Archmage is NOT the Problem

"

Nerim#4690 wrote:

"

Here are the scaling factors based on your own numbers:

The relationship you showed here with your percentages is exponential.

Consider the following simpler example:

2 (100% increase from last)
4 (100% increase from last)
8 (100% increase from last)
16 (100% increase from last)
32 (100% increase from last)
64 (100% increase from last)

While more extreme, this is the most basic example of exponential growth, and your values reflect what I'm observing.

It's the same idea as compound interest in a bank account, which is definitively exponential.

Saying that it's linear is just plain wrong.

"

Nerim#4690 wrote:

TL;DR: “+N Level to (Cold/Fire/Lightning/All) Spell Skills” resulting in an exponential increase to the base value of skills is overpowered, as nearly all other modifiers are scalar.

Disclaimer: 99% of my time played is on Stormweaver Sorceress. While it’s true that I know very little about the scaling of other classes, I believe that most of what I have here applies to them as well. I also have no experience in PoE 1, so I’m unaware if these problems have existed before.

Let me get straight to the point with the so-called “hot take”:

Archmage is NOT the problem.

Instead, a single modifier is dominating everyone’s itemization choices, warping the in-game economy and significantly stifling build diversity. “+ N Level to (Cold/Fire/Lightning/All) Spell Skills” (“+Levels” for short) is extremely unbalanced and should be removed from the game almost entirely. The only exceptions that should be made are for uniques, such as Fireflower or Levinstone, or notables on the passive skill tree.

To understand why +Levels is such a problem, we must understand the nature of damage scaling with respect to skill level. If you look at nearly any ability in the “Elemental” category on PoE2DB, you’ll notice that skill base damage scales exponentially with skill level.

At lower skill levels, this is less of a problem. After all, exponential scaling is, by nature, less noticeable at lower variable values. However, as skills reach level 20 and beyond, the difference becomes staggering.

However, most (if not all) other scaling is multiplicative; scalar. Consider the following example with the other directly damage-related modifiers: “+% (Cold/Lightning/Fire/Spell) Damage”, using the max value for a single modifier:

Max Level 20 Spark Damage (per projectile)

No Modifier = 229
120% = 503
+1 = 257
+2 = 289
+3 = 324
+4 = 363
+5 = 406
+6 = 455

These values look fine, right? The thing is, these modifiers are not applied in a vacuum during normal gameplay. You must factor in the bonuses provided from the passive tree, which are almost all scalar in nature. These bonuses cause the exponential base scaling from +Levels to be overwhelming. Consider the same example as above, with an easily-attainable total of 100% Increased Spell Damage from the passive tree:

Max Level 20 Spark Damage (per projectile)

No Modifier = 229
120% + 100% = 732
+1 + 100% = 514
+2 + 100% = 578
+3 + 100% = 648
+4 + 100% = 726
+5 + 100% = 812
+6 + 100% = 910

Again, this is with just a 100% passive modifier, not including what comes on gear with other modifiers and from other passives. Consider another 100% bonus from passives and/or other gear modifiers (again, easily attainable):

Max Level 20 Spark Damage (per projectile)

No Modifier = 229
120% + 200% = 961
+1 + 200% = 771
+2 + 200% = 867
+3 + 200% = 972
+4 + 200% = 1089
+5 + 200% = 1218
+6 + 200% = 1365

Thus, it becomes obvious that +Levels is far more impactful than any other single damage-scaling modifier on gear because it exponentially modifies the base damage of the skill. This is reflected by the severe economic skew towards wands, staves, and other weapons with the max +Levels modifier on them.

But this is not the only problem with the +Levels modifier. A not-so-obvious consequence of using +Levels is that the mana costs grow alongside the base damage of the skill at a similar rate. At first glance, you might think that this is the balancing trade-off. I would agree, but I think this makes matters even worse.

Because +Levels is so overwhelmingly strong, players are inclined to build around it and mitigate its sole downside. This has resulted in mana-modifier gear being disproportionately valuable versus other modifiers.

Enter Archmage, a build strategy that encourages the very same mana-modifier stacking that mitigates the downsides of +Levels, while multiplicatively scaling the exponential increases granted by +Levels by a huge amount. This is why Archmage appears to be overtuned: because +Levels benefits from Archmage significantly more than any other single effect or passive.

Now that we’ve addressed most of the problems, let’s talk solutions. I believe the only way to balance +Levels is to remove it from all non-uniques and replace it with additive or multiplicative modifiers.

“Just change the scaling of skill gems!” - At first, this seems like a viable solution. However, it’s likely that GGG has balanced the game progression around the current scaling of skill gems. Changing it could result in a worse leveling and/or endgame experience.

“Reduce the exponential scaling at higher levels!” - I don’t think this is the solution either. The exponential scaling should be something attainable with restriction, such as using a unique or passive tree notable. It should be special and powerful, but not available on every weapon in the game.

I think the best solution is to replace +Levels with a scalar modifier of the same tier. This will cause only a mild diminishment in the economic value of the items, as the modifier slot still provides the same benefit as before: more damage. Scalar modifiers also have a finer distribution, while +Levels only has N possible values.

But please, let me know your thoughts! Do you like +Levels, and believe it should stay? If so, why?

"

Sutopia#6937 wrote:

Problem is obviously MOM.

Anything that can scale both attack and defense is going to be blatantly OP period.

"

laurent67#6296 wrote:

hum, a bit of poe 1 and archmage history maybe ? ( and selfcast problem)

archmage was put ingame for solve an issue, selfcast.

with time each league patch different spell was nerf and werf again.
because trigger, when use trigger ( like cast on, spell totem ..etc) you can did you spell a lot of time per sec, maybe 4,5 or 10 much more than selcast.

result nobody want to play selcast because they have so low dps ( with succesiv nerf), and at the end we struggle to make a red map ( t15 or t16).

the solution ? -> archmage, to buff that's selcast can use in a way that's the trigger way don't work.

so here in poe 2, just for fun , swith off archmage buff and enter you favorite map for fun and chech you dps ..

ps: if want to go deeper, could check the old Enki-arc build guide and their update league post. ( it's a guys who focus to make selfcast build against GGG nerf patch during a lot of years and league, and at the end he FF)

and archmage is really not an issue. just try selfcast spell ....


edit:

just for data, add my stat lvl94 spark ci-mom-eb sorc ( around 30 div gear)

spark tooltip dps with archmage 37k, without 7 k
for context, in t16 +2 i feel ok around 27k dps, and feel nice from 35k

but without archmage, with 30 div gears, i'm not sur to end a t5 white map xD
glhf for have 30 div when end campagne . hahaha

just for saying, if remove archmage base dmg of ALL spell MUST have a MASSIVE BUFF. or all selcast just dead ( like in poe 1)

"

Rotgarg#0212 wrote:

The problem isnt with Archmage or Sorc. It is litterally that there are only a few builds that can get to endgame. Even fewer that can clear trials with moderate ease. I am not worried. This is early access. We are absolutely getting a wipe atleast soon. I want BUFFS to power, not NERFS to others. We are complaining all wrong.

Name one time GGG nerfed something and buffed everything else? Never. Look at melee. Please spread this message on. Buff more classes. No nerfs.