MaxQuest

Valid point. The selected granularity difference is much bigger than between 2H and 1H_FAST weapons, but your example helped me understand how to compute the average dps loss from overkill. It's actually simple : avg_overkill = avg_dmg / 2 avg_overkill_dps_loss = 100% * avg_overkill / enemy_hp And concrete examples: v1: hp = 141 (grubs) 2H: 18-23 -> avg = 20.5 -> avg_overkill = 10.25 -> avg_overkill_dps_loss = 7.2% 1H_FAST: 8-9 -> avg = 8.5 -> avg_overkill = 4.25 -> avg_overkill_dps_loss = 3.0% v2: hp = 638 (titan) 2H: 18-23 -> avg = 20.5 -> avg_overkill = 10.25 -> avg_overkill_dps_loss = 1.6% 1H_FAST: 8-9 -> avg = 8.5 -> avg_overkill = 4.25 -> avg_overkill_dps_loss = 0.066% Conclusion: 2H weapons lose on average ~4% more dps due to overkill than 1H_FAST weapons.

If you make 4 swings with DW_FAST it's overkill by 20 dmg. If you make 3 swings with DW_FAST it's underkill by 10 dmg. It may feel so. But the whole thing depends on: is enemy hp divisible by your weapon damage. And I don't know yet how to mathematically compute this possible discrepancy.

^ Good) but most likely on weekend. Btw, I have noticed yesterday that beta3 actually changed the weapons base damage. It shaked things up a little. Have updated the respective list.

You probably wanted to say that weapons with higher base damage have an effectively slightly less dps, because part of the damage is lost due to overkill The thing is: you make a valid point. But there is more to it. Imagine also such an example: - enemy has 100 hp - you have a 2H weapon that deals 50 dmg in 4s (i.e. 12.5 dps) - and you can DW_FAST that deal 30 dmg in 2.4s (i.e. 12.5 dps) The 2H setup will kill the target in 8s. (2 swings) The DW_FAST setup will kill the target in 9.6s (4 swings) What I want to say is that besides overkill there is also underkill damage; which partially equals the scales. I strongly disagree. For me 2H MUST be the best DPS. 2H > 1H Slow > 1H Fast > DW Slow > DW_Fast > > > Shield_Fast > Shield_Slow Since with my mod, fast_recovery_category is now 1.5s, and thus: - fast_melee_weapons have: 0.5s attack + 1.5s recovery - other_melee_weapons have: 0.7s attack + 3.0s recovery it's understandable that their damage requires adjustment. I've decided to start with the following base values, and see how it goes in order to build a deeper understanding over this topic: - 2H: 17-22 - 1H_SLOW: 13-17 - 1H_FAST: 8-10 Additionally am thinking to: - buff the innate bonus for single-wielding from +12 to +15 acc - buff the default bonus of a shield to +15 deflection This way: - single-wielding increases your dps by ~30% - dual-wielding increases your dps by ~30% (it's actually bouncing between 40-20% depending on other variables, but on average it's ~30%) - wielding a shield, reduces attackers dps (vs deflection) by ~30% So far, so good. Now I'm going to examine how those values play in different situations: Auto-attack DPS: - 2H: (17+22)/2 = 19.5; 19.5/3.7 = ~5.27 dps - DW_SLOW: (13+17)/2 = 15; 15 / 2.81 = ~5.33 dps - DW_FAST: (08+10)/2 = 09; 09 / 1.55 = ~5.80 dps - 1H_SLOW: (13+17)/2 = 15; 15 * 1.3 (from +15 acc) = 19.5; 19.5/3.7 = ~5.27 dps - 1H_FAST: (08+10)/2 = 09; 09 * 1.3 (from +15 acc) = 11.7; 11.7/2.0 = ~5.85 dps - SLOW_SHIELD: (13+17)/2 = 15; 15/3.7 = ~4.05 dps - FAST_SHIELD: (08+10)/2 = 09; 09/2.0 = ~4.50 dps PrimaryAttack DMG (without damage bonuses): - 2H: 19.5 dmg in 3.7s - DW_SLOW: 15.0 dmg in 2.81s - DW_FAST: 09.0 dmg in 1.55s - 1H_SLOW: 19.5 dmg in 3.7s - 1H_FAST: 11.7 dmg in 2.0s - SLOW_SHIELD: 15 dmg in 3.7s - FAST_SHIELD: 09 dmg in 2.0s FullAttack DMG (without damage bonuses): - 2H: 19.5 dmg in 3.7s - DW_SLOW: 30.0 dmg in 3.51s - DW_FAST: 18.0 dmg in 2.05s - 1H_SLOW: 19.5 dmg in 3.7s - 1H_FAST: 11.7 dmg in 2.0s - SLOW_SHIELD: 15 dmg in 3.7s - FAST_SHIELD: 09 dmg in 2.0s As you can see DW_SLOW results in highest amount of damage from Full Attack. But (!) they inflict it over 3.51s, while faster weapons do it much faster, and there is more time left for auto-attacking. Let's check how much damage such attacks deal over 3.7s: PrimaryAttack DMG (over 3.7s!)(without damage bonuses): - 2H: 19.5 dmg in 3.7s - DW_SLOW: 21.69 = 15.0 dmg in 2.81s + ~6.69 (auto-attack dmg, as there was 1.25s remaining) - DW_FAST: 21.47 = 09.0 dmg in 1.55s + ~12.47 (auto-attack dmg, as there was 2.15s remaining) - 1H_SLOW: 19.5 dmg in 3.7s - 1H_FAST: 21.64 dmg = 11.7 dmg in 2.0s + ~9.94 (auto-attack dmg, as there was 1.7s remaining) - SLOW_SHIELD: 15 dmg in 3.7s - FAST_SHIELD: 16.65 dmg = 09 dmg in 2.0s + ~7.65 (auto-attack dmg, as there was 1.7s remaining) FullAttack DMG (over 3.7s!) (without damage bonuses): - 2H: 19.5 dmg in 3.7s - DW_SLOW: 31.1 dmg = 30.0 dmg in 3.51s + ~1.01 (auto-attack dmg, as there was 0.19s remaining) - DW_FAST: 25.25 dmg = 18.0 dmg in 2.05s + ~7.25 (auto-attack dmg, as there was 1.25s remaining) - 1H_SLOW: 19.5 dmg in 3.7s - 1H_FAST: 21.64 dmg = 11.7 dmg in 2.0s + ~9.94 (auto-attack dmg, as there was 1.7s remaining) - SLOW_SHIELD: 15 dmg in 3.7s - FAST_SHIELD: 16.65 dmg = 09 dmg in 2.0s+ ~7.65 (auto-attack dmg, as there was 1.7s remaining) And let's make one more experiment. For example imagine that Primary and Full attacks come with a +30% damage bonus (like FoD): PrimaryAttack DMG (over 3.7s!)(with 30% bonus damage on PA): - 2H: 25.35 dmg = 1.3 * 19.5 dmg in 3.7s - DW_SLOW: 26.19 = 1.3 * 15.0 dmg in 2.81s + ~6.69 (auto-attack dmg, as there was 1.25s remaining) - DW_FAST: 24.17 = 1.3 * 09.0 dmg in 1.55s + ~12.47 (auto-attack dmg, as there was 2.15s remaining) - 1H_SLOW: 25.35 dmg = 1.3 * 19.5 dmg in 3.7s - 1H_FAST: 25.05 dmg = 1.3 * 11.7 dmg in 2.0s + ~9.94 (auto-attack dmg, as there was 1.7s remaining) - SLOW_SHIELD: 19.5 = 1.3 * 15 dmg in 3.7s - FAST_SHIELD: 19.35 = 1.3 * 09 dmg in 2.0s + ~7.65 (auto-attack dmg, as there was 1.7s remaining) FullAttack DMG (over 3.7s!) (with 30% bonus damage on FA): - 2H: 25.35 dmg = 1.3 * 19.5 dmg in 3.7s - DW_SLOW: 40.1 dmg = 1.3 * 30.0 dmg in 3.51s + ~1.01 (auto-attack dmg, as there was 0.19s remaining) - DW_FAST: 30.65 dmg = 1.3 * 18.0 dmg in 2.05s + ~7.25 (auto-attack dmg, as there was 1.25s remaining) - 1H_SLOW: 25.35 dmg = 1.3 * 19.5 dmg in 3.7s - 1H_FAST: 25.15 dmg = 1.3 * 11.7 dmg in 2.0s + ~9.94 (auto-attack dmg, as there was 1.7s remaining) - SLOW_SHIELD: 19.5 = 1.3 * 15 dmg in 3.7s - FAST_SHIELD: 21.64 = 1.3 * 16.65 dmg = 09 dmg in 2.0s+ ~7.65 (auto-attack dmg, as there was 1.7s remaining) So what can we tell now that we have these heuristics: that the initially selected values were pretty close, but 2H is asking for a 10% buff; while 1H_FAST for a ~10-11% nerf. Meaning that following minor adjustments: - 2H: 17-22 -> 18-23 - 1H_SLOW: 13-17 -> 13-17 - 1H_FAST: 8-10 -> 7-9 should be pretty close to a really balanced state. I'll rerun the complete numbers later, but rejoice Balthazar, from a pure auto-attack dps we get: - 2H: (18+23)/2 = 20.5; 20.5/3.7 = ~5.54 dps - 1H_SLOW: (13+17)/2 = 15; 15 * 1.3 (from +15 acc) = 19.5; 19.5/3.7 = ~5.27 dps - 1H_FAST: (07+09)/2 = 8; 8 * 1.3 (from +15 acc) = 10.4; 10.4/2.0 = ~5.20 dps P.S. Gee... didn't expect it will turn into such a wall)

Btw, a question regarding weapons base damage balancing. I see two ways: v1. all weapon groups have same dps: 2H == DW_FAST == DW_SLOW == 1H_FAST == 1H_SLOW. And in this case Primary and Full attacks are adjusted in order to make all options somewhat equal. And ideally you can take anything, but the choice won't matter much. v2. all weapon groups shine in some of their respective domain. For example fast weapons are best for plain auto-attacking. 2H for Primary Attack. DW_SLOW for Full Attack. But roughly they are ~20-30% atmost apart from each other in efficiency. And it's up to the player to figure out which is better suiting his concrete build. Thoughts?)

I am thinking about spells more from a point of view of possible builds and archetypes. Imagine that some player wants to build a aoe nuker. He take a wizard in PoE1 and happy progresses through the game. Usually 1-2 fireball is enough for his contribution, but if his party stumbles upon a not that numerous, but high-hp monsters, he knows that he will be able to pour 4x35 damage. A similar wizard, in Deadfire, has only 2 spellusages per rank. And we need to make sure that he has enough spell damage to seal the encounter. 35 damage in Deadfire can not be enough; thus we increase the damage, and cast time to compensate for it. Another nice thing is that heavy-hitting but slow fireball, additionally achieves the following: - it enables a new way to play a spell nuker - alpha-strike nuker. Opening from stealth with a big surprise can be both efficient and fun. Also think of either single-class evoker; or the synergy with assassin/deathblows. - it also enables speedy-nuker - a wizard that increases his action speed (via dex/frenzy/swift-strikes/bloodlust/potions) and can potentially reduce the 6s cast under 3s, when he really wants to invest all available resources into a fight, tipically a boss fight. And if such a speedy gonzales had only fast spells, he would quickly run out of stuff to cast. - it makes interrupt a more meaningful part of the game. Can't mod those via editing gamedatabundles. Technically that armor recovery modifier can be added to STR, by injecting custom code into Deadfire's dll. And if IE Framework works for PoE2 it can be done; but it is quite time consuming. So can two evokers) But frankly speaking the situation has been quite ameliorated, my melee dual-wielders no longer stormed through enemy ranks in under 6s, in beta 3. Exactly) you only take it if matches your playstyle and your party composition. The task is to make sure that it is efficient enough to be viable; and that there is a way to build your party to support you, and protect from interrupts and hard-cc. And that if you want to play that way - it would be good enough. But if it doesn't match your playstyle - there should be decent alternatives.

^ I understood that, and really like the idea, at least on paper. There could be some occasional inconvinience in practice, where lighter-armored offtanks could get ahead of the main tank, if you have the following formation: _______main_tank_______ offtank_________offtank But generally all of my offtanks (aside from 2 barbarians) were in plate armor anyway. So yes, I like the suggestion. And although can't implement it through editing the gamedatabundles, we can still discuss the corresponding values. Something like -1 stride for heavy armor, and -0.5 stride for medium armor? With Armored Grace taking care of this penalty as well?)

I am glad you had a blast) But looking at your party composition, are you sure the mod played a decent role in your enjoyement? I am asking because: - I haven't touched enchanters much, aside from speeding up Merciless Gaze and Eldritch Aim from 3.0/0 -> 0.5/0. - the mod also indirectly sped-up the recovery durations of weapons closer to their beta2 values: 4s (for average weapons) -> 3s; 3s (for fast weapons) -> 1.5s (that's also why I mentioned adjusting weapon damage for next mod iteration; and yeap I have checked KDubya's weapon related threads ^^) Weapons use the same recovery categories as spells (in bb_attack.gamedatabundle). So average weapons already should have 3s recovery, while fast weapons - 1.5s. I'll will double check this in the evening. Agreed enemies run too fast in combat. I haven't found a global setting, but there were a set of movers per creature_type. Have temporary set Lagufaeth running speed to 4 (down from 6) in BetaBalanceMod. Interestingly it feels much better (and need I say it was hilarious at value: 1). On one hand I understand that lagufaeth are designed to be fast. On the other - it doesn't feel good, when they stampede you. And on the... third hand they can use wild sprint and fleet feet, since there are many barbarians and wizards between them.

I am glad you had a blast) But looking at your party composition, are you sure the mod played a decent role in your enjoyement? I am asking because: - I haven't touched enchanters much, aside from speeding up Merciless Gaze and Eldritch Aim from 3.0/0 -> 0.5/0. - the mod also indirectly sped-up the recovery durations of weapons closer to their beta2 values: 4s (for average weapons) -> 3s; 3s (for fast weapons) -> 1.5s (that's also why I mentioned adjusting weapon damage for next mod iteration; and yeap I have checked KDubya's weapon related threads ^^) Weapons use the same recovery categories as spells (in bb_attack.gamedatabundle). So average weapons already should have 3s recovery, while fast weapons - 1.5s. I'll will double check this in the evening.

Am thinking that benchmarking welcomes a complex approach. Check a spell vs dummy; in combat vs grubs; vs titan; vs terror; vs one-two groups of lagufaeth.Compare it to the other spells of the same rank, as they compete for a spellusage. Try it with empower and without. And ofc take into account action economy. A spell might have a very good efficiency ratio. But if deals low-damage despite of being super-super-fast, you might end with no spells to cast. Usually none. But if you are very limited on the amount of spells you can cast (think of low-level wizard); he might simply have no slow and powerful single-target nukes/dots yet. So he might take Fireball because of it's high base damage. Fireball in PoE1 and Fireball in Deadfire are different things and serve different purpose.In PoE1 you could cast 4 fireballs and deal enough aoe damage that would surpass the max hp of enemies in the group. If it was the same in Deadfire, you would cast 2 fireballs and be thinking: now what? they are still alive. Another thing to take into account is that now we have interrupt mechanics. And the whole point of slowing down the casting durations, was partially related to make interrupting more.. prominent and easier to use (by having larger time windows when it can be applied). Atm many spells spells had their effects being subpar for a 5-8s cast. And this means that if we want to have a few 6s spells, we have to select a few which to buff. Fireball looks quite fitting for me. As it also enables new ways to play. Now the thing is how exactly to balance it. We have to take into account damage boosters (think of empower, evoker-assassins, damage multipliers on crit and overpenetration), speed boosters (dex, bloodlust, frenzy/swift-strikes/potions), the feel (the effect should match the duration spent casting), the frustration (if it fails), the overpowerness (if it oneshots), and so on But basically if a spell enables a new play-style that it is worthy the spent effort, and feels well and can appeal to a set of players; but at the same time it is not something that they will repeat over and over in 100% of cases (because it turned out to be... too efficient) - then the mission is achieved) I haven't touched Fireball damage in BetaSpeedMod. While it's cast/recovery duration is 4.5s/1.5s, which is kinda close to what you are proposing. As for BetaBalanceMod - I'm still testing it. And must say that it feels different on Normal vs PotD. I've increased the base damage to match the increased casting time. But increasing it further is a quite arguable path, since there is also Empower; and we should take into account even the most unexpected parties, like 5 evokers opening with empowered fireballs from stealth. What fireball asks in my opinion is actually more reliability. It would be frustrating to graze or especially miss, after a 6s cast. So I'll likely add acc bonus in future iteration.

The game will quite change till it gets released, plus there will be a ton of extra info available for us to take into account, so the mod will likely become obsolete, unless re-iterated I'd recommend checking it in the current beta) Gee, you selected the perfect enemy for an evoker. A monk/mage-slayer with high defenses) If you multi-class, you have 2 rank 1 spells and 1 rank 2 spells. Even if he doesn't interrupt or stunning blow you, and even if you hit with the spells, the combined spell damage is not enough to deal with his 257 hp. And... it would be pretty OP it if would) Have tried also single-class evoker. This way you get 1 extra spell of rank 2, and 1 of rank 3. Technically an empowered fireball could deal up to 150 dmg on crit, and finish him of via rays of fire. But grazes, grazes and stunning blow. No success so far. P.S. I've tried Fine Brigandine for survivability, as it helps better than Plate, since terror wields clubs. Also Spirit Shield helps a ton... but it takes the precious spell usage; and there is little left that you can use to bait his interrupts. P.P.S. It's hard to get enough armor on lvl 6 vs terror's 9 penetration; I'm thinking what will be when weapons will get +2 PEN more... P.P.P.S. Had a good laugh looking at your Shepherd kitting with animal companion) And one more thing: it's an interesting experiment, and I have tried pure-evoker and also a few multi-classed variants; but one single high-hp/def enemy is not really good target for low-level evoker benchmarking, as his strength is more related to dealing with groups of low-average hp'ed monsters (and either from stealth; or assisted by his party). Nice. I will think about it. Although can already say that I crit way less often than the enemies on PotD. So it will actually buff them more than the player) Btw it's important to remind that base acc is now lower than in PoE1. My lvl 6 single-class evoker has +15 acc from level, while it would be 20+3*5 in PoE1. And MinimumRollToGraze is now 25, compared to 16. P.S. if you want to try +0.5 crit damage/duration sooner, you can edit the following in bb_global.gamedatabundle: CritDamageMult: 1.25 -> 1.50 CritEffectDurationMult: 1.25 -> 1.50

Cipher's 2nd rank power: Mental Binding - instead of one roll vs Will for paralyze check, rolls two: - first one checks if the ability hit (vs Will) - and second as a confirmation for paralyze (vs Will) Screenshot Expected behaviour: - one roll vs Will for paralyze (for main target) - one roll vs Fortitude for immobilize (for all enemies in aoe) The fact that you have to confirm your roll, severely decreases your chance to succeed, and the actual usefulness of this power.

There is also a problem that Animal Companions lose their selected AI state on save-load.

Well, you have picked the spells, I haven't touched yet (see the changelog). I'm not surprised you are not happy with them I do completely agree that Wall of Flame, Essential Phantom and Prayer of the Body are underpowered when compared with their long casting durations, and the spells that you could cast instead (due to 2 spellusages per rank limit) Every spell is a double-edged sword. If you can cast the Necrotic Lance, so can lagufaeth wizards. Also having a few long-casting spells enables a multi-class wizard that learns only a handful of spells and heavily invests in passives instead. I think it's ok to have a very few 4.5s single-targeted spells as long as their effect matches such long cast-time. Also we should keep in mind the effect of DEX, +% Action Speed alchemy potions/Frenzy/Swift Strikes, Bloodlust; and the fact that we can have a pure spell-nuker with dumped STR. Ideally there would be several categories of damaging spells: - relatively quick - easy and comfortable to use, but not fight-deciding - slow - with big oompf damage, heavy backlash in case of interrupt; and where you have to build your party around them > basically if you want to cast a 6s spell, you: - really want to invest in DEX (being heavy buffed, you can reduce 6s to ~2.5s) - or use it only once from stealth; and actually invest points in stealth - have a party of debilitators and concentration-providers that would cover you - want to decide which race with which affliction resistence to take apriori but there is also a sense of satisfaction when this spell comes off, as it has the potential to seal your victory > combat duration is quite tricky. With a good melee party on Normal, combat often ends in 9-15s. While with a random party on PotD it can often take up to 60s. Also we should take into account nuking alpha-strike parties. Imagine a group of 5 stealthy evoker wizards on hardcrack, that open up with 5 empowered fireballs and rest after each encounter. > agreed of self-buffs. And agreed that prayer of the body needs a speed-up. But speaking of the "boost" - I haven't explicitly changed Prayer of the Body. I had no time to go through each priest spell one by one (except Spiritual Weapons), yet) You can see the powers and spells I have explicitly examined in the changelog) > agreed on ciphers and chanters. As you have probably mentioned ciphers have -1.5s recovery on majority of their powers compared to vancian spells. Also chanter summoners now have more time to do other stuff.

The idea behind BetaSpeedMod was to: - slightly speed-up casting and recovery categories - tune-up casting cipher - increase the speed at which you can conjure summoned weapons and check if it feels better than in current beta) so far it's 5 vs 0, that it indeed feels better, with on emphasis on "much" and "significantly". I am happy with that) While the idea behind BetaBalanceMod is to actually improve: - consistency - viability of underused play-styles But it requires a much bigger set of changes; and also extra ingame testing from my part. So it's still a WIP, and actually quite at the start of it. You have probably noticed that I haven't had the time to touch priests and druids yet; nor rank3+ on wizard. I don't think I have abberated too much from that table) (looking at the spells listed in changelog). As sometimes spell's effect is too weak or too strong for their matching cast/recovery; and I feel that table could be a bit expanded. But I'll re-iterate on it, in order to bring better / more intuitive grouping. Glad to hear it) Regarding evoker: it looks like he has to heavily rely of Fan of Flames, Rolling Flames and narrow passages in the early game. Which single-class character had easy-time where evoker couldn't advance?

I am missing a few AI settings in current beta, that were there in beta 2. Specifically "Auto-Attack" behaviour and "Behavior Priority". Compare: beta3 with beta2. Also I'm struggling with creating the most simple auto-attacking profile via behaviour editor. It would be nice if along with Abilities we could also select Auto-Attack: example

^ Well, for 5 days that's for sure) And you might want to check BetaBalanceMod now P.S. Gonna take a sleep. It's a hella time consuming)

So, as Josh stimulated us to try different casting speeds ourselves, and since I've already been tinkering with those for awhile (props to Andrea for the idea) I think I would share these modifications for everyone willing to try it. I've made 2 mods: - BetaSpeedMod - it changes the values of casting and recovery categories, and also adjusts several selected spells - BetaBalancingMod - it includes the above, and also has a few balancing changes. This is still a WIP and I am experimenting with these. Here's the full changelog: - BetaSpeedMod: - BetaBalanceMode: Here are the download links: - BetaSpeedMod: download v2 - BetaBalanceMod: download v2, download v3 How to install: - go to <Pillars> folder (e.g: E:\Steam\steamapps\common\Pillars of Eternity II - Public Beta) - create 'override' folder in <Pillars>/PillarsOfEternity2_Data/ - choose one mod (BetaSpeedMod OR BetaBalanceMod), and unzip it to /override folder - you shall have the following path: <Pillars>/PillarsOfEternity2_Data/override/BetaXMod/gamedata/ What to do now: - if you are interested: try it and share how it feels) P.S: I am thinking now: - of iterating through all weapon types, as their base values require a few adjustments - re-testing few spells in combat (especially fireball) - adding -20 respective defense malus to rank 3 afflictions (still gotta figure how to best do it) - and perhaps later writing down the reasoning behind these changes. But am also curious: - Josh mentioned that they are going to increase the penetration of weapons by 2. Do you want this implemented in the mod?

Good point. Than mana + cooldown? Bah, now we have DA:O system which I never liked very much. Cooldowns are also so-so of an idea as they don't play well together with the current DEX. You might want to build a quick caster... and then having to wait for cooldowns in each encounter, because you were too fast. Sure DEX could lower the cooldowns as well. But personally I don't see yet why do we need mana pool in Deadfire in the first place. If a player is repetitively using the same tactic with his spellcasters - it's either that some spells are clearly better than the other (and can be fixed via balancing), or there are repetitive enemies who act in repetitive manner and that's on encounter designers. Imagine if enemies mainly used quick paralyze to interrupt you (like lagufaeth in current beta). Player tactics would quickly shift towards making wood elves and using slow spells. And than suddenly there are enemies who would use stun instead... or enemies with really high reflex or immunity to fire.

Speaking of low levels, lets compare a level 5 wizard in PoE1 and Deadfire: PoE1: - has access to rank 3 spells - has 4 spell-usages of rank 1, 4 of rank 2, and 2 of rank 3 - has 2 talents, 10 learned spells + x from grimoire + arcane assault Deadfire (single-class) - has access to rank 3 spells - has 2 spell-usages of rank 1, 2 of rank 2, and 1 of rank 3 - has lets say 2 talents, and thus 3 (or 6 in preBeta4) remaining points for spells + up to 6 spells usable from grimoire In regular fights my low-level wizard was making 1-2 casts per encounter; and was feeling like a liability. While in harder fights, such as Raedric, he was unleashing his whole arsenal (10 casts) and was feeling like an mvp. Deadfire somewhat balances this. A lvl 5 single-class wizard can cast 5 spells. And here he can choose to take mostly long cast spells; or he could focus of faster spells, with intent to switch to auto-attacking after that; or simply invest less in DEX and stretch his arsenal along longer time span. But yes, I would still advocate for ability to learn a bit more spells, as it increases versatility of our casters. And also few talents, like "+1 bonus spell-usage of rank 1 spell" that you could take on power level 4; "+1 bonus spell-usage of rank 2" that you could take on power level 5, and so on; if there are no items ingame that provide bonus spell-usages. Agreed, if we'll have to repeat the same approach in every encounter, again and again, it gonna feel repetitive not fun. A few things could partially help with this: increase the noise parameter on the spells that take more than 3s to cast. It's ok if you can stealth-fireball an eoten, but if there are some scout enemies or rangers with animal companions, they would get more alert and will have a chance to interrupt you. Hence we might want to use another opener. decrease the maximum range of ranged aoe offensive spells while out of combat. To let's say 7m. This will make opening with the slowest aoe spell - situational; and also help combat potential exploits, when killing enemies not yet visible, but known to be there. make different approaches equally viable: start with slow fireball from stealth - and have a support toss aoe concentration once spotted start with slow fireball from stealth - and let your dedicated debilitators take care of enemy interuptors start with quick hard cc - and let glass-cannons engage start with quick buffs - follow up with soft cc - and brute force stuff start with dots - follow up with heals, defense buffs - and just wait aoe immobilize and blind enemy backline (for at least 12s) - this way you will be out of reach for them - and will be able to deal with enemies not all together, but one group, than another. aka divide and conquer aoe immobilize or hard cc enemy frontline - and let assassins deal with their backline If I will have mana pool on casters, there is a high chance I'll just take a lot of passives, while spamming the 1-2 most efficient spells for that character. For example: Cleansing Flames -> Shining Beacon x N (while there is mana). I mean there is a chance that this change will increase the repetitiveness, not hinder it.

That's great! And it's not like there will be many pre-release mods anyway. Thank you a lot for the pointers! Will definitely attempt to make a sidekick at some point. *scratching head* while looking at the characters.unity3d filesize of 0.98GB... is there a specific reason, why you guys decided to bundle into such big packs, instead of lets say: /assetbundles/characters/character_a.unity3d /assetbundles/characters/character_b.unity3d /assetbundles/characters/character_c.unity3d ... because atm in order to add/edit at least one character prefab, modder would have to unpack/do_stuff/repack the whole characters.unity3d and if he's successful to somehow also redistribute it. Am not even sure if nexus even permits such big sizes; plus repeating all this on each update :shiver: Also, with smaller .unity3d files modder can tinker with the help of AssetsBundleExtractor. While if the file is big, you are literally getting lost there ^^ (having to literally check all monobehaviours when you want to find a specific one). Sure there is also DevXUnity-Unpacker Studio (which looks more user-friendly, and supports repacking), but it is far from free. Was the copying speed somehow of a problem due to the big amount of files?

BMac, can we also have some possibility to set the load order of our mods? For example, I have: \Pillars of Eternity II\PillarsOfEternity2_Data\override\[MOD_A] \Pillars of Eternity II\PillarsOfEternity2_Data\override\[MOD_B] \Pillars of Eternity II\PillarsOfEternity2_Data\override\[MOD_C] and would like to add an (optional) file: \Pillars of Eternity II\PillarsOfEternity2_Data\override\mod.order.json { "mods":[ {"name":"MOD_C", "enabled": 1, "params": {}}, {"name":"MOD_A", "enabled": 1, "params": {}}, {"name":"MOD_B", "enabled": 0, "params": {}} ] } This will really help in future, if there will be some partially conflicting mods. Also, can you give at least some general, vague pointers on how to create a custom, sidekick-like hireling, and attach custom barks and on-event banter?

Best: - graphics upgrade - substantially better character models, textures and animations (but!) need more good face presets - the idea of multi-class - the idea of subclasses - exploration and setting Worst: - (by far) missing, confusing, innacurate or even lying tooltips - erratic/chaotic combat - unbalanced state between single-class and multi-class, between phys dps and spell dps, between dps and cc, between melee and ranged - underpowered crowd control - mushy power level - the occasional frustration from "if only I had 1 more AR right now" and "oops, these two don't stack" - lack of versatility. It's ok to have really limited spell-usages, but the amount of spells and powers you can learn is a bit overlimited. Unsure: - the game often makes me wonder: why exactly those formulas are used TL.DR: + setting and atmosphere - lack of balance and transparency

Already experimenting for a week now. I've tried setting different categories for many individual spells. Summoned Weapons even at 6.0s/0 -> 1.5s/0 do not feel OP right now. Or for example Eyestrike 5.0s/2s -> 3.0s/1.5s doesn't make cipher Supreme. Or Merciless Gaze being kinda bad at 3.0s/0; (when taking into account it's effect and comparing it with opportunity cost) it's a very situational spell of choice even at 0.5s/0. And these changes really feel good, despite of having an amplitude bigger than the mentioned 1s. As for mentioned cast categories per se (average and slow), I technically only changed the slow one from 5.0s -> 4.5s

I.e. they will increase base focus gain from 0.25 to 0.35 like in PoE1? Can someone explain this? As it rises quite a few questions. They increase PEN of weapons with 5 penetration to 7. They increase PEN of weapons with 7 penetration to 9. What about: - weapons with PEN 9, e.g. estocs? - fist damage? Does this affect non-kith enemies as well? Do I understand it right that it basically buffs weapons in comparison to damaging spells (whose PEN is not affected by this change)? Or the status-quo is kept the same, provided that we take the buffed empower into account? Also does Empower increase the penetration of the spells? And will Empower start affecting abilities like FoD, Barbaric Blow, etc?