A Defenses and Attributes Minifesto (Or, What Do We Do With Dexterity and Armour?)

TL;DR: GGG should make defenses primarily come from Attributes, with further defensive nodes being pushed towards the outer ring and edge of the tree. Armour should be reworked into % total damage reduction with the amount needed to reach any given % reduction scaling with the level of the monster dealing the hit.

The Problems
As mentioned by Jonathan and Mark in the interviews, the attribute bonus from Dexterity has been a consistent pain point for GGG. There's always been a question, apparently, of what an appropriate attribute bonus from Dexterity should be. Building a dexterity character in PoE2 caused me to come to a similar question; What exactly I'm building all this Dex for? It only gives accuracy, as compared to Intelligence and Strength which give mana and life, respectively. Of course, it also unlocks gear, but it still ended up feeling more like an obligation rather than someone I was excited about.

For players, PoE2's balance of Armour as a defensive layer directly competing with evasion and energy shield has drawn many complaints from both players and high profile streamers regarding its lack of effectiveness defensive layer, especially without the additional layering of Fortification and % damage reduction from Endurance Charges. With the changes to Energy Shield no longer being bypassed by Chaos Damage and Evasion no longer being bypassed by Spells, only large AoE attacks (and Acrobatic further mitigating that for high-investment Evasion characters), Armour's lack of changes coming into PoE2 makes it inherently in a worse spot compared to the other primary defensive layers even as compared to PoE1, which I believe is what many players are feeling when they make these complaints.

Related to point 2, GGG's decision to remove life nodes from the PoE2 passive skill tree has made investing in survivability heavily slanted towards energy shield, as %increased and +flat ES is the most effective way to increase a character's base numeric life pool. To make matters worse, the ES nodes are heavily populated in the 12 o' clock direction on the skill tree, corresponding to the Witch/Sorceress starting location, and wholly absent from the lower half. While this naturally makes sense for ES as the primary defensive layer associated with Intelligence, it also heavily slants building for defenses towards building ES and Intelligence, making it harder for non-Intelligence based classes to build survivability. In PoE1, these characters would just take more life nodes to make up the difference, but with the removal of life to being purely a reward for Strength stacking and the weakness of Armour as a defensive layer, Intelligence characters are observed to be by far the most effective characters at building defensive layers and survivability.

GGG has also expressed that defensive nodes being must-takes in PoE1 does not align with their vision for how they want players to interact with the skill tree as part of their justification for removing life nodes. However, some amount of defensive nodes are always going to be must-takes, or come off as must-takes, so long as they exist and the game has consequences for dying. Removing the consequences for dying is obviously a non-starter. Players also deserve the opportunity to invest in their defenses and determine the balance of survivability vs damage, and assume the level of risk that their skill allows.

The Role of Attributes
There already exist nodes that every single player takes on every character; the attribute highway nodes. Every single player is going to take at least the minimum number of Attribute nodes required to use their gear and active skills. Thanks to the flexibility added in PoE2 to decide which attribute a highway node gives you, GGG has an opportunity to move defenses to attributes, giving players a choice of which layer they want any given node to advance regardless of their start location on the web, but still limited by their intended build its attribute requirements. I propose:

>1 Strength gives X Armour
>1 Dexterity gives Y Evasion
>1 Intelligence gives Z Energy Shield

I use variables here to reflect the fact that I do not have the formulas that GGG uses to determine how much EHP each defensive layer gives; But in general, each point of an attribute should give approximately the same amount of EHP as expressed through Armour, Evasion, and ES, respectively (this means it will probably take more Intelligence per point of ES as compared to the other two stats and their defenses, but it is what it is). This helps solve the 'Ok but what does Dexterity actually do' issue, establishes a baseline level of expected defensive capability that should be consistent across all characters, and establishes a clear, specific role for attributes as both the primary source of the primary defensive layer associated with that attribute as well as the key to unlock gear that also gives more of that same primary defensive layer.

What Do You Mean By Primary and Secondary Defensive Layers?
By primary defensive layer, I refer to Armour, Evasion and Energy Shield. Under my proposed system, these would be distinguished from secondary defensive layers (e.g. block, regen, leech, etc.) by the fact that they are tied to Attributes. This means that, since Attributes are necessarily limited, for a given level of investment in defenses the more of one primary defense you build, the less of another you are able to. This stands in contrast to the various secondary defensive layers which are only limited by the bases they appear on as explicits and their position on the passive skill tree, but are at least theoretically accessible to all characters. Block is a great example of a limited secondary defense, being effectively only accessible to melee characters, but providing significant added value to all three primary defensive layers; The type of value Block gives to each form of primary defense is even differentiated by the reason a character using each layer might want to invest in Block.

Why Should Strength Give Armour? Doesn't Armour Suck?
Armour, coming from PoE1, has been designed as a means to mitigate a large number of small physical damage hits associated with large packs of 'trash' mobs. And, it has consistently compared poorly to the other two defensive layers; Uniquely, it only applies to physical damage, with significant tradeoffs or investment required for it to apply to elemental/chaos damage. The effectiveness of Armour as a mitigation also scales with the size of the hit mitigated, not with the level of the area or the monsters within. These combined leave Armour ineffective as a defensive layer exactly where it is needed most to fulfill the fantasy of the Warrior: Trading blows with powerful enemies and coming out on top due to superior endurance and strength.

These issues were patched in PoE1 very quickly. Determination Aura, Grace Aura combined with Iron Reflexes, and Granite Flasks allowed players to stack sufficient Armour to start to mitigate even large hits, and over time the Armour co-efficient was reduced to allow less Armour to mitigate effectively. The implementation and increasing accessibility of endurance charges and Fortification, which applied pre-Armour physical damage reduction as well as increasing elemental resistance and providing flat overall damage reduction, respectively, further improved survivability for Armour-based characters. However, as Fortification became more accessible, it also starting being picked up by all characters as a general secondary defensive layer, leading to the Fortification stacking rework in PoE1 as a way to make it less accessible to ranged characters and spellcasters that it was not designed for. This rework reflects a legitimate concern with overall % damage reduction being accessible to characters that shouldn't have access to it, which I believe is effectively addressed by tying it directly to Strength, which those characters tend to not invest in as it does not help them meet the attribute requirements for their active skills.

Since Determination, Grace, Fortification and Endurance Charges are not in PoE2 and do not provide inherent bonuses, respectively, GGG has a golden opportunity to rework Armour into a new paradigm: Scaling % damage reduction, like Fortification with a modified formula reflecting its role as a primary defensive layer. Flat Armour rating, as a raw stat provided by gear and, under my proposed paradigm, Strength, would be plugged into a formula to spit out a % damage reduction that applies to all hits from enemies (but not non-hit damage or hits that the player deals to themselves from items such as Fireflower) based on the level of the monster dealing the hit.

I cannot determine what such a formula would look like, as I do not have access to GGG's internal math, but whatever the formula is it ought to be pegged such that equivalently accessible amounts of Armour and Evasion would give equivalent levels of EHP. As an example, comparing two characters with 10,000 Armour and 10,000 Evasion, respectively. On the Riverbank, the earliest level zone, 10,000 Evasion would be sufficient to reach the 90% cap, and I would presume that Armour would, under this paradigm, also have the same 90% damage reduction cap that it currently has. Therefore, each character would multiply their damage taken over 10 hits reduced to 10% of that total damage; The Evasion character by only taking 1 of the 10 hits for the full damage, and the Armour character by taking 10% of the damage of each of the 10 hits, taking the same amount of damage that the Evasion character took, but over time instead of one lump. In a hypothetical higher level area, where 10,000 of either defense is only enough for 75% damage reduction or dodge, respectively, the same effect would apply, but at a ratio of 25%, and so on. Armour and Evasion would then be able to be differentiated further by how players choose to mitigate the damage that does come through and what secondary defensive layers are most effective.

Wouldn't Making Armour Fortification Be Too Good?
Armour as direct damage reduction before resistances would be very strong, but that is why I compare it to and believe it should be directly kept even with the EHP increases provided by Evasion. Evasion already provides direct damage reduction, on average, set to a certain percentage based on enemy monster accuracy, which is a direct reflection of enemy monster level. Without Fortification and Endurance charges to further scale the reduction, GGG can essentially lock in how much expected EHP they want characters to have in any given area based on its level by adjusting the formulas for the primary defenses.

I would also expect that, like evasion, there will be and already are methods for powerful enemies to mitigate high levels of Armour stacking and continue to present a threat to players. The 'Splintering' enemy mod that already exists and allows enemies to break Armour in the same way players do would have an escalating effect on players relying on Armour as their primary defensive layer, causing each hit not just from themselves but their minions to deal more and more damage under this paradigm. I would also expect that some boss attacks would overwhelm some or all of a character's Armour, especially if that player gets hit by the center or 'sweet spot' of the attack. For example, the Executioner's big slam I would expect to only apply Armour to the shockwave from the slam, but if the character gets hit by the axe itself, then they would apply significantly less or even no Armour to the hit. This can then be further balanced around by giving players opportunities, similar to the Acrobatics keystone, to provide further mitigation from Armour at the cost of having to make significant further investment.

Where Can We Get Base Life From, If It's Not On Strength? I Want to Invest My Passives In Life!
By moving raw Life out of Strength, there is a second order issue created of players being unable to invest passive points in raw life. While I do think that this ultimately reflects GGG's vision for how players should interact with the skill tree (taking things they want instead of things they feel obligated to take), some players really do want to invest in raw life. My proposed solution to allow players to directly invest in life is to create 12 (and only exactly 12, no more) life rings, situated at each half-hourly position on the skill tree, forming a bridge between the middle ring and the outer ring with the notable for each ring being one or two nodes off of the main bridge. This way players are presented with a choice between taking the efficient route using the attribute highway bridges at each hourly position, or taking a more inefficient route that allows investment in life. This will also ideally help reduce incentives for over-investment by life by limiting the number of rings and spacing them out from one another significantly. This will displace a number of existing bridge rings, but I believe making those current bridge rings attached to only the middle or outer ring at their current attribute node is an acceptable solution.

What About Stat-Stackers?
Stat stacking is a high investment strategy that I think should remain in the game as a source of both survivability and damage for players that are willing to invest in it. GGG has effectively already solved this issue in the Dexterity section with Falcon Technique and Falcon Dive, letting players scale their attack speed with Dexterity and Accuracy (which scales with Dexterity), respectively. I would expect to see, under this paradigm, similar nodes in the Intelligence and Strength sections of the tree allowing for scaling of other offensive abilities like mana, physical damage, and so on associated with those Attributes.

Ok, I Read That Whole Novel, But I Don't Really Get It. Can You Sum Up?
Both players and GGG have consistently expressed a handful of pain points with the way defensive scaling plays out as player builds develop past the early game. GGG as also expressed difficulty pinning down exactly what the Attribute bonus for Attributes should be, especially Dexterity which struggles to establish an identity for itself. My solution to these problems is to 1) establish an identity for each attribute's bonuses in relation to a primary defensive layer associated with that Attribute: Armour for Strength, Energy Shield for Intelligence, and Evasion for Dexterity and to 2) take advantage of Early Access and the lack of damage reduction forms associated with Strength to rework Armour into a form of % damage reduction, comparable to Fortification in PoE1, that scales with monster level at the same rate as the average damage taken over time as Evasion. Taken together, I believe these two suggestions form a relatively elegant solution to both establish a clear mechanical identity for the bonuses each Attribute gives and establish parity between the three primary defensive layers that are given as implicits from gear.

This proposed solution would shoehorn characters even more into specific builds, by forcing them to take only defenses associated with their attribute type. It would also increased the power of stat-stacking builds, which while usually expensive are some of the strongest in the game, by giving them huge defenses with no trade-off, since they want to stat-stack for damage anyway.

"

This proposed solution would shoehorn characters even more into specific builds, by forcing them to take only defenses associated with their attribute type. It would also increased the power of stat-stacking builds, which while usually expensive are some of the strongest in the game, by giving them huge defenses with no trade-off, since they want to stat-stack for damage anyway.

On the first point, you're right; The intent is to provide additional defensive baseline for playing building what they need to unlock the gear they want anyways, but I did forget to mention (and this is on me) that I'd also like to see additional keystones like Iron Reflexes converting Armor into ES and ES into Evasion at whatever the appropriate ratio to be relatively EHP neutral would be to help address the issue of shoe-horning.

Regarding stat stackers...It's a difficult issue, for sure. Though I'd just in general see damage taken out of stats as defenses are put in which would hopefully help mitigate the issue, with additional avenues for further investment in stat based damage available but expensive. Pillar of the Caged God remains a problem however, so that is a legitimate point that I'm not sure how to fix.