I made a mistake and told Makiah about sludge beating warrior and it got him from 1100 to 270 in less than a day xD
And there I was wasting time on EU, struggling on 1.5k
And I’m over here still trying to make highlander DH work… but I’m stubborn 
(Hint, it’s not good)
3 Likes
Are you that different?
You’ve made plenty of broad statements in the past about pallies being fact deniers and would not. That’s the sort of conclusion one can only make if they only pay attention to the SgtCreamSoda’s of the pally world, but not say… yourself, who I assume you aren’t counting as a fact denier.
Or… are you counting yourself as a fact denying paladin? Then we go back to my question in the beginning 
I’m drastically different.
Want proof? Pick a class. Any class. At some point that class was over tuned and I’ve called it out.
For him, show me a single post from him where he complains Paladin is OP.
I’ll wait.
I mean any deck that can burst damage early is going to have a favorable matchup against warriors. Pushing all the sweepers back a turn makes it nearly impossible for warriors to do anything early. Against sludge it’s probably a worst case scenario with the only real early removal helping sludge decks.
Hunter is doing basically the same thing to the warriors. If the agro priest was still a viable deck the warriors wouldn’t even exist.
Maybe this will force the meta to evolve now.
LOL, he has 72% winrate on sludgelock and is now rank 69 xD
FFS what a highroller
How does that prove anything?
The point of contention, what I quoted from you, is whether you only pay attention to things you want to hear.
How many classes you pay attention to doesn’t say anything about that.
I’m the one waiting bro. You the one who accused him of not paying attention. I’m asking you how can you demonstrate that you are different.
Like… pay attention to what people are writing/saying 
Apparently, and unfortunately, looks like we’ve exterminated all the warriors from higher legend ranks
Back to finding another viable deck, then 
I’m so pissed off by all these streamers saying shaman is still working, while I still haven’t seen a single one of them (or anyone else for that matter) trying to make it work
It’s not working. Stop lying to people. It’s so bad, Norwis played 3 games on Shaman after patch and hasn’t been online since. That’s how bad it is.
Anyway, enough rant about Shaman from a person most vocal about nerfing it.
But the fact stands, we need something viable to counter the druids and rogues on the ladder.
Updated stats
D4-1
| Archetype | WR | Pop |
|---|---|---|
| Zoo Hunter | +9.4% | 11.5% |
| Sludge Warlock | +5.8% | 0.8% |
| Aggro Paladin | +5.4% | 2.9% |
| Highlander Hunter | +3.9% | 0.2% |
| Spell Damage Druid | +3.2% | 1.9% |
| Highlander Warrior | +3.1% | 33% |
| Excavate Paladin | +1.1% | 1% |
| Insanity Warlock | +1.1% | 0.1% |
| Painlock | +0.4% | 0.6% |
| Handbuff Paladin | +0.3% | 0.1% |
| Dragon Druid | -0.1% | 1% |
| Shopper DH | -0.4% | 1.3% |
| Zarimi Priest | -0.5% | 0.7% |
| Snek Warlock | -0.7% | 2.9% |
| Highlander Shaman | -1.2% | 2.2% |
| Secret Hunter | -1.3% | 0.1% |
| Highlander Druid | -1.9% | 6.3% |
| Highlander DK | -2.9% | 1.1% |
| Thief Rogue | -3.1% | 0.9% |
| Elemental Mage | -3.1% | 0.4% |
| Ogre Rogue | -3.2% | 0.1% |
| Plague DK | -3.3% | 4.3% |
| Wheel Warlock | -3.4% | 0.8% |
| Drum Druid | -3.6% | 0.1% |
| Rainbow DK | -3.8% | 3.1% |
| Overheal Priest | -4.1% | 0.2% |
| Mech Warrior | -4.2% | 0.8% |
| Drilling Rogue | -4.3% | 4% |
| Frost Mage | -4.4% | 0.1% |
| Elemental Shaman | -4.6% | 0.1% |
| Rainbow Plague DK | -4.7% | 0.1% |
| Cutlass Rogue | -4.9% | 0.6% |
| Naga Shopper DH | -5.1% | 0% |
| Pip Priest | -5.2% | 0.3% |
| Mystery Egg Hunter | -5.3% | 0.2% |
| Rainbow Mage | -5.5% | 1.7% |
| Coin Rogue | -5.9% | 0.7% |
| Control DK | -6.1% | 0.2% |
| Taunt Warrior | -6.4% | 0.1% |
| Cycle Rogue | -6.8% | 0.2% |
| Rainbow Excavate DK | -7% | 0.3% |
| Highlander Priest | -7.1% | 2% |
| Highlander DH | -7.3% | 0.3% |
| Ramp Druid | -7.4% | 0.1% |
| Handbuff DK | -7.6% | 0.3% |
| Spell Mage | -8.4% | 2.3% |
| Highlander Mage | -8.4% | 0.5% |
| Riff Warrior | -8.4% | 0% |
| Highlander Warlock | -8.6% | 0.4% |
| Highlander Rogue | -9.2% | 0.3% |
| Spell Damage Shaman | -9.8% | 0.2% |
| Lightshow Mage | -9.8% | 0.2% |
| Pirate Rogue | -9.9% | 0.7% |
| Hero Power Druid | -10% | 0.5% |
| Virus Rogue | -10.1% | 0.1% |
| Highlander Paladin | -10.8% | 1.1% |
| Nature Shaman | -10.9% | 0.1% |
| Cycle Odyn Warrior | -11.6% | 0.2% |
| Odyn Warrior | -11.7% | 0.2% |
| Owlonius Druid | -12.4% | 0.2% |
| Excavate Mage | -13.9% | 0.1% |
| Excavate Shaman | -14.4% | 0.1% |
| Orb Mage | -15.6% | 0.1% |
| Big Demon Warlock | -15.6% | 0.1% |
| Rock ‘n’ Roll Warrior | -16.2% | 0% |
| Murloc Shaman | -26.8% | 0.3% |
T1KL
| Archetype | WR | Pop |
|---|---|---|
| Handbuff Paladin | +16% | 0.8% |
| Zarimi Priest | +6.9% | 3.7% |
| Sludge Warlock | +6% | 1.8% |
| Aggro Paladin | +4% | 2% |
| Zoo Hunter | +3.2% | 6.7% |
| Highlander Warrior | +3% | 25% |
| Thief Rogue | +2.7% | 0.6% |
| Coin Rogue | +2.2% | 0.3% |
| Dragon Druid | +2% | 1.6% |
| Mystery Egg Hunter | +1.4% | 0.3% |
| Shopper DH | +0.8% | 0.7% |
| Naga Shopper DH | +0.5% | 0.5% |
| Wheel Warlock | +0.4% | 1.6% |
| Spell Damage Shaman | +0.1% | 1.4% |
| Excavate Paladin | 0% | 0.4% |
| Nature Shaman | -0.2% | 1.2% |
| Spell Damage Druid | -0.3% | 6.1% |
| Overheal Priest | -0.6% | 1.3% |
| Drilling Rogue | -0.8% | 16.2% |
| Painlock | -1% | 1.4% |
| Owlonius Druid | -1.3% | 0.8% |
| Cutlass Rogue | -1.6% | 1.3% |
| Highlander Paladin | -1.9% | 0.5% |
| Highlander Shaman | -1.9% | 3.1% |
| Snek Warlock | -3.1% | 4.5% |
| Rainbow Mage | -3.1% | 0.9% |
| Plague DK | -3.3% | 0.5% |
| Rainbow DK | -3.8% | 2.3% |
| Cycle Rogue | -4.2% | 0.5% |
| Highlander DK | -5.4% | 0.9% |
| Highlander Druid | -5.9% | 2.6% |
| Cycle Odyn Warrior | -6.5% | 0.6% |
| Highlander Warlock | -8.4% | 0.3% |
| Spell Mage | -9% | 3% |
| Pirate Rogue | -11.3% | 1.5% |
| Highlander Priest | -11.7% | 0.6% |
| Mech Warrior | -13.4% | 0.7% |
| Hero Power Druid | -15.4% | 0.4% |
Handbuff Pally seems to be doing godlike in T1KL, but it’s still only 0.8% of games so it is too early to assume that those numbers will hold up. However, with 3.7% of games over four days, I think Zarimi Priest is the real deal in T1KL. Sludgelock is also looking very strong. I think after the next VS report comes out, a lot of Diamond players are going to be familiarized with top Legend meta decks and it’ll shake up D4-1 a lot.
That’s less than 1 game in 100, no wonder I haven’t seen any
With that low of a sample, mathematics does its’ job - the most extreme values always belong to the lowest samples.
That winrate over popularity ratio is literally meaningless. You don’t see a single pally, and when you do, you get scammed and lose or beat him before you find out if he’s handbuff, highlander or flood with a slow hand.
Also, this distribution is a bit sus. Only 6,2% druids? Na-ah. Every 2nd game in the last 24 hours was druid on NA and EU, and my sample is in hundreds of games.
This might be from 2 days ago
1 Like
Nope. Remember, d0nkey data is tracker side only. It literally doesn’t look at what cards the opponent plays at all. The stats are 100% for Handbuff Pally players with the tracker installed, so we know that they were all Handbuff Pally because tracker sees their whole deck.
Again, I’m estimating popularity here by tracker side data. Maybe Druid players disproportionately don’t use trackers. I can normalize for winrate (weighted average of IRL winrate is tautologically 50%), but there’s no way to make such an adjustment for deck popularity.
1 Like
Anyway, in top 1k druid and rogue are proving to be a constant menace
We see less and less warriors as time goes bye as people found counters (I played a dozen of sludgelock mirrors yesterday, for example, and a few today)
In top legend (top 50-100) mostly highlander decks are played - Shaman and Pally. Highlander control Shaman, which sucks against warrior but beats almost everything else, should also rise in popularity once people catch on it and warriors thin out completely.
So spell dmg Druid, scamdrill Rogue, control Shaman, sludgelock and Reno Warrior so far, are the decks most promising to compete for the rock/paper/scissor trio.
Also, haven’t seen a single Zarimi priest on the ladder in 2 days. I’ve seen a few control ones, though
I took the liberty to standardize Scrotie’s list from a few days ago, to show you how the data should REALLY be interpreted if you want it to be useful.
Following is the same table for winrate and popularity for Top 1k legend, with one difference - new column has been added with Winrate*Popularity factor as a measure of true deck’s strength.
Why?
It’s simple. In economy that’s what we do when we’re trying to compare variables along 2 dimensions - we put them all on a same scale/variable/dimension, in this case combined factor of winrate AND populartiy.
Just winrate tells you NOTHING about the deck’s strength, because the LOWER the popularity, the HIGHER the extremes in that variable (the lower its’ predictive power), so the higher the popularity, the HIGHER the predictive power of the deck is.
The real list goes like this (and it fits into my personal experience and data - what a surprise!)
Picture/Table (nicer aesthethically)
https://ibb.co/ZGXdCNX
Table only (bad aesthethically)
Summary
|Deck|Winrate|Pop|W/Pop|
|---|---|---|---|
|HL Warrior|53%|25%|13,25%|
|Drill Rogue|49,20%|16,20%|7,97%|
|Zoo Hunter|53,20%|6,70%|3,56%|
|Tempo Druid|49,70%|6,10%|3,03%|
|Snek Warlock|46,90%|4,50%|2,11%|
|Zarimi Priest|56,90%|3,70%|2,11%|
|HL Shaman|48,10%|3,10%|1,49%|
|BS Mage|41%|3%|1,23%|
|HL Druid|44,10%|2,60%|1,15%|
|Flood Pally|54%|2%|1,08%|
|Rainbow DK|46,20%|2,30%|1,06%|
|Sludgelock|56%|1,80%|1,01%|
|Dragon Druid|52%|1,60%|0,83%|
|Wheellock|50,40%|1,60%|0,81%|
|OTK Shaman|50,01%|1,40%|0,70%|
|Painlock|49%|1,40%|0,69%|
|Overheal Priest|49,40%|1,30%|0,64%|
|Cutlass Rogue|48,40%|1,30%|0,63%|
|Nature Shaman|49,80%|1,20%|0,60%|
|Pirate Rogue|38,70%|1,50%|0,58%|
|Handbuff Pally|66%|0,80%|0,53%|
|Rainbow Mage|46,90%|0,90%|0,42%|
|HL DK|44,60%|0,90%|0,40%|
|Owlonius Druid|48,70%|0,80%|0,39%|
|Shopper DH|50,80%|0,70%|0,36%|
|Thief Rogue|52,70%|0,60%|0,32%|
|Odyn Warrior|43,50%|0,60%|0,26%|
|Mech Warrior|36,60%|0,70%|0,26%|
|Naga Shopper D|50,50%|0,50%|0,25%|
|HL Pally|48,10%|0,50%|0,24%|
|Plague DK|46,70%|0,50%|0,23%|
|HL Priest|38,30%|0,60%|0,23%|
|Cycle Rogue|45,80%|0,50%|0,23%|
|Excavate Pally|50%|0,40%|0,20%|
|Coin Rogue|52,20%|0,30%|0,16%|
|Egg Hunter|51,40%|0,30%|0,15%|
|HP Druid|34,60%|0,40%|0,14%|
|HL Warlock|41,60%|0,30%|0,12%|
Results:
Tier 1 - HL Warrior
Tier 2 - Drill Rogue
Tier 3 - Zoo Hunter, Tempo Druid, Snek Warlock, Zarimi Priest, HL Shaman, Spell Mage, HL Druid, Flood Pally, Rainbow DK, Sludgelock
Judging by the 100% decrease in winrate*popularity factor from HL Warrior to Drill Rogue, it’s safe to say only Highlander Warrior is Tier 1, followed by Drill Rogue alone in Tier 2.
Tier 3 shares a lot of decks, but not as many as tier 4 (every other deck which you can’t see on the picture, but you can see in the table).
1 Like
by the way: if you divide percentage with percentage, you get a factor not a percentage.
I like how you think; I also conjure special factors of that sort; but you have to always be aware those are approximate factors that are always unclear if they mean something exactly for several reasons e.g.:
a) do I just multiply/divide here or do I have to also give more weight to one of them?
b) is this multiplication/division the best or are there other ones being more important?
you can also do that stuff inside tables of mulligan/drawn/kept/etc. btw.
Popularity acts as a weight here - if a deck has a high popularity but lower winrate, it makes the deck stronger than its’ winrate shows, which is more accurate, but not completely accurate (just because people prefer to play Spell Mage doesn’t mean it’s viable, it just means that the deck’s fun over winrate ratio is high)
That’s how classes SHOULD be balanced - balancing by winrate or popularity alone guarantees you lack of balance in the future.
I can’t, because I don’t have the data. I can be thankful for the table Scrotie provided, because that’s the only thing I can work with.
Keep in mind that he previously went through the effort to normalize the stats because they were one-sided tracker data. I don’t even know how I’d go about doing it (nor can I guarantee that he did a good job with it), but I must admit that my table looks much more like everyday ladder situation than his original one does, so his original work must have some merit to it.
If you have access to additional data, please do share them, and I’ll do my magic.
When I mentioned “inside tables” I meant it as a side note for a different purpose. E.g. I took the d0nkey Card Stats for a specific deck and put the mulligan/kept/drawn/count onto a Spreadsheet. I started doing stuff like “mulligan+drawn” or you can use Count multiplied by Drawn to do what you do but at the Card-of-a-Deck-of-the-Patch-level instead of the Deck-of-the-Patch level.
Thanks, I’ll give it a try when I find some free time to play with the stats.
The thing is, I don’t trust those unpaid stats on any sites.
Most of them (all of them) are tracker data so they’re one-sided only, and I don’t know how to normalize them.
On top of that, I don’t really trust drawn/mulligan winrate stats because the game is highly complex and full of synergies between the cards, which means that if a card has 51% mulligan winrate, it doesn’t mean much to me.
For example, it can mean be 51% in general, or it can be 20% in general but 70% when combined with another card it synergises well with.
Or it can be low, because other cards have even higher mulligan rate, so if I wanna see how useful that mulligan rate is I have to compare it to all the other cards in that deck’s mulligan rate.
Or, it can have 51% win rate on average, but 70% winrate in a specific matchup (like how you keep Crescendo against hunters only in mulligan) but I don’t have that data and if I wanted to have that data, I have to search for specific mulligan stats for that specific matchup.
So, hopefully now (you and others) understand how meaningless the data in a vacuum are, when not put in a proper context and compared to other data, as well as how much work there is to analyze and interpret data correctly.
And hopefully, in the future, when I post my experience on the ladder and specify the context, you should all give more weight to it than the simple data in vacuum.
Why do you think you need to normalize them?
I doubt there’s a significant variance between the true stats and samples as large as hs replay uses.
When you try to “normalize” things in stats you often create more problems than you solve.
Outside of blizzard giving us population statistics to compare to the data collection sites’ numbers, we don’t have a clue about what normal looks like.
This is facts. Averages are very sensitive to outliers.
1 Like
Could be true for most of the decks, yeah, but it could really be true that players using the tracker are better players overall (and it makes sense, I wouldn’t be able to play Sludgelock and Rainbow Mage as good without it, for sure) and it could also be true that people who use tracker play a few decks disproportinately (for example, if I’m playing a ramp druid, I really don’t need a tracker, I just ramp and then play my I win cards), while if I’m playing a sludgelock or rainbow mage or something which requires counting, I definitely need that tracker.
So, if we wanna really be precise, we do need some normalization.
I trust Scrotie knows what he’s doing with that because the table I got by analyzing his table conforms to my personal experience on the ladder.
If it didn’t, I would have to assume that he didn’t do his job properly, but I really can’t assume that now.
Personal bias is impossible in my case, because my sample is way too high to allow big mistakes 
Again, it’s about sample size. As large as the data sets are over several days, it’s extremely likely that your numbers are close enough that it doesn’t matter in any real sense.
He doesn’t, not really. He read a wiki and thinks he’s smarter than everyone.
Anecdote is not data.
If you really think this, then I think you should continue to subscribe to uber stats and good luck.