Raenvan

on condition the progression pace is the same. It all depends on whether ability trees are front-loaded or back-loaded. The original intention must have been to compensate the overall slower progress of multiclasses towards higher levels with wider pool and greater number of abilities. But many melee ability trees turned out to be particularly front-loaded, so something needed to be done... Now single-classes gain slightly more ability points. As long as low-level multiclasses can combine strong initial passives from both sides, the current system seems ok to me.

When single-classes had gained one new ability per level in previous beta versions, many complained that they had been much behind multiclasses (who had got an extra one at every third level, that is +33%). Now single-classes receive two at every new power level (50% more), and become suddenly overwhelming?

Your argument makes sense, although I wonder how common it might be to go for the attrition route. By the way, Resolve isn't the best label for such a characteristic.

If most players (including me) keep Resolve at neutral 10, it becomes a constant "attribute". Which would be equivalent with a removed attribute, except for the major unforseen consequences

Even so, they have got rid of PoE Concentration and substituted Resolve conversation checks with skills. Almost nothing has survived, just the attribute itself (before the Might split).

Of course not, yet it's a big struggle to fill Resolve with value! It seems strange for me that the number of attributes had been set in stone long time ago, before defining their effects (and what they exactly represent).

+1 to ditch Resolve, and to revert Strength to Might. I'd rather give the Deflection bonus to Perception, as it seems to be somewhat inferior to Dexterity right now. I can't see a good place to the lost Will bonus. It might be easier to restore some kind of symmetry by removing the Fortitude bonus from Might, and Reflex from Dexterity.

It seems that Perception granting +1 Accuracy is a bit weak (about +2% DPS), but it would be overpowered with +2 Accuracy (+4% DPS). On the other hand, STR and RES affect weapon/spell damage similarly (+3% DPS), but are otherwise asymmetric because RES also increases Deflection. Moving +1 Deflection from Resolve to Perception (on the top of +1 Accuracy) could remedy both problems. In fact, such an effect should be naturally expected from Perception.

This idea has been already discussed in the following thread: https://forums.obsidian.net/topic/95062-simple-rebalance-for-resperdexmight. Proposals included +1.5 accuracy per Perception, or increasing Crit damage bonus.

Going back to one of the first posts... The original intention must have been to compensate to overall slower progress of multiclasses towards high-level abilities with more weaker ones. Beta versions fail this last aspect, as low-level melee passives can be overpowered. As long as some low-level abilities remain stronger than high-level ones, nothing justifies the extra ability points of multiclasses. There are two possible fixes: The elegant, but hard way. Ensure that all low-level abilities are weaker than higher level ones. Simply cancel the extra ability points of multiclasses.

You wouldn't have to. Single classes gain a new power level at every 2nd character level (except end-game), multiclasses at every 3rd (I don't see why end-game is different, though.)

I expected exactly the same linear scaling with power levels, and still believe that it would be best option. On the other hand, it seems that today's incomplete implementation has caused so much confusion. Some recommend removing power levels completely, or making them irrelevant (like +1% bonus per level). Multiclasses having zero level passives larger than 50% causes significant disparity only if those passives are overpowered. If, for example, paladins defenses at first level were +2 instead of +20, nobody could complain about serious imbalance. That's the most important point in power level scaling! As of now, many low-level passives grant (nearly) 100% of their bonus immediately. If this particularly front-heavy behavior (100%+0% per power level) were changed to linear or even slightly superlinear, multiclasses would become much less attractive. A moderately front heavy curve (like 30%+7% per power level) might still favor multiclasses, but the difference would be much smaller.

Straight reduction for multiclasses seems easier to code indeed, unless devs have a working system for power level scaling at hand. I'd miss some secondary goals of power scaling though (improved balance between single classes, partial regain of the lost 30% on higher levels, etc.) That said, I can imagine a minimal (not optimal), easy-to-implement plan that addresses the most severe single-/multi-class imparities. That would reduce the bonuses granted by few, specific abilities for multiclasses to around 60-70%. This reduction should be limited mostly to innate abilities of some classes and to selected low-level, easily accessible passives like Paladin's defense boosts and some Paladin subclass bonuses, Barbarian Carnage and Berserker inspirations, Rogue's Sneak Attack/Backstab, Cipher's Soul Whip/Biting Whip, Fighter's Constant recovery. Other multiclass passives wouldn't be affected, neither extra ability points of multiclasses. Such a restricted rationalization might solve a notable portion of the problem to my hopes.

Multiclass characters can't even get class-based passives that provide resistances until level 10 at the earliest (barring console hijinks) in most cases since they don't become available until power level 4. One Dozen Stood Against the Power of a Saint is a notable exception, though, since it's power level 2 and thus available to multi-class characters at level 4 rather than 10. Perfectly agree. I asked about low-level abilities specifically, because usually those make multiclasses superior. Medium level abilities matter much less by definition, since multiclasses get them only several levels later. These can form the core list of passives that should be rationalized first. I'd add Barbarian's Carnage, and Berserker's inspirations that grant extra penetration and/or armor rating given their primary importance. Since all of these abilities are obtained at low levels, the only option to introduce power level scaling is to reduce their initial effects. On later levels, however, they could scale up to (or even exceed) the current bonuses.

In the long term, ideally all abilities depend on power levels, of course. What I meant above is that a small number of low-level passives tend to make few multiclasses superior in the actual betas. Those selected ones could be rationalized and tested in relatively short time.