Analyzing the new Itemization rules

[leftayparxoun]

The plan for a new Itemization revamp was recently announced and this means it's time to dig in the numbers!

EDIT: I missed the notice on the wiki that only item Stat Total is affected by the item Rarity Scalar. As such, most of my observations regarding Armor, Resists and wDps were incorrect. I have edited the post to update those sections, so if this is the first time that you are seeing this EDIT message, feel free to reread the post for the updated info.

WARNING: LONG POST AHEAD
There is no TLDR. I got tired at the end and decided to not make one. Maybe I will edit the post later.

Let me take you along to an in-depth* look into the upcoming itemization changes.
Brace yourself because, in the words of Tom Cardy, "this next one is a bit of a doozy."


Part 1 - The Changes, a Wiki page and Collective confusion

Our first introduction to the upcoming itemization pages comes from the dev post I linked above. However, what a lot of people seem to have missed is that, in that same post, we were linked a brand new wiki page which contains the details of said itemization plan.

In that wiki page we learn all* of the information needed to calculate the new intended stat total of basically any item in the game, however it is a bit disjointed.
As an upside, in that page it is stated that:

This page is for documentation. For application of the system, we refer to a Google Sheet where one can easily find the values of items and stats they are looking for.

As a downside, that Google Sheet is not yet available. So, with the early April patch looming over the horizon, I decided to dive into the calculations myself and see how the new system is looking like in advance. After all, it's better to catch mistakes in the making rather after they have been implemented.

Despite the dev forum post claiming that

If you’re interested in seeing the new math and the values we’re using going forward, you can have a look at this new Wiki Page which contains all the reasoning and the values associated for stats going forward.

there are a lot of missing pieces of information in the page itself. Understandable (due to it being WIP probably) but imo it should have NOT been released in an incomplete form.

Let me begin with some statements regarding the information contained in the wiki page:

• There is no information regarding rounding (up, down, nearest) or at which decimal it is done for each calculation/variable. This is very important in stuff like weaponDps calculations. In the results I will be presenting later, I ASSUMED that all numbers are always rounded to the nearest integer. If that is wrong then various figures I will show you might be off by a couple points.

• There is no information about which Procs belong to what tier (weak/moderate/strong). Granted, they might change if the procs themselves are readjusted numerically, but still.

• There is no information regarding the "worth"/"cost" of passive bonuses like the Sentinel set 7pc bonus or the Tank Darkpromise 7pc Bonus, nor regarding passive bonuses on gear such as the Bloodlord weapons' one or the Blood-Iron gems' one.

For the above, I simply ignored them in my analysis and left them open-ended.

• Moreover, there is also no information regarding the worth/cost of +1 or +2 bonuses in mastery trees from set bonuses (e.g. Warlord/Sov/offSov 6pc) or for the cost of granted abilities (8pc Sov/offSov/Warlord, 7pc Onslaught).

• Generic crit%, as in both heal and offensive, is also not listed in the stat table despite existing in some DOK/WP pieces. It might end up in the chopping block once the rework is done though.

• As for the reasoning behind the Stat Costs for the various stats presented in the Tables 5. and 6. there is no REASONING provided (as was suggested by the dev post). Hopefully it will be added sooner or later, but I will also be discussing it in detail later.

• Finally, I strongly suspect there has been a mistake in Table 1. and specifically in the column providing the Scalars for weapon dps. The values of 10.00 and 14.50 would lead to absurd wDps values if taken as is. For example, the 1h fortress weapon would have ~670 wDps compared to the ~58 they have now. After messing a bit with the numbers I ASSUMED that the intended values are 1/10th of those listed (so 1,00 and 1,45). That led to realistic values for weapons across all tiers. If that assumptions is wrong I will have to manually correct the following results I will be presenting for weapon wDps changes so try to keep that in mind.

Having said all that, it's time to show you the calcs.


Part 2 - Comparing the Formulas with current Items or: How I Learned to Stop Worrying and Love the manual Addition

My calculation file is open for everyone in this link.

• The first sheet ("Formulas") contains the summary of the wiki page, with certain formulas having been simplified for brevity and visual clarity and with certain "comments" of mine (or worries if you'd rather) in red text.

• The second sheet ("Scaling Stats Costs") contains the calculations of the adjusted stat "costs" of the stats whose "value" depends on the Career Rank (CR) Requirement of an item.

• The third sheet ("New theoretical itemization") has the calculated table of each item type's (Armor, jewelry, weapon, shield) stats as a faction of item level (ilvl) and item rarity (Common, Uncommon, Rare, Very Rare, Mythic). I only calculated stuff for Rare (blue) and Very Rare (Purple and Gold) items since these are our main focus when it comes to gearing. The CR Requirements listed on that sheet are not always accurate and can be ignored (in that sheet).

These numbers standalone don't mean too much unless we compare them to the actual currently available gear.
To do that for ALL gear in the game is impossible (for one person and within a day). I deliberately chose to check the most relevant gear for the Choppa class. While it is not a 1-to-1 mirror with other classes (when it comes down to the current stat "cost" of their gear), it is a generally good indication for the changes the new Itemization system will cause.

Originally I was planning to cover 1 tank and 1 healer class too but if you open the calculation file you'll probably understand why I decided against that...

• The fourth and arguably most important sheet is "Choppa comparisons". In it you will find the direct comparisons between the predicted Stat Totals of Choppa gear (according to the information from the Itemization wiki page) and the current stat totals of the same gear (if calculated by the stat "values" introduced in the same wiki page). The former were essentially taken from the 3rd sheet calculations while the latter where manually calculated for each, gear, jewelry, and, weapon, piece (while also adjusting the stat "values" according to the CR requirements of each item).
  
  Besides the stat comparisons, I also chose to calculate the expected Set Bonus values for each set by ASSUMING that all set bonuses are calculated with the same Set lvl; that being the maximum ilvl between the gear pieces comprising the set. For example, for the Conqueror set (Chest = ilvl 53, Gloves = Boots = ilvl 51, Helmet = Shoulders = ilvl 52, Belt = ilvl 50) I assumed that all set bonuses are calculated with a Set lvl of 53. If that assumption was not correct (no info on the wiki) then the actual set bonuses' values of those sets might be a couple points lower.

Here is how it looks for the Choppa Vanq set for example:


It's quite hard to decipher what's going on with a first glance, so, as someone who just spent the better half of his day working on this thing, let me summarize all my findings and try to explain what is going on overall with the new system:


Part 3 - Key Takeaways and tangential thoughts

Let's begin with the most generalized observations one can make:

• Armor values and resistance values remain essentially unchanged, with only endgame sets (Triumphant+) ending up with slightly higher Resistance values (between +10 and +20 per resistance).

• The weapon Dps (wDps) of almost all weapons remains nearly the same as before. On the negative side of things, most of those weapons seem to be overstated according to the new itemization rules. Especially the weapons boasting procs on them (e.g. Subjugator, Fortress, Sentinel Weapons) are quite likely to lose even more stats according to the dev post:

  For Weapons, Procs (the extra trait or effect) historically did not count as having any value or weight in item calculation. This was a bit strange, as having extra damage, absorbs, etc is most certainly valuable for players, so many weapons with Procs will still retain those, but will see their stats reduced to balance this added benefit. Across the board, mid to end-game Weapon stats will generally be going down.

  The specifics are still obscure though.

• Regarding individual set pieces, there is no specific pattern. Some pieces are understated, other overstated and some just about right. The only observable trend is that, on average, the early T4 sets (from Ruin up until Vale-Walker) are slightly understated in their pieces and should be getting stat buffs when reworked.

Keep in mind though that a stat increase is not necessarily a buff (and the opposite goes for a stat decrease too)

This is because not every stat is equal. If a dps gear piece has 30 STR, 20 WS and 10 TOU and that specific piece gets reworked to 25 STR, 15 WS and 30 TOU, that's a net 10 stat point increase which negatively impacts the viability of the dps item. That is to say that even pieces who get their stat total lowered might end up better with better stat allocation (if that is within the scope of the rework).

• Regarding set bonuses, we are informed about the following from the wiki page:

  This value is applied to each bonus on a set, no matter the required number of pieces equipped.

  This essentially means that, unlike currently, all set bonuses within the same set will be standardized at the same value. This introduces a HUGE issue in itemization, but for now I want to focus on the minor implications of it. For almost all sets where the 1st and 2nd set bonus provide raw stats (think Triumphant with +STR and +WOU respectively), the newly standardized value lies somewhere between the 2 former values.
  In the specific example, Triumphant bonuses were +71 STR and +78 WOU and should now both become +74. The same follows for the latter set bonuses.

• Specific pieces and set bonuses that are overstated in a major way compared to the newly enforced stat totals have 2 main causes:
  • The main culprit is Auto Attack Haste/Speed. You will notice that at CRreq = 40 it "costs" 4 stat points per 1%. This means that a piece which has +10% Haste on it (which is a quite common occurrence for physical dps) will basically already have a +40 stat cost on it without considering anything else. To put things into perspective, Vanq gloves/boots now have a 75 Stat Total and you would be spending more than half of it on Haste. This is also an issue in the Redeye set where the current 25% Haste set bonus ends up costing 100 Stat Points (while the set should have 65 in the set bonuses).
  • The second problem-maker is Armor, specifically on set bonuses. It by far exceeds the available Stat Totals for set bonuses in all armor sets it's found (on Redeye it's 140 points and on off-Sov it's 184). Judging how all gear sets will be getting a slight armor increase, it doesn't hurt too much to have those Set Bonuses lowered, but perhaps altering the Armor Stat Cost (as extra stat) could be an option.

• There is also the opposite issue too:
  • Pieces with Dodge or Disrupt on them are quite outdated and in general score quite low for their Stat Total. Replacing Dodge or Disrupt with a combination of both (e.g. 3% Dodge --> 3% Dodge + 3% Disrupt) is an easy way to fix this issue, but I'd perhaps recommend removing them altogether from dps sets.
  • As for set bonuses, there's a lot of cases of current itemization where crit (heal, melee, ranged, magical) or strikethrough does not nearly cover the expected Stat Total. The former cost 10 stat points per 1% while Strikethrough (parry, dodge or disrupt) costs 9 points per 1%. When it comes to high ilvl gear, this will introduce heavy powercreep. Sov now should have 79 points in all set bonuses. That would basically mean having a +8% crit as a 5pc Set Bonus. Even lower sets like Onslaught have 70 points to spare --> +7% crit? Is that intended? Meanwhile, strikethrough is at the opposite end of the spectrum. If you decide to dedicate a set bonus to it (see Invader 4pc on Choppa) then are we really gonna have +8% parry strikethrough as a set bonus? The stat is kinda worthless already (unless you are facing tanks), and +4% is usually way more than enough to reach the strikethrough cap vs most enemies.

Part 4 - Oh Boy...

Will try to keep things brief.

Do you like Westerns? I like Westerns.

THE GOOD

• Normalizing stuff and have rigid guidelines for itemization is a good thing, especially for new player retention. Same for proper documentation of said itemization (although that still needs some work on it, as mentioned earlier).

• Adjusting stat costs of stats that are not bolstered or benefit players less at early levels is a great idea. It should be a good help for mid-tier players while catching up to T4.

• Normalizing set bonuses is a really good way to spice up build variety. That is because set bonuses are usually the #1 factor that makes or breaks a build.

• I personally REALLY like the Talisman correction approach, as it addresses certain issues I had with itemization up until now.

THE BAD

• The BIGGEST issue imo stems from the standardization of gear Set Bonuses. Adjusting simple stat set bonuses like STR, TOU, etc. is trivial. However the problems begin when non-base stats are involved. As mentioned earlier, in this proposed itemization we are looking at +7% and +8% crit chance set bonuses. We are also looking at wonky +range set bonuses. The current +10 ft Invader set bonus for rdps should become +8 ft according to the new rules (I can picture 100 Magus players pulling up their calculator again trying to find out what +25% to 65 + 8 ft is ).
  
  It gets even messier though when it comes down to proc set bonuses (and passives and Mastery Tree bonuses and granted abilities). The wiki states that the cost for procs will be either 10, 20 or 40 depending on the tier of the proc. However, what happens to gear sets whose Set Bonuses are supposed to have total values different that those values (e.g. Sov with +79)?
  - Are the extra points completely wasted?
  - Is it possible (in terms of game-engine) to have more than 1 proc in the same Set Bonus (e.g. 2 strong procs for 80)?
  - Is it possible to have stats added on top of the proc in the set bonus to make up for the difference? (e.g. 1 strong proc + 39 stats = 79)?
  
  The wiki mentions nothing on the matter which is quite concerning.

• The second biggest issue also comes from the standardization of set bonuses (believe it or not). Since all Set Bonuses now do not differentiate between set item rarity or set item type, this introduces a HUGE power creep to the value of Jewelry Sets.
  Let's take the offensive Genesis set for example:
  
  Let's even ignore the individual set pieces for now (which are overall getting major buffs for even meta pieces like the Genesis Fragment.
  Due to the Set Bonuses being dictated by the Set lvl (substituting ilvl), this means that Genesis, being an ilvl 45 set, will have a Set Bonus of 63.
  This essentially means that if nothing changes, the set bonuses would look something like this:
  
  (keep in mind that it is a set with a CR Req of 36, so that's why strikethrough and crit are a bit discounted)
  
  Now you might say that it's still fine (since strikethrough is a bad stat), however I would like to remind you that similar changes would happen to EVERY jewelry set, including the T1-T3 (and later T4) exploration + Book of Boom sets, to the Tome Unlock sets and even to the LOTD sets.
  Plus, take a look at a similarly adjusted Healer genesis set:
  
  
  Granted, the itemization update might be accompanied by a rework of many gear sets, adjusting the type of stats they provide and even their CR requirements and item/set levels. However, due to the overall boost that early T4 sets (ilvls 35 - 55) are receiving by the nature of the itemization formulas, you'd have to SERIOUSLY bump down various jewelry sets (to e.g. ilvl 36) to not seriously devalue standalone jewelries (like the Annuli from the Twillight's Tide event).

and at last


THE UGLY

• Why is Parry and Parry strikethrough worth the same if the latter is subject to a hardcap? At least discount it a bit. For comparison Block costs 12 and Block Strikethrough 6.
  Even worse, in the case of Dodge Strithrough and Disrupt Strithrough, 1% of either also costs 9 stat points, but for 9 points you can get both 1% Dodge and 1% Disrupt.

• AA Haste is basically a much "costlier" stat after the itemization changes. While it, mirroring the AA Damage stat cost makes sense, I'd argue that 4 per 1% is a bit too high. 3 would personally seem more suitable for such a stat.

• For whatever reason, the Twillight's Tide Annuli are ilvl 40 while the Lost Vale Rings (Annulus-lite in stats) are ilvl 50.
  Here is what would happen if the ilvls are not adjusted:
  
  
  Essentially LV rings would significantly outperform Annulus rings, whose only niche would be relegated to having Armor available instead of Resists in Iron rings (Blood-Iron and Bled Blood-Iron rings are also ilvl 43). Hopefully me and a lot of other players did not waste our summer vacation farming Rings that would become outdated 1-2 years later (this point also goes for what will happen with Jewelry sets as mentioned above).

• I would like to also bring your attention to one notorious Jewelry set, the Cursed Rings of the Apostate. That set is most commonly known as the Chaos God band set and is mainly used only as a 1pc for the extra WOU the LV ring brings. Lets take a look at the stats these two rings have now and their predicted stats according to the new itemization:
  
  
  Both rings have a -120 resist total currently. Post rework they would be intended to have 183 and 176 resist totals respectively.
  If the -120 resists are kept, then that would mean that the rings should have a +76 and +74 base stat surplus after their rework (converted via resist stat costs).
  Their set bonus would also be increased.
  Basically, the rings would look something like this:
  
  I probably don't need to explain why neither of them would be balanced.
  
  As a potential solution their resist values could also be increased by a bit to get back to normal base stat amounts and/or their ilvl could(should) drop.

• The total work required to rework even just the most commonly used sets (ignoring Influence/Quest rewards and overworld drops) is INSANE. I spent about 15 h to do everything required for this post and I barely even scratched the surface of 1 class.
  
  While a gradual rollout (weapons first, then e.g. jewelries, then gear sets) makes sense from a working perspective, it leads to 2 very unfortunate consequences:
  • If even just one part of the entire new itemization needs to be reworked due to being unbalanced (e.g. Jewelry set bonuses, or Crit stat costs) then you'll have to REDO everything you have already done. It is much better to finalize the system first and once people have raised their worries and you have taken a good look at the total system again, to then start adjusting items. If anything, you should probably understand that issue firsthand since the Lair items and weapons you reworked some months ago will need another hand to comply with the new system...
  • If you break it down in multiple parts it will introduce changes to the META with every single new patch. This can be a headache for both the balance team since they will be balancing based on incomplete data and will potentially have to keep revisiting previously reworked classes, and also for people who make/update guides on the game (I am begging you here. For the love of god, please don't roll it out in 5 different phases )

That is all from my end for now and, while I would like to hope that those hours I spent were not for nothing, the new system seems to have quite a few "flaws".
To close things up, I would like to dedicate the final section to the itemization dev team and to give a recap on this point and a couple recommendations:


Part 5 - please, please, please, please, please, please, pl...

• The wiki page is quite incomplete:
  • The Google Sheet for calculations is missing.
  • A lot of formulas could be further simplified. Feel free to take a look at the first sheet of my Calculation file for some suggestions.
  • Check Part 1 of my post again for the rest of missing stuff

• PLEASE properly discuss the itemization system before the weapon rework is out (April patch). Less work to redo if so.

• Ideally release the reworked items as bunched up as possible. A gradual rollout will be a big pain both to the balance team and to guide makers.

• PLEASE PLEASE PLEASE PLEASE Document everything PROPERLY. If possible please make copies of the item database from before a large itemization patch rolls out. That way it will be easy for people to look up the item changes without you having to write a 12.000 word changelog each patch. It will be extremely important especially if you choose to...

• ...readjust individual set pieces in terms of stat ratios. As also mentioned by players in the other thread, I would love to see random stats removed from gear with other purposes (e.g. TOU from offensive dps sets, AA Haste from AOE dps sets and/or anticrit from off-sets focusing on armor).

Would love to hear what players have to say about these findings, but this post was mainly aimed towards the devs.
In any case, I'd like to thank everyone for reading my post.

Until next time,

[nocturnalguest]

First of all, very very big "Thank you, sir" for doing all this work. It pinpoint exact flaws and missing pieces of puzzle RoR team gave us.

Your findings shared publicly with google sheet further expose incompleteness and confirm feelings of many who are already convinced just by their "feelings" that most likely nothing good may happen with this "project" going live. It also shows in details why some people are worried just by reading incredibly blurry and ill written RoR team post on forum.

It also adds a bit into bowl of weights of why revamp is better not to happen at all assuming all other factors too (potential RvR reworks, RDPS balance patch, other finetuning etc).

After reading all this i feel that new "system" (if ever sees the light of day) will be no different than rest of itemization "systems" done by different people (each i can name, those are no secret who is behind what exactly) as i strongly believe that current balance team will just manualy "finetune" most META things and create even more outliners instead of reaching initial goals of unification. I highly doubt anyone would launch a single piece of jewelry with +135 ini following published logic or lmao @ curative genesis.

As for the rest, i have nothing to add besides what i mentioned already in original thread (viewtopic.php?t=60487&start=50#p593466). Thankfully leftayparxoun had time to create constructive arguments to support worthless theses of "feelings", "worries" and other things some people voice.

As for the rest of crucial thesis:
a) boost for <40 CR can be done in many other ways, its not essentially necessary to remake items, e.g. bombling made some good suggestions on a matter few years back
b) TTK is already got way too many gaps in different game modes and environments and thus fragile enough to further (in)directly affect it with itemization
c) RvR rework and RDPS rework should have much higher priority then itemization. Need to test out, establish new meta and then finetune it with itemization as one of possible mechanism.

[Blkfiredrag]

Great write up, curious to see how it plays out. Thanks Only

[Lescargo]

I would love to read this if game was not become a crash fest for me since last patch.

I mean there are priorities.

[Culexus]

Incredible work Only. Thanks for taking the time to put this all together.

[leftayparxoun]

I just updated my calculations and the main post because I missed the notice on the wiki that only item Stat Total is affected by the item Rarity Scalar. As such, most of my observations regarding Armor, Resists and wDps were incorrect. I have edited the post to update those sections, so feel free to reread the post/check the calculation sheet for the updated info!