Dr. Hieronymous Alloy

They really need to replace intimidate with explosives on the dissident.

One thing that could help somewhat: If we returned to a 50% bonus for critical hits, as it was in the first game, rather than a 25% bonus, that would increase the per-point value of perception up to 2.4% rather than 2.0% flat. Which doesn't solve the whole problem but does bring things in closer parity -- it's only 0.3% below the value of a point of Dex at that point. It might even actually completely solve the issue depending on how the falloff values for each change as you add additional points of per/dex. In a way, Per has been nerfed twice: once by the removal of interrupt, once by the shift to a 25% critical bonus. So I think that's my recommendation for this problem: return to a 50% bonus for critical hits, rather than the current 25%.

Sure, that's a fair comparison. But druid spiritshift forms. . .are instant-cast, right? (I haven't bothered to try a druid in Deadfire yet). And they give really big bonuses, in exchange for short durations? If they gave milder bonuses, they'd need longer durations . . .

A summoned weapon really isn't a spell, though. It doesn't target anything, it doesn't *do* anything, until you attack with it like a weapon. Functionally it's equivalent to an extra weapon set, not a spell or a buff. Think about it this way: something like Arcane Shield operates as a buff on top of your currently equipped gear. It doesn't replace your currently equipped shield. If there were a spell that summoned an equippable shield, yeah, the spell should either have an instant cast and be level equivalent -- with some sort of nice bonus effect to justify the cast -- or if long-cast / short duration, be markedly better than other level-equivalent gear.

Historically speaking, in the time period that this game is roughly set in, you wouldn't try to reload a pistol in combat -- they were considered one-shot deals, you'd fire them first then charge in with a sword afterwards. Period accounts of naval combat especially are full of "discharged my pistol, threw it aside, and charged with my sword" type stuff. If you were going to be reloading, you'd be in formation with muskets and bayonets. Thing is saying you couldn't reload guns and crossbows in melee would be a balance problem.

I'm assuming those peasants died in vain.

there is many a cautionary tale 'bout the granting o' wishes. genie in the lamp. monkey's paw. etc. grant wish and make instacast, but then reduce extreme long duration to put more in line with other spells? I don't really follow what you're saying here re: "wishing." All the spells need numerous balance changes to durations, cast times, effects, etc. right now., and yeah, there's an obvious bug with the lashes boosting weapon damages and I don't think anyone's saying that should persist. That said, i'm not sure why the duration of the summoned weapon particularly matters. "Long enough to last the whole fight" seems reasonable if it's otherwise equivalent to an appropriate-for-level found weapon. If it's better, then maybe shorter duration makes sense. *Regardless* though it should be an instant-cast because . . . it's a melee weapon, you have to cast it then attack with it. If it's roughly equivalent with other found items (as I *think* you're arguing for) then a fast cast time and a very long duration would both be indicated, in order to maintain that parity. With longer cast times than "instant" or shorter durations than "the whole fight," it's no longer equivalent to a found item, so you'd need to give the weapon a comparative boost in effectiveness to balance that out.

I agree on points one and two, even though I always found interrupt rather underwhelming in PoE1 I don't really see how the third one is an argument against perception. So cc reliant characters shouldn't try to buff their accuracy at all? I am very happy for any bonus that I can still get. -10% chance to miss for +10% chance to crit equals an expected increase of 24% in cc duration Fair enough. It's basically a utility argument, not a math argument: i.e., "each point of strength helps you jump 1% higher" vs "with a strength value of X, you can jump over this wall." If you can make it over the "always hit" threshold by stacking accuracy that makes CC more useful generally, if you can't, Per/Acc still helps, just . . . not in the same way.

Yeah, I agree on this also.

Sure, but even in those instances, I'd argue that those are the fights where you're supposed to be using debuffs to lower enemy defenses, or buffs to heighten your own. That's why I said "gameplay error" above -- going head-on against a thirty-point defense advantage is the equivalent of trying to rush Raedric's Keep instead of sneaking in through the sewers. The game allows you to do it, but it's . . . not intended to be the optimal approach. And again, once you start using buffs and debuffs, the relative value of perception drops a lot, faster than the relative value of additional points of Might or Dex. Sure, but. . .hrm. There were a few advantages to stacking Per / Accuracy in the first game. The first was the one we've mostly been discussing -- the statistical boost to hits, etc. That has diminishing returns etc. as we're discussing, and you're right that the relative value of each point of accuracy is mathematically greater now that there are fewer sources for it. Second was that stacking Interrupt could be really useful. Third was that once you stacked a LOT of accuracy, you could rely on your CC effects; get your accuracy high enough and they would always land and frequently crit, which makes them a lot more useful utility wise. In this game, we still have the first advantage; the second has been removed. The third one is what I'm talking about re: the inability to stack accuracy -- it appears you can't get your accuracy high enough to guarantee or near-guarantee your spells will always land. Like, even with max starting Per, there's still a 15% chance you're gonna whiff against an equivalent-defense enemy, and you only have a 10% chance to crit.

you could theoretically make it a special Weapon Style, just for (one handed melee weapon) + pistol/blundy

Can we all at least agree that the summoned weapons should be instant-cast? No longer than it takes to do a weapon switch?

Yeah, Perception helps a lot against very high defense targets, but why are you facing very high defense targets? Fighters (and I think everyone else? but I checked with a fighter) only get +3 Accuracy per level. Base accuracy for everyone is 20 now; max accuracy for fighters is then 80 at level 20. The accuracy of 50 in your example is equivalent to the standard accuracy for a level ten character; for that same character to reach an 80 accuracy, equivalent to the defense in your example, he or she would have to be level twenty, max level. So yeah, Perception does help a lot if you're trying to face stuff waaaaay over your level, but if you try that everything else about trying that will make sure you get crushed, so there's little point in it. I mean yeah you've got a point that most of this argument is based on numbers theorycrafting perception vs. equivalent defenses, but there is a reason I've been making that assumption --- this game has level scaling. Most of the time, you're going to be facing enemies at rough level parity, not way higher or lower level. (It's also true that enemy defenses vary a lot -- but if an enemy has one defense thirty points higher than "level norm," they'll have some other more vulnerable defense that's lower that you're "supposed" to be attacking instead. Trying to hit defense 80 with an accuracy of 50 is a gameplay error). As to PotD, you're right that Per is marginally more useful on PotD, but PotD is only a fifteen point increase in enemy defenses, so additional points of Per don't make as much difference as you'd think -- I posted the math in another thread, but basically at -15 to hit, an additional point of accuracy IS relatively more valuable for generating grazes and hits, but you lose Critical Hits at the top end, so the overall benefit stays around 2% per point. (from this thread : https://forums.obsidian.net/topic/94949-should-might-stay-multiplicative-or-return-to-additive/page-4?do=findComment&comment=1960507 ) So that's . . . 2.1% damage increase for the additional point of Per on PotD, still below the benefit from Might or Dex. You're right that Accuracy was very effective in PoE 1 but it's a lot harder (by design, apparently) to get accuracy bonuses in Deadfire, so I'm not sure it's going to be possible to (for example) push your accuracy high enough to ensure you're guaranteed to hit / crit with CC attacks, etc., like you could in the first game.

I don't mind shooting the pistol in melee but reloading it seems like an issue.

Other side though: if summoned weapons *aren't* superior . . . why do they even exist? You're using an ability slot for them and you have to take time to cast them. If they *aren't* superior to whatever loot you've found, why are you bothering with them? I'd say that at minimum they should do an unusual damage type (Burning, Corrosive, etc.) and have some kind of accuracy / damage bonus at least slightly above typical loot for equivalent caster level. Otherwise there's no point in having them at all.

Good point! Yeah, one problem with "breaking out the math" is that it doesn't cover things that aren't math -- like, we can break the math out for Might vs. Per, but that will only show the damage numbers, not the non-damage stuff. Right now / next patch, Might / Resolve / Strength will give the biggest bonus -- three percent per point -- to all damage, multiplicatively. When you curve this out, each additional point of Might gives a slightly smaller relative damage boost (because 3% of base is a smaller fraction of the total with each additonal point), but roughly speaking it's a straight 3 % damage boost. Importantly, now that Might is multiplicative, though, the absolute damage boost from Might remains constant -- each additional point of Might gives you 3% additional of your base pre-might damage total. Dexterity and Perception, though, both boost all attacks / spells / actions, including crowd control, etc. Dexterity reduces action speed by 3%, which is roughly equivalent to Might's 3% bonus, except . . well, it gets really complicated, partly because you can only reduce action speed so much, and there is a minimum number of frames any action can take (and Dex doesn't reduce those five or six frames). So in practice, Dexterity gives somewhere between a 2.7% and 1.7% bonus per point, while also comparatively boosting CC etc. (see chart here: https://forums.obsidian.net/topic/86684-mechanics-the-big-attack-speed-conundrum/?p=1808225) Perception, like Dex, boosts everything -- but because of the way the attack roll is calculated and the statistical effect of an additional point of accuracy, gives at most a 2% or so damage bonus; also, it drops off faster than Might or Dex -- once you get above +25 accuracy, you're only adding .8% damage with each additional point of per / Accuracy (because you're only converting grazes to crits instead of converting misses to crits, and because crits only give a 25% damage boost, while going from a miss to a graze gives an additional 50% damage). So yeah, you're absolutely correct, the real comparison isn't Per with Might, it's Per with Dex, because those are the two stats that boost everything, not just damage. Dex gives a slightly smaller damage boost than Might, but gives a damage boost ranging from about 2.7 % to about 1.7 % per additional point .. while Per gives between 2.0% and 0.8% over similar ranges. That doesn't seem much, but it adds up, so that someone who stacks Per at character creation ends up doing roughly 10% or so less damage overall than someone who stacks Dex. If I were *really* smart, I'd figure out how to generate a graph that showed the relative bonus-per-additional-point of Might, Dex, Per, and Str/Res. [ I'm *not* that smart though: I'm pretty sure my recommended percentage adjustments above are wrong, because I calculated "graze to hit" and "hit to crit" percentages off the overall attack roll, not as percentages of the grazes on the attack roll or percentages of the hits on the attack roll. Correcting for that, for my rec's above to work, you'd need something like 6% graze-to-miss per point of resolve, and 6% hit-to-crit per point of Per, and those would quickly get silly I fear -- at 20 Per, you'd have sixty percent hit to crit! which is just getting silly. So I retract my above recs -- I've got no idea how to solve this problem! :sheepish: ] Thanks for reading all our armchair rambling!

i haven't tried the combo yet myself but it would seem logical for each to get the one-handed bonus.

That was the original proposed pistol modal. I'm presuming it got nixed for some reason

good summary. Just to make things more complicated now that you've already typed the nice summary: 1) working out the math on a critical hits / miss proposal. What would it do to the math if Perception got an additional +3% hit to crit bonus? As before, the standard equation when Accuracy and [Defense] are equivalent is 0 + (25/2) == 62.5% expected damage on a normal attack vs. equivalent defense, adjusted for accuracy / miss / graze rate With the additional +1 Accuracy that a point of Per currently gives, that shifts up to 50+ (25/2) + (1*1.25) == 63.75% with an additional point of accuracy (as one miss shifts up to a graze, graze shifts up to a hit, hit shifts up to a crit). That means that the additional point of Per/Accuracy has shifted your expected damage value per swing up by 1.25, or 2%. If you add +3% hit-to-crit (so a point of perception is giving +1 Accuracy and also +3% hit-to-crit conversion, that becomes 47 + (25/2) + (4*1.25) == 47 + 12.5 + 5 = 64.5 % expected value per swing, so a difference of 2% flat, which is .032 or 3.2% of the previous expected value. close to the values you'd expect from Might or Dex but a little higher. (Once you have accuracy bonuses over +25 though that advantage will drop off). As to Resolve: If we gave Resolve Graze-to-Miss, that would be a reduction of . . Again, if we figure each incoming attack is made by equivalent perception . .. Normal: 50 + (25/2) == 62.5% expected damage on a normal attack vs. equivalent defense, adjusted for accuracy / miss / graze rate If we adjust that by, say, 4% Graze-to-Miss, that becomes 50 + ( 21/2) == 61.5 = 2 point drop == .032 or a 3.2% drop in expected incoming damage per attack This is because Graze-to-Miss is a more powerful effect (counterintuitively) than hit-to-crit is, because a graze reduces from 50% damage to zero, while a crit increases from 100% to 125% damage. point being, giving perception +1 accuracy and +3% hit to crit is (roughly) equivalent to giving Resolve +4% graze to miss. So that's my recommendation: add +3% hit to crit for each point of Per, and +4% graze to miss for each point of Resolve. EDIT: that's all assuming that a "%" of hit to crit or graze to hit or whatever is a percentage off the attack roll, not a percentage of the crits, etc. 2) One downside to current Perception one problem with current perception: it now has the same problem mechanics had in the last game -- You have to have one party member with high perception, and more than one party member with high perception is useless. Suggestions: allow traps to be found with either/or perception and mechanics, allow pooling of Per for trapfinding, allow pooling of Per and Mechanics for trapfinding, etc.

I don’t think symmetry is required, personally. A different effect for Resolve will give it more identity than anti-Perception. Yeah, that's fair, I'm just trying to make my suggestions minimal so they have as few unintended effects as possible, and if pet and res balance each other out, roughly, it should help prevent unintended secondary balance problems. Or that was my thought pattern anyway.

Eh, my preferred solution would be to reverse Str/Res move and go back to Might. I really don't like Str/Res for a lot of reasons, partly that Might was thematic with the setting, partly that the change makes hybrids much less viable. I've also never really liked the "heap extra defensive bonuses onto a stat" method of balancing them. Every stat has some defensive bonus already. Heap deflection and (is it Reflex?) both on Perception and you're just making Perception a kind of flavorless empty stat for tanks (like Resolve was previously). That said I admit I don't have an answer to this puzzle that I'm fully happy with, either. What I'd prefer to see is something that gives both Per and Resolve distinct, opposing "roles", like they had previously with interrupt/concentration (i.e., "bonus critical hits" vs "incoming critical hit reduction" etc).

I'm hesitant to support changes any bigger than the minimum necessary to get things in rough balance . . . the more that gets changed, the more secondary effects there are going to be (for example, increase Might's bonus too much, Barbarians start getting way out of hand due to readily available Might bonuses, etc.). One thing that could be done is giving Perception and Resolve non-numeric bonuses. Perception already has a bit of this in that it helps find traps (although that's relatively minor and should probably be combined with the mechanics skill *and* per both contributing to trapfinding). If they figure out a way to make Concentration & Interrupt benefit from Res & Per again, that could be a major help.

Adrian Tchaikovsky's Shadows of the Apt series does this; "magic" works, but only when it isn't being scientifically observed; some races "get" technology, others "get" magic; a magic-using race literally can't understand how to work a crossbow, a technological race can't understand a spell; the more ignorant and fearful you are, the more vulnerable you are to magic.

Well, it could be extremely cynical fiction where the protagonist is crushed by an uncaring and incomprehensible universe! Who says the protagonist has to win?

Yeah, Cthulhu Mythos is another good example of a system where the "unknowableness" is the point. In a sense we're re-treading theological ground ("Is an omnipotent being bound by the rules of logic?" etc.) Ultimately for purposes of a video game though yeah you need clear rules. If only because you can't program a computer to be ineffable.