State of Set Balance Report

General Discussion
60

I did a partial update on State of Set, including all Barb + Wiz sets, AoV, Akkhan, Crusader No Set, Marauder, Inna, Monk No Set, Inarius, Trag’Oul, Rathma, Necro No Set, and all WD sets except Jade and Helltooth. So: most sets. I’ll try to get to the remainder soon.

Edit: finished updating the rest. So they’re current as of 12/27/2022.

Here are the standings of the current top 10:

Set/Build Peak Adjusted Clear Average Adjusted Clear in top 10 Adjusted 100th place clear
Wiz Tal Rasha 153.7 152.7 150.6
Barb Raekor 151.9 151.3 148.6
Wiz No Set 151.7 149.6 144.0
Necro Trag’oul 151.6 151.0 148.4
Crusader Akkhan 151.2 150.6 146.7
Necro Rathma 151.2 149.8 142.4
DH Marauder 151.0 150.2 147.3
Monk No Set 151.0 150.4 147.6
Necro No Set 150.6 149.8 143.4
Barb No Set 149.0 147.7 138.3

We’ve got 9 sets that can break adjusted 150.0, and 8 that can break 151.0!

Also noteworthy: Wiz No Set (LoD Meteors) is ahead of Trag’Oul. Meteors, in general, are very powerful…the 100th place Tal clear would tie for 9th place with Necro No Set! (both 150.6 adjusted).

4 Likes
61

I’m looking forward to the new PTR invalidating everything we know about set strength. How about you Rage?

1 Like
62

Haha, yeah, of course! Lord only knows when we’ll see it, though.

2 Likes
63

Well you two keep up the enthusiasm, it might be your hard work pays of in case the devs are peaking in to your work.

Merry Christmas.

2 Likes
64

Seems like there may be something wrong with how you are calculating these, unless I’m misunderstanding something. (which is totally possible)

But for instance, Monk peak adjusted clear is 151 for a 2:13 at 9191, and Tal’s is 153.7 for 5:17 @ 9195.

How does that make sense. More than twice as fast, at the same paragon is a lower adjusted clear.

65

These rankings only count non-seasonal power. It’s like the “base” power of each set, not counting how much the seasonal themes jack them up. So basically nothing from Season 27 is included (the only exception is Akkhan, which basically gets no benefit from any of the angelic crucibles).

I did some adjusted clear numbers for all builds in season 27 a few weeks back, you can check that out here: Adjusted Clear numbers for all builds, season 27

A few of these numbers are sure to have changed since then, but I’m sure it’s overall much more in line with what you were expecting.

1 Like
66

Oooh, and that is what I was misunderstanding. doh

Yeah, 160 for monk, that seems better, lol.

2 Likes
67

I was thinking about doing another pass on the season 27 numbers for all builds, and then comparing those to the most recent State of Set numbers to see how many tiers each crucible ability was adding. Any interest in chipping in on that? It would be good to have another pair of eyes and another brain making sure that the correct Angelic power for each clear has been correctly identified…

Also, I sent you a message.

1 Like
68

Hi Rage, sure. I’ll check for the message.

2 Likes
69

Check again, some stuff to consider.

1 Like
70

Is it time to necro an old Rend buff post yet?

71

Hmm… I dunno. I figure that at this point, most of the changes that will show up in the PTR later this month are mostly set already.

If we did manage to convince them Rend needs a buff, I think the most likely thing at this point would be an application of a certain crucible power onto some piece, or set, of gear. And I think we’re probably agreed that that would really not be the best thing… either for Push Rend or for Zbarb.

1 Like
72

post removed by author

73

Yes it would. Every build is capable of the same max clear but hardcore you take more defense whether it be gear, paragon point distribution, or talents. At the very least you are more careful in hardcore. Moving out of damaging effects non hardcore players would normally just stand in. This effects your overall clear potential. Also there are a few other factors that play into hardcore clears. If you die you keep your paragon and items in your stash. So when someone dies they effectively need to use a backup set or start completely over if a back up set has not been established. Either way they take a step back. And lastly there is drastically smaller pool of hardcore players posting clears. This leads to a far lower average clear across the boards.

5 Likes
74

Hey guys, any plans to update this NS set report? Or it is good to use it for some PTR purposes?

75

Yeah, I updated it on 12/27/22, and will probably do a selective one that hits H90 / Raiment / Natalya a few weeks after the new patch goes live.

3 Likes
76

Hey, let’s check in about some stuff related to State of Set… and other stuff…

77

State of Set has been updated! Many thanks to TinneOnnMuin, who helped me with this update.

Here’s the current state of the top 10:

Set/Build Peak Adjusted Clear Average Adjusted Clear in top 10 Adjusted 100th place clear
Wiz Tal Rasha 153.9 153.3 151.3
Necro Trag’oul 152.6 151.3 149.0
Barb Raekor 152.0 151.4 148.7
Wiz No Set 152.0 150.4 144.5
Crusader Akkhan 151.4 150.5 147.3
Necro Rathma 151.2 149.8 142.8
DH Marauder 151.0 150.2 147.3
Monk No Set 151.0 150.4 147.7
Necro No Set 150.8 149.9 144.5
Barb No Set 149.0 147.9 139.1

It has not changed much since last time, though Trag’Oul Necro did jump up an entire tier to take the #2 spot. For the full set of numbers, head to the full spreadsheet.

There are some places where these numbers are sure to be incorrect, i.e. for sets that only recently got buffed (Nats, Unhallowed, PoJ, Raiment, Masquerade, etc) and so haven’t had time to show their strength in a non-seasonal setting.

So: to see what would happen, I did a cross-check against the S28 numbers for each build:

Set State of Set Peak Clear Season 28 Peak Clear Difference S28 Peak minus Average Altar S28PMAA vs State of Set
Barb H90 143.6 154.2 10.6 147.0 3.4
Barb IK 148.6 152.8 4.2 145.6 -3.0
Barb Wastes 148.6 153.6 5.0 146.4 -2.2
Barb MOTE 143.9 152.7 8.8 145.5 1.6
Barb Raekor 152.0 159.2 7.2 152.0 0.0
Barb No Set 149.0 154.2 5.2 147.0 -2.0
Crusader AoV 146.6 153.5 6.9 146.3 -0.3
Crusader Seeker 138.9 144.9 6.0 137.7 -1.2
Crusader Roland 138.3 140.0 1.7 132.8 -5.5
Crusader Invoker 144.7 152.4 7.7 145.2 0.5
Crusader Akkhan 151.4 156.5 5.1 149.3 -2.1
Crusader No Set 147.2 153.4 6.2 146.2 -1.0
DH GoD 144.5 151.7 7.2 144.5 0.0
DH Natalya 148.2 157.7 9.5 150.5 2.3
DH Unhallowed 143.4 152.3 8.9 145.1 1.7
DH Shadow 144.2 151.5 7.3 144.3 0.1
DH Marauder 151.0 157.2 6.2 150.0 -1.0
DH No Set 138.3 135.5 -2.8 128.3 -10.0
Monk PoJ 144.4 151.5 7.1 144.3 -0.1
Monk Inna 145.3 152.4 7.1 145.2 -0.1
Monk Uliana 144.3 152.4 8.1 145.2 0.9
Monk Monkey King 144.4 150.4 6.0 143.2 -1.2
Monk Raiment 140.6 146.3 5.7 139.1 -1.5
Monk No Set 151.0 156.8 5.8 149.6 -1.4
Necro Masquerade 146.9 155.7 8.8 148.5 1.6
Necro Pestilence 135.0 141.8 6.8 134.6 -0.4
Necro Inarius 147.1 155.6 8.5 148.4 1.3
Necro Trag’oul 152.6 159.0 6.4 151.8 -0.8
Necro Rathma 151.2 159.1 7.9 151.9 0.7
Necro No Set 150.8 157.2 6.4 150.0 -0.8
WD Mundunugu 144.9 154.2 9.3 147.0 2.1
WD Arachyr 148.1 156.3 8.2 149.1 1.0
WD Zunimassa 145.3 151.5 6.2 144.3 -1.0
WD Helltooth 142.9 149.8 6.9 142.6 -0.3
WD Jade Harvester 139.0 143.4 4.4 136.2 -2.8
WD No Set 144.1 150.8 6.7 143.6 -0.5
Wiz Typhon 147.4 154.9 7.5 147.7 0.3
Wiz Tal Rasha 153.9 162.3 8.4 155.1 1.2
Wiz Delsere 148.2 152.4 4.2 145.2 -3.0
Wiz Vyr 144.7 153.7 9.0 146.5 1.8
Wiz Firebird 148.0 156.0 8.0 148.8 0.8
Wiz No Set 152.0 159.0 7.0 151.8 -0.2

The 5th column, here, subtracts the average buff from the altar from the top adjusted clear for each set from S28. In other words, it tries to put it in the same ballpark as the numbers from State of Set. I looked at this “average altar buff” 3 ways.

First, just an average of all the numbers in the 4th column (which is the difference between the S28 clears and State of Set). This came out to +6.7 tiers. Second, I averaged the numbers only from builds that were either recently buffed (Nats, Typhon, etc), S-tier (Raekor, Marauder, etc), or the best set for the class (Arachyr). This came out to +7.6 tiers. Third, I looked at just the “S or Best” builds (S tier builds plus Arachyr). This came out to +7.2. So, I settled on +7.2 as a good average figure.

I think the overall takeaway here is that yes, a few builds are stronger than they appear in State of Set, most notably H90 Frenzy, which has barely budged since its damage was doubled (went from peak clear of 143.5 to 143.6). That’s just because people haven’t pushed it hard yet in non-season. In S28, it’s about where you would actually expect it to be.

5 Likes
78

Great minds think alike, I was just looking at this myself. Wanted to get an idea what the average S28 gain is looking like.

DH no set looking pretty sad :frowning:

I’m surprised at the difference between Vyr’s and DMO. Vyr’s benefits strongly from the season theme, but there’s no real reason for DMO to be that far behind the average in season 28. Might just need some more seasonal DMO pushers. Or maybe it’s just Ronin single-handedly dragging the non-season number for DMO up.

2 Likes
79

I’ve actually been pretty surprised all season that the gains are as small as they are. The largest gain for any set that didn’t get buffed is for Vyr, at +9.0. Mundu, Nats, and H90 are all higher than this, but all 3 of these got buffed/reworked.

I guess I overestimated the Altar value when I went over the numbers pre-season…

Yeah, the absolute “floor” for a buff from the Altar ought to be something like +4.5 tiers. That’s about what you get from the stuff that is just “baked in”, like straight buffs to damage, and not reliant on RNG at all (primarily the potion abilities, but also the “extra orb from elites” which can vary a lot from rift to rift).

Yeah, single players can definitely change the numbers a lot. With Raekor, for instance, Enryu has 3 of the top 10 clears. With Rathma, vhsculture has 4 of the top 10, including the top 3. For Delsere, the top 20 has only 11 players, with 3 clears by EasyFrank, 3 by LittlePanna, and 5 by Ronin (and 1 by TinneOnnMuin, of course).

With Immortal King, Seeker, Natalya, Uliana, and LoD WD, the top Adjusted Clear is 2.0+ tiers ahead of the 2nd place clear.

All that is why DMKT and I built in the 3 right-hand columns (showing peak-avg, standard deviation, and avg-100th place). The greener those are, the more accurate the overall prediction is likely to be. That still seems to be working pretty well: the analysis is quite confident (green) for Raekor, Tal, and Arachyr, but not confident at all (orange-red) for, say, Nats.