yeah idk
fire felt really really strong if there wasn’t a hunter in the lobby
not in the same way boomkin/ret/ww/warrior felt strong with frontloaded overwhelming damage
but fire really couldn’t die to anything non hunter and would either bait people into thinking they could die long enough to win
or would be left alone long enough for 0 coordination to let a db go uncontested at a point in dampening that it was gg
Thats what everyone plays for though and enjoys. No one is playing a DPS spec to drag people to darkest depths of dampening like oldschool affys and frost. Its not fun or enjoyable to do half the damage or less of everyone else.
This season fire is going to be really strong and a real DPS spec though.
Mage itself mechanically is never gutter tier. How a spec plays though and its damage can feel pretty gutter tier though.
10-14k fireballs, 2k scorchs and 30-50k pyroblast average last season was pretty horrendous for a DPS spec. If I ran rune of power and on use badge it took 9 seconds of someone entirely AFK to die to a combust. Thats atrocious levels of balancing.
If we had the same changes we got this season as last fire would have been a solid A tier all season long. Not OP but decent.
Now I press combust and can wallop someone for half their HP in a global with tinderball fb fb pyro alone.
My fireballs with glass are hitting 80k regular crits 40-45k noncrit, pyros are hitting 50-75k crits, scorch is critting for 50-65k execute range. Like a real DPS spec. Enormous difference. The new reworked glass cannon and 55 yard range and 15% damage buff if no one is near you with flamecannon is huge pressure too.
Can’t even compare them if you were saying fire last season is really strong then it must be ret level rework powerful rn if you compared them side by side.
The DPS side of things fire was gutter tier as heck last season. Mechanically its strong like any other mage spec in terms of control and survivability.
Yeah, I agree that’s what data should be used for, but not the data you’re giving here. You’re assuming the top 50 players are also at an even skill level and the specs have had the same level of balance throughout the season. But this particular data doesn’t take into the whole of the season (it shows no highs or lows) and it doesn’t show when these ratings were achieved and rate of gain for each.
I think Blizzard uses lots of data to make its adjustment decisions (a lot more than this), but I am pretty confident they look at more than the top 0.1% of each spec.
Balancing specs is very difficult because of situations like Fire Mage. You have a spec that needs “a bit more help from its teammates” (ie to be carried or paired with players in RSS that know what to do when teamed WITH Fire Mage but not what to do AGAINST Fire Mage). If its performance relies on teammates, then adjusting OTHER specs is going to affect Fire as well as adjusting Fire Mage itself.
And for Boomie, I don’t think it was this supposed high-performing spec all season. Why would it end with a higher score than it’s supposed direct counter, Ret, when Ret was not only its counter, but the most OP broken RSS spec in the game for over a month (and is still top tier)? It is likely because Ret was super popular and played at the lower levels but a lot of folks got their 2400 and stopped pushing. Others that did push, eventually hit the high skill levels where skill is more important than spec. The result is more players in the Ret pool, which means more in the 0.1%, and more players to lower its average score in that 0.1%. I am fairly confident it is a smaller, and tighter grouped set of players in the Boomie 0.1%.
I’m not saying this isn’t usable data, but it requires a whole lot more context than saying Unholy is 15% worse than Boomie. All adjustments need to take the current iteration of the game in context and be made understanding how they could affect all levels of play.
how to not prove a point with stats 101
Ring of Fire doing some hard carrying here.
Arcane strat: Be rodeo clown until dampening is high, ring of fire, win.
do you think everything will even out if you 10x the sample size?
that specs like demo will become average and specs like fury will move out of dumpster tier?
but it also means that people were able to get arcane to 3500 rating.
i can promise you that you weren’t queing into fury war or unholy dk at 3500 last season as their cap was well below that.
what i find hilarious is that the only people saying the data is invalid are those that are playing over performing specs…
you just can’t make this up lmaooo
Some of that not really applicable to the majority of the player base. Some people will definitely misinterpret the data, either way…
Always personally thought it was odd to use the far right side of the bell curve and toss “average” around lol.
so youre saying that if all the boomy players swapped to fury with the same skill level, the fury cutoff would go from dead last to the top?
no data set will ever be fully accurate but at some point we need to get a grip on reality and understand what’s within reason.
the top fury war peaked at 3300 cr, and that’s after he gapped every other warrior by 300 cr. he’s clearly pushing the spec to its limits and its limit as his score and OP’s data suggest, is welllll below 3500
These stats mean less than a single player’s experience, really.
What you want is to nerf/buff based on large swings in ladder top 90th percentile for a spec, relative to some historical average, not the r1 cutoffs.
Any top player could go through spec by spec and say which high cutoff was based on it being overturned and which is based on the 1-10 people that if they see in a lobby they know they will go 5-1 or 6-0 unless they play around their skill instead of treating them as a botting AI.
And that’s not even getting into 3s balance.
Fistweavers Disc and Preservation are significantly better in shuffle than 3s by design. That design gap should be addressed (have dampening affect pain suppression and barrier, fistweaving healing to be nerfed vs non player damage, buff evoker mana and increase nullifying shroud cooldown to 2 or 3 minutes in shuffle, etc).
They’ve done a very good job in terms of balance pre ret fiasco. I think if they have a rules based approach based on r1 cutoffs balance would get worse not better.
This data isn’t statistically significant in any manner due to the lack of weighting spec participation.
No it’s not this data sucks, doesn’t paint a complete picture at all
You’re telling me Fire Mage out performs Arms? Ok buddy
Yeah I mean, I def didn’t feel like fire was insane last season, but Volkazar definently made it look less oppressive then it has been for a few expacs. But I could name worse specs than fire, simply because fire could still do a burst go into targets, which is more ideal than playing a rot spec like affi in SS
Its funny you say that because caster lobbies with affy are the one thing I dread. Don’t do enough damage to kill them and its hard to win as affy/fire.
I had a lobby with numbing on arcane, dewndoritos on affy and aboni was the SP with rdruid healer a couple months ago. I think renner was the rdruid there and meep was on his mw.
most miserable 0-6 of my RSS season. I’ve had quite a few affy caster cleave lobbies i cant stand them. This season with fire having some dmg it might not be so bad tho.
They may have issues in melee lobbies but I think affy into other casters is a tough time
interesting statistics but you can tune just off of this
Dude, yea fury warrior is an easy example where its not very good.
But by your data, Arms warrior was “average.” But If it only had 3 spots like arcane / fire for example, its cutoff wouldve been 3511 and I can garuntee you that those warriors wouldve qued a bunch more to pass eachother and inflate that cutoff.
Youre ignoring a huge chunk of information. I agree with you they need more frequent small tuning week to week but using cutoff information as a marker of balance is not the way to determine that lol.
Yeah, I figured this would be an issue for some of the less popular specs. Arms out here with like 40 slots and Balance has like 5 and Arcane has 3. That’s…insane. 
Fury is also a weird thing where it was good, got kneecapped, and everybody swapped to Arms so Fury retains its tons of slots from initial participation.
If they did weekly balancing, players couldn’t possibly keep up with the fotm rerolls without buying character boosts!
Maybe then Blizzard could pay their QA team more to afford weekly rebalancing.
I mean this list jives with everything I have seen and experienced in shuffle for the most part so it’s not like it’s totally inaccurate. Bottom line for me is data like this points you in the right direction to look for issues and experience let’s you make the final decision.