.
One immediate alternative approach to an Umezawa build you might think of is to pair a Streetfighter with an Illusionist, or at least a Wizard. You get a lot more opportunity-cost-worth-it damage spells and ways to afflict enemies. You also get Infuse With Vital Essence with is like a super-charged version of Prayer for the Spirit (since you are mostly using it as a self-buff in this build anyway). Plus, the Wizard has access to Wizard's Double, which, with a sufficiently high deflection, its duration-less unconditional +40 deflection is the best defensive spell you can use.
The problem is that the Wizard actually has its defensive spells inefficiently distributed for our purposes. Both Arcane Veil and Mirror Image are at PL2, which means a Priest of Wael actually effectively gets twice as many casts as the wizard, since for the priest Arcane Veil is at PL1 and Mirror Image is at PL3. The fact that Arcane Veil is a PL1 is also a bigger deal than you may think. For one, early-game (Port Maje) 1.1 Path of the Damned is a fair challenge, and a wizard multiclass won't get two casts of PL2 spells until level 7, which is after when you probably most desperately need it. In addition, there are several resting effects in the game that give you a bonus +1 level spell cast (The Wild Mare and the Luminous Adra Bathhouse immediately come to mind). This is great for a Priest of Wael because it potentially means 3 casts of Arcane Veil, plus an additional 2 upon a self-empower. There's basically no equivalent for a wizard. And while Wizard's Double is good, it won't be that good for much of the early game because your deflection just won't be high enough to really milk it for what it's worth, and by the time you can take advantage of it, the Wael version will be picking up new tricks.
Speaking of which, wizards lack the following spells that really help tie the Umezawa together: Despondent Blows, Devotions for the Faithful, Barring Death's Door, and Salvation of Time. It makes for some great general party utility and combat versatility. While I'm sure there is a great Streetfighter/Wizard build out there, I am fairly confident that for this specific playstyle (as opposed to an immortal high-deflection riposte build, which a wizard could do better) Streetfighter/Wael is the way to go.
The Umezawa build as of now is also not a solo PotD build. A soloable version of Umezawa would be closer to an immortal, high-deflection riposte build, because the way this build is now the Umezawa is a team player. Frankly, without a lot of delicate pulling of enemies or just constantly equipping a large shield, there's just no way that the Umezawa can sustain all the enemy hate in the world because while your deflection is high to mitigate a significant portion of damage, it's not going to be high enough. Even with access to Barring Death's Door, encounters would probably just take too long and you'll be left out of steam with enemies still standing. Umezawa can work great in smaller-than-5 parties, but a completely soloable build would likely be unrecognizable to the one being presented here today.
NOTES FOR MAGRAN'S FIRES
Abydon: not recommended. This build leans hard on certain unique weapons and armor and you can really run up an expensive repair bill versus other characters that can use generics instead.
Berath: nothing relevant to worry about.
Eothas: nothing relevant to worry about.
Galawain: be on the look out for Unstoppable and Bullish enemies. Unstoppable can't be afflicted, but they can be flanked; important to keep in mind for keeping your DPS up. Bullish enemies interrupt and knock back at will, so make sure you don't drop your best deflection bonuses (Arcane Veil or Escape) before trying to cast something at point-blank range.
Magran: if you can pull off this build along with a party on this challenge you should probably quit your job and become a professional DOTA or Starcraft 2 player.
Skaen: if you want to use this character to help illuminate things, the sabre proficiency is more important early on, and then you can use the sabre-torch. If you don't have access to the sabre-torch, well... *shrug*. Normal torches are still usable but represent a huge DPS loss.
APPENDIX: LINEAR RETURNS
I made the assertion earlier that speed adjustments offers linear returns. I've fought similar debates re: World of Warcraft and Diablo 3, and I'll fight it again now. But before we go onto my analysis and conclusion, we need to be clear what we mean by "linear returns."
If you've taken calculus, then the easiest way to express what "linear returns" means is that for a given differentiable function f where f'' is the second-derivative and f consumes a stat x to yield a metric y, then if f(x) = y, ∃x₀: ∀x > x₀ f''(x) = 0. Putting it into words, we mean that after a certain point for x the second-derivative of f(x) is 0. Analogously, "increasing returns" is when f''(x) > 0, and "diminishing returns" is when f''(x) < 0.
Put in less math-y speak, a stat has "linear returns" when for a given absolute change in that stat, the resultant metric always yields the same absolute change as well, regardless of what our starting point was. By contrast a stat has "increasing returns" when for a given absolute change in the stat, the resultant metric yields continually larger absolute changes the higher our initial stat was. Similarly a stat has "diminishing returns" when for a given absolute change in the state, the resultant metric yields continually smaller absolute changes the higher our initial stat was.
A key point to this is properly identifying what the "metric" is. The poster child for this is resolve(deflection) and perception(accuracy). I've talked to and read posts by people who assume that because +1 deflection gives the enemy a -1 penalty on the attack roll and a +1 accuracy gives you a +1 on the attack roll that deflection and accuracy have linear returns. Nothing could be further from the truth. Because the actual metric is "effective health." That is, how much effective health do you have? And as this is the true metric it becomes very clear that deflection/resolve has increasing returns, because at very low deflection another point hardly matters, but at the top-end each point of deflection is so important that eventually it gives you infinite survivability. Accuracy, by contrast, has diminishing returns, because it's the flip side to that deflection equation. When you go from missing all the time to grazing some of the time, you just got an infinite increase in your damage potential (which was previously zero). However, when you already have accuracy so high that you are critting all the time, another point of accuracy will do literally nothing.
With regards to action speed and recovery times, a lot of people get hung up on the fact that when it comes to speed adjustments, the more bonuses you have the smaller your reduction in your action time and recovery time, and they therefore conclude that therefore the returns are diminishing. This is wrong for two interrelated reasons. One, the action time/recovery time is not actually the metric. It is in fact just a mere component of our true metric, and is in fact the denominator. Two, because it is the denominator for our true metric, the smaller our starting value, the smaller the change needs to be to accomplish the same net effect. For example, reducing your recovery time by .1s when you're starting at 1.5s is way better than reducing your recovery time by .1s when you're starting at 5s. So just the mere fact that you get less recovery time reduction the more action speed you already have does not, by itself, mean you have diminishing returns. Instead, you have to look at the true metric.
So what is the true metric? Basically, damage per second, or damage over time. More generally, it is "how many things can we accomplish in a given amount of time?" Now, does this metric get linear returns from speed bonuses? Or is it diminishing? Or is it increasing?
Let's imagine a hypothetical scenario where you attack and on average do 15 damage to the target, and you do so every 5 seconds. Using the action speed equation, we can draw a graph of how much damage you are capable of doing for different levels of +action speed. See that smooth diagonal line going up(*) in the graph below (click to enlargen)? That's literally the definition of linear returns.
graph.jpg 36.18KB 3 downloads
(*) Note that for a game like Diablo 2 or Diablo 3, it's not quite this simple. At a certain point, because attack animations are connected to whole numbers of frames (frames as in "frames per second"; whole numbers as in no fractional frames) you start running into a "breakpoints" where you actually get 0 returns for a while until you get enough accumulated attack speed improvements to "round" down to the next lower number of frames per attack. So at a certain point it stops being a smooth diagonal line and starts becoming a steadily embiggening staircase. It averages out to be linear returns, but in truth it no longer really is any sort of well-defined returns because the function is no longer differentiable at that point, which was an important part of defining any kind of returns above. It is possible that at the extreme, Deadfire hits similar issues, but in truth it is clear that Deadfire retains recovery time to two decimal places (even if it rounds to one-decimal place in tooltips), and it is likely that it is impossible to accumulate so much action speed so as to hit "breakpoint" issues. Plus, if Deadfire allows for fractional attack frames, then breakpoints are a non-issue altogether.
Now one hiccup here is the fact that the way Deadfire treats adjustments to your action time and recovery time is that it has different native units of measurement depending on whether it's a bonus or a penalty (malus). In effect, the native stat for bonuses is action speed. The native stat for maluses is action time. So when you have a -50% recovery time bonus, it is actually truly a +100% action speed adjustment. However, if you have a -20% action speed penalty, it is actually truly a +25% action time adjustment. Why is this relevant? Because it affects how Deadfire combines the numbers behind the scenes.
When you're combining bonuses, you translate anything that's not an action speed into an action speed adjustment and then just add them up; you then optionally reconvert it into whatever unit (recovery time adjustment or action speed adjustment) the tool-tip requires; e.g. a -50% recovery time penalty and a +15% action speed bonus becomes a +100% action speed adjustment and a +15% action speed adjustment which becomes a +115% action speed adjustment for the recovery, which turns into a -53.5% recovery time bonus (and just a -13% action time bonus). When you're combining maluses, you translate anything that's not a recovery time penalty into a recovery time adjustment and then just add them up e.g. two -20% action speed adjustments become two +25% recovery time penalties that add to become a +50% recovery time penalty, which you can then reconvert back for display purposes into a -33% action speed adjustment if needed. When you're combining bonuses and maluses, you convert all bonuses into positive action speed adjustments and sum them, and then convert all maluses into negative recovery time adjustments and sum them, and then you subtract the latter from the former even though they are two completely different units of measure; so a -50% recovery time bonus and a -20% action speed adjustment becomes a +100% action speed adjustment minus a +25% recovery time penalty. The resulting number's unit depends on its sign. If it's positive, the resulting answer is determined to be an action speed adjustment. If it's negative, the resulting answer is determined to be a sign-flipped recovery time penalty. This is needless to say weird.
Anyway, this is to say that yes, penalties can drag your numbers down because a -20% action speed is actually much more powerful than a +20% action speed. But unlike what MaxQuest says in his otherwise really useful action speed post, it's not just maluses that do this because of double-inversion (unlike damage penalties). For similar reasons, a -50% recovery time bonus is much more powerful than a +50% recovery time penalty. It's because you have to first convert to the "native units" at which point you see that a -20% action speed is actually a +25% recovery time penalty, which is more powerful than the +20% action speed; and the -50% recovery time bonus is actually a +100% action speed bonus which has a larger magnitude than the +50% recovery time penalty. Yes, again, this is weird and confusing.
This odd use of native units of measurement also has an effect on the full understanding of our linear returns. When all you have are action speed bonuses, the linear returns are easy to see as a diagonal graph going up and to the right. When all you have are recovery time penalties, it is trivial to see that e.g. each +10% recovery penalty you add to your 5s recovery is a flat .5 second, which is also linear. But when you are combining bonuses and penalties, things get a little harder to compute and things don't add up as trivially. But if you remember about "native units" then it looks a little less weird that a -75% recovery time bonus can cancel out as much as a +300% recovery time penalty (which can make it intuitively feel like there's increasing returns to time/speed modifiers) or that a -20% action speed penalty outweighs a +20% action speed bonus (which can make it intuitively feel like there's diminishing returns to time/speed modifiers). And this is why it's so important that you get x and y right when analyzing whether for an f(x) = y, that f''(x) = 0 .
Now, astute observers will note that with linear returns, the more you have of something, the less any further gain is worth relative to what you already have. MaxQuest has called this "intrinsic diminishing returns" for lack of a better term. I don't quite like the terminology because "diminishing returns" has a specific meaning, but I also don't have a better suggestion.
Anyway, it is certainly true that if you have +500% action speed, another net +25% action speed is going to be a relatively less impact than when you had +0% action speed. This point is relevant if you're trying to decide between investing in damage or investing in speed, because to maximize total damage over time you want to balance out your bonuses as much as possible (for the same reason that given a rectangle with sides a and b with a fixed combined length of a + b = c, to maximize the area, a = b. In other words you get more area from being a square than from being an extremely skinny rectangle). But if you're not actively concerned about a tradeoff between damage or speed (such as when choosing how to allocate points between might or dexterity) the fact that the relative gain is less and less is completely irrelevant to whether or not you get linear returns, and sort of an orthogonal point altogether.
Why? In general, a good way to think about increasing, linear, and diminishing returns is this: if something has increasing returns you generally want to invest a lot in it; if something has diminishing returns a little bit of investment might be worth it but it's definitely not worth it after a certain point(*); if something has linear returns, you're always going to get value out of it. If instead of this being Deadfire and us talking about speed bonuses, we were instead talking about Baldur's Gate and equipment that gives you bonus damage to weapons (like a pair of gloves that might give you +1 to weapon damage rolls), literally no one would be saying "oh I guess you have +5 total damage bonus now. Not worth getting more" unless it was some sort of tradeoff between equipment that gave you +1 damage or one that gave you +1 extra attack. In fact, everyone would probably be trying to stack on as much weapon damage bonus as possible after maximizing their attacks per round. It's the same thing here. When you're not trading off for damage, there's basically no reason to not get more speed if you can.
(*) This is very simplified and glosses over a lot of nuance. Something can have such extremely slow increasing returns that it may never be worth investing in, and something can have such extremely slow diminishing returns that over a reasonable range of stats it may never not be worth investing in. Similarly, if every action speed and recovery time adjustment in Deadfire were suddenly decimated to literally one tenth of their current value, even though speed adjustment would still be linear returns, the gains would be so small that even though they would still be linear returns, you would be getting linear crap returns. The linear returns for investment in speed works out for us players in Deadfire because in large part Obsidian purposefully balanced speed with damage (see the Might/Dexterity correspondence).
So that's it. What are you still doing here? Show's over! Hope you found this useful, entertaining, or at least educational!
Edited by thelee, 10 October 2018 - 12:54 PM.