Raenvan

Since power levels are intended to balance single and multi-classes as of now, they should ideally apply for melee types as well. A fourth, partial solution might be based on power levels, if innate abilities and at least low-level passives were linked to them. For example, the Berserker's penetration bonus would be capped at his power level (and increase on relevant level-ups), and the Paladin's defense bonuses shouldn't be higher than an appropriate multiple of his power level. In this way, multi-class characters would advance somewhat slower than single-classes.

The problem is that both Paladin and Berserker have very strong low-level abilities, particularly innate powers. Combining two strong abilities into a first-level character is always better than going single-class, unless they provided reduced effects separately at low levels. Making low-level abilities depend on power levels might be a step towards the solution. For example, Berserker's penetration boost should not exceed his power level, or Paladin's defense boost should not exceed some multiple of his power level. Although such adjustments would also hit single-class characters, multi-classes would have at least some disadvantage at low and/or medium levels.

The conclusions about dual wielding and armors seem correct, something needs to be done to make two handed weapons viable. Your experiment had one more important ingredient: how much stronger the heavy armored Berserker/Goldpact with all those nice stuff is than a regular warrior. If one plays such an invincible character, (many of) his attributes hardly matter at all. It was proposed in this topic to decrease his Constitution, since he doesn't need as much health. True. On the other hand, he doesn't need higher Dexterity, Might/Strength, Perception either, as these would only shorten his fights by roughly 3% per point, but time isn't relevant for him, having been unhittable.

Would you share Dex and Con of the two warriors, and what you meant under "super heavy armor guy did way better." It's no surprise he wasn't penetrated, but I suppose the heavy armor guy must have had even higher Dex if their DPS matched.

Nice idea and clear examples! Haven't checked the math, but this comparsion might work similarly between heavy/medium armored high Dex guy and medium/low armored high Con guy. Given the current speed and health bonuses and armor/penetration values, various characters can benefit from moving 6 points of Dex into Con. Including those who plan to wear any light/med/heavy armor, have at least about 10 Dex and not more than 12 Con. As pointed out, this change can grant similar DPS after armor switched to one or two grades lighter, because lower overall speed is compensated by lighter armor. gives 30% larger health, which usually overcompensates the possibly larger penetration from enemies. Exceptions include purposely very fast characters who usually wear only robes, and those who already had high enough Constitution. This observation can be a good argument against treating Con as a (second) dump stat.

I admit that the exaggerated example was indeed oversimplified, and mostly irrelevant in party-based combat. Consider now typical group encounters. One might want to analyze what happens if Constitution of all characters were raised by one point at the expense of one Might or Dex, while strategy remained the same. Simply speaking, all his damage dealing slows down by 3% (provided that spells hold out) and all his characters' HPs increase by 5%. It would be be surprising if AI adapted anyhow, so his characters can last 5% longer in average. If the encounter was victorious with original stats, odds are high to win with modified ones, too. More precisely, odds are at least as high as in a replay with original stats, given the random nature of rolls. If the encounter was lost with original attributes, one usually replays with revised strategy. That's fine, and can always work. However, if the encounter was only marginally lost, there's a slight chance that it could have been just doable with the same strategy and modified stats. At least, odds can be slightly higher than odds of a replay with unmodified stats. Few words on criticism. It was claimed that high Con was unnecessary because healing is so abundant. It can be true, but as shown above, raising Con doesn't make things worse, either. -1 Per or Dex decreases the frequency of inflicting criticals and afflictions like stun, sure. Might to Con redistribution seems to be neutral, though. It's very intuitive to think that further increase of Constitution is meaningless, if encounters are mostly won. However, raising it while decreasing Might (Str, or maybe Dex) doesn't hurt, either. The threat to very large Con might be the chance to run out of spells and abilities during the extended combat phase. In this case, further bumping is not advised. PS. The intention here is neither demanding radical change of an approved attribute based on fictive examples, nor pushing full Con parties. Just to show that raising Constitution is a very viable approach.

No offense, but I'd like to explain once more why Con is OP to my opinion. We've run through numerical examples in replies no. 3 and 4, here let's replay a 1-to-1 fight along your lines. I've learnt the lesson that only one thing matters: who survives with at least one HP left. Assume two pure warriors (no buffs, spells, whatever): Durable frontliner first with 15 Con and 20 Might/Str, and Meatbag second having 20 Con and 15 Might/Str. Durable delivers 13% more damage than Meatbag (+15% nominal), and Meatbag has 20% HP more than Durable (+25% nominal). In contrary to what should be the "good" outcome, it's (mostly) Meatbag who wins with few HPs left. While I admit that such 1-to-1 encounters are not and should not be typical, I do believe that a well-balanced game should give some adequate answer to this problem. Anyone knows an answer?

As far as resting isn't limited, you are absolutely right... Thanks, these arguments make lot of sense, but I still miss something. To your opinion, 15 Con is almost always enough, but other attributes are worth raising beyond 15. It may be translated to such a statement that +5 Con is as effective as +6/7/8... Might. Which means that +5% Health bonus is yet overpowered

Your 15 Con frontliners make sense, but 20 Con dwarves would serve even better, given the 5% HP bonus. This example also proves that Con is overpowered. Yes, Con is less inbalanced than current Resolve, but I guess that some fine-tuning is possible without increment problems. After base bonus relaxed to +3% Health, one of following options might be implemented: Current health loss per injury is 25%, and will be cut to 15% in next beta. These seem just arbitrary numbers to me. How about Constitution-dependent attrition, such as (30-Con) or (25-Con) percent? New beta will somehow decide what is really bad injury. Let the damage threshold depend on Con, give the same +3% or +5% per point bonus to it. These yield 130% or 150% threshold for a 20-Con frontliner, and 79% or 65% for a 3-Con lurker. Roughly half for the lurker.

Nice examples, although you assumed equal bonus from Might and Con. This is balanced face-to-face, but the offensive one is better against multiple foes, as you said. In this beta, B gets +50% Health from +10 Con (150 HP, 10 DPS), while A gets +30% damage from +10 Might (100 HP, 13 DPS). B kills A in 10 seconds, and is alive with 20 HP. Considering 1:1 fights, B is definite winner. My question is: What little advantage should Con provide on the top of the balanced +3% Health per point (face-to-face case) to compensate for the disadvantage in group fight? To my opinion, it could be related logically to the new Health/Injury system.

The point is that Constitution could be the most overpowered, and at the same time underrated attribute in Deadfire! Offensive stats like Might, Dex and Per helps to hit harder, faster or with higher probability with up to +3% per point, and easily catch your eye. One extra point in Con doesn't increase DPS, but allows to last (and deliver damage) for 5% longer time. Comparing to the other defensive attribute, current Resolve, the difference grows even larger, as one point there saves only 2% damage in average. I'm afraid if Constitution bonus were smaller (like in early PoE1), almost nobody would consider raising it. On the other hand, the current bonus seems to favor pumping Con more than anything else. I'd prefer if the current Constitution bonus were relaxed to comparable levels (say 2-3%), and some new factor could be brought into play. Like effect on number of allowed injuries, or on health loss percentage after injuries, etc. These have been already mentioned in other topics. Any thoughts?

We should compare this optional 30% bonus to the normal 25% damage bonus of crits. And it's extremely situational: you need a critical hit, and a Penetration that falls between 133% and 200% of your opponent's AR. Constant +10 Might is much better, isn't it?

The estimated effect of the +50% penetration bonus is +16.6% damage in average, assuming very simplified conditions. It gives a huge boost only if your original Penetration was low, which every damage dealer should avoid. In all other cases, you may have some chance to do +30% damage. Following your arguments, criticals' damage bonus should be somewhere between 50% and 100% (the latter without penetration bonus) to get balance with Might/Dex.

As you proposed earlier, 4x larger probability of crits with 25% damage bonus, or equivalently, same probabilty with 100% damage bonus would give near parity with Might. This assumes no penetration bonus, no interrupt bonus and +1 Accuracy per Perception. So your 50% bonus is absolutely needed, unless Perception grants +1.5 Accuracy (which I still prefer). If Perception were scaled up with real criticals, Resolve would be balanced automatically to my opinion. No though work required. If somebody dumped Resolve, he'd suffer from enemies' crits on him. As for suggesting interrupting criticals, I meant not only some chance to interrupt, but inherently 100%. Which is one more argument not to dump Resolve. I may have a very simplified damage estimation for criticals' penetration bonus. Let's assume that penetration and armor rating have uniform distribution in the same interval. In this case, the crit's penetration bonus upgades no penetration to normal one with 16.6% probability (-70% damage restored), and normal to full also with 16.6% (+30% damage). In average, the damage made is increased by 16.6% due to penetration bonus. Which is smaller than the 25% damage bonus we have now.

Exactly! And I'd consider adding an inherent Interrupt to critical hits, as well. If I'm right that's equivalent with 1 of the old 3 points from PoE1, so not overpowered. 50% critical bonus + auto-interrupt + penetration bonus (whose expected extra damage hasn't been calculated to my knowledge) should be attractive enough to raise Perception. Or to raise Resolve to avoid received crits.

Fractional values, say +1.5 Accuracy, do matter. Imagine an encounter where your Accuracy equals your opponent's Deflection. A roll of 50 means you just grazed. Had you +1.5 Accuracy, it would be a hit similar to the effect of +1 or +2. In another fight, your opponent's Deflection is one higher, and you again grazed with a roll of 50. Had you +1.5 Accuracy, you would still graze (51 is the lower limit of Hits) similarly to +1, but different to +2. In the long run, +1.5 is equivalent with +1 in half of the fights, and with +2 in the other half. Having attributes integer modifiers is only for beauty, not for game mechanics. Edit: The above explanation is incomplete, it only shows why +1.5 is not equivalent to +2. To prove the other part, that it can differ also from +1, one assumes that enemies can have integer+half Deflection. Like -1 lower base, and +1.5 due to Resolve. In an encounter where your opponent's Deflection is higher by 0.5 than your Accuracy, you graze with a roll of 50. You would still graze with +1 Accuracy, but hit with +1.5 or with +2.

Resolve would be automatically balanced In that case, since Deflection has opposite role to Accuracy. It's been discussed in the other thread https://forums.obsidian.net/topic/95062-simple-rebalance-for-resperdexmight/. Before splitting damage to Str/Res, of course.

Let me summarize what has been already found. The change should be as minimal as possible, so Might and Dex bonuses remain +3% per attribute point. Perception and Resolve bonuses could be boosted various alternative ways, but preferably symmetrically (as they have been after early PoE1 versions): Restore original Interrupt and Concentration mechanisms and modifiers. Increase Accuracy and Deflection bonuses to +1.5% per point. Rounding is probably not a real issue in long term. Secondary defenses (Will, Reflex and Fortitude) might be scaled, although they are already larger than the necessary 1.5% to my opinion. Adjust other combat mechanisms that amplify the gains from unmodified +1% Accuracy and Deflection increments. Like Graze interval and Crit bonus. This idea involves more variables, and could be more difficult to balance than option 2. Invent new unique and/or separate values to Perception and Resolve. It will be tough to balance them with other attributes, and to get widespread approval (see the case with Might to Str/Res split). I've read many suggestions for Res, fewer for Per.

In order to systematically exploit the rounding issue, you'd have to know and adjust to the Deflection bonus of all your enemies. It's the actual Accuracy-Deflection what matters, not pure Accuracy. So it's impossible to exploit rounding except single occasions. It doesn't matter in the long run. I can see the rationale in the second argument. Accuracy and Deflection bonuses were equal in PoE1 for few years, and it was widely accepted. But I'm not sure how Accuracy bonus was, and should be related to other defense bonuses. That's an option too yeah but I'm hesitant to downgrade Dex because it worked fine at 3% for the whole first game and I don't want to muck with something that works; so I'd rather bring everything else up to the 3% point if possible. edit: hrm What about if you changed critical hits to +50% damage, instead of +25%? Then, either (miss to graze + Graze to hit) OR (hit to crit) would be mathematically, roughly, against equivalent defense, equal to +(one half an accuracy point). So you could give Per an additional +1% hit to Crit, and give Resolve (+1 downshift everything)+(+1 crit to hit) and they'd be equivalent. Perception and Resolve have been nerfed compared to PoE1 with the removal of Interrupt and Concentration bonuses. First simple rebalance option is compensation by increasing the remaining bonuses by 50%. Second simple option is to nerf all other attributes by 33%. If you don't like either options, how about boosting everyting, although with different amounts? For example, Resolve gives 2% Deflection, Perception gives 2% accuracy, Might gives 4% damage, Dexterity gives 4% action speed, etc.

If you don't like non-integer bonuses, let's tone down all other attributes by 33%. So +2% damage per Might, +2% action speed per Dexterity. One can find proportional (possibly rounded) values to Int and Con, as well.

I assume this proposed change is instead of the might>strength change. Fix perception and resolve independently rather than changing Might. You are right, damage bonus should be kept at original Might. Resolve would be overpowered with damage there.

The purpose of this post is to rebalance Resolve and Perception, which both have lost important effects (Concentration and Interrupt) since PoE1, with unmodified attributes such as Might and Dexterity in a very simple way. There are many ideas to improve Resolve, which was considered almost useless without Concentration modifier. Options include damage split to Str/Res, kind of shield against new Interrupt, negative duration modifier to afflictions, moving duration from Int to Res, etc. None has been thoroughly tested so far, while some of these seem to be overpowered or ambiguous. The new proposal is based on a different perspective, on the almost equally nerfed Perception. Without Interrupt modifier, Perception affects only Accuracy and the related hit chance in combat. +1 Perception grants 2% additional DPS in a typical case (Accuracy = Deflection) plus maybe some extra Penetration, clearly inferior to Strength and Dexterity, which give 3%. The 2% DPS modifier comes from the 1.25% miss-to-crit conversion and 62.5% expected damage (50% hit and 25% graze), and slighly varies with different Accuracy to Deflection scenarios. This weakness of Perception has been hardly recognized and discussed so far. The simplest solution would be equalizing the DPS effect of Perception/Accuracy with that of Might and Dexterity. Simple math results that Accuracy should be increased by +1.5 per Perception point instead of only +1. Since the combat effect of Resolve is the inverse of Perception, Deflection should be increased also by +1.5 per Resolve point. Res/Per/Dex/Might are mathematically equalized in a such simple way! I can imagine that many players would still not be satisfied with Resolve giving just 50% more Deflection, but this feeling might be related to the fact that defensive playstyle is generally unattractive. The real question to my opinion is whether Resolve/Deflection is balanced to Constitution's +5% Health increment. However, this question has been already there in PoE1...

Right, but that additional 1.25% is roughly 2% of your total damage -- i.e., if accuracy and deflection are equal, the "expected value" of a weapon swing works out to be around 62.5%, and the additional 1.25% from +1 point of accuracy ends up being about 2% of that 62.5% total, so 2% additional damage. That's like the best case scenario for the expected value of an additional point of Per, too. Accuracy gets above 25, relative value drops fast. Might and Dex both have similar diminishing relative returns for each additional point, but the curves start much higher and for most in-game ranges you're going to get between 3% and 2% dps increase per additional stat point, whereas for Per it's like between 2% and . . . probably somewhere under 1%, I haven't crunched the top end numbers. Thanks for the correction. At least, Perception is less negligible this way than I thought. If it granted +1.5 Accuracy instead of +1, its effect on DPS (3% per point) would be similar to Might and Dex for a typical scenario (Accuracy = Deflection). As you said, Perception becomes weaker against weaklings, since you gain either 1.25% or 0.75% (miss-to-crit or graze-to-crit) in the current beta and your expected damage increases. It can get weaker or stronger against bosses who have high Deflection, because crits would be almost impossible and miss-to-hit gives only 1% extra damage, but your expected damage also drops considerably below 62.5%. Good idea. If somebody doesn't like +1.5 Accuracy/Deflection per attribute point, all other attributes should be toned down proportionally. Might and Dex gives only +2% damage/speed, for example.

As Concentration was removed from Resolve, most of us think it's pointless. Interestingly, although Perception was also deprived of Interrupt, nobody complains. Perception has become almost equally worthless (at least, for a defensive character) to my opinion! +1 Accuracy is just the offensive version of +1 Deflection, after all. Unless Interrupt and Concentration are restored, or Resolve and Perception receive something really valuable, both attributes could be expelled from Deadfire. If somebody complains about the lost symmetry of the defenses, let's just remove Fortitude from Might, too. It isn't a pure physical attribue, anyway. Yeah, I didn't realize how problematic Perception was now till I crunched the numbers and realized that each point of Per gives you at best about 2% additional damage (by changing misses to crits mathematically) whereas increasing might or dex (or even str/res) gives you a 3% bonus (roughly). Actually miss-to-crit gives only 1.25% additional damage, plus extra penetration and duration. Nevertheless, we have three or four (with new Resolve) attributes that increase DPS various ways. This is a bit redundant, and Perception is most probably the weakest one.

As Concentration was removed from Resolve, most of us think it's pointless. Interestingly, although Perception was also deprived of Interrupt, nobody complains. Perception has become almost equally worthless (at least, for a defensive character) to my opinion! +1 Accuracy is just the offensive version of +1 Deflection, after all. Unless Interrupt and Concentration are restored, or Resolve and Perception receive something really valuable, both attributes could be expelled from Deadfire. If somebody complains about the lost symmetry of the defenses, let's just remove Fortitude from Might, too. It isn't a pure physical attribue, anyway.