So I fed all Mar 15 Vicious Syndicate Standard data into an Excel spreadsheet that models future meta shifts — for each simulation cycle, decks with winrates below 50% lose players at a rate proportional to how much below 50% the winrate is, while decks with winrates above 50% gain those players proportional to how much above 50% their winrate is.
Given that, here are my predictions:
1. Warlock, Hunter and Mage numbers will continue to decline, while Priest numbers grow. Eventually Priest would grow to 40% of the meta.
2. Spiteful Priest is in decline and would eventually be forced out of the meta completely. Priest players will abandon it for other Priest archetypes.
3. Overall Paladin numbers will remain constant but in 2-3 weeks virtually all Murlock Paladins will become Dude Paladins, continuing the trend of the last few weeks.
4. Big Spell Mage will hold onto its niche, not really growing in popularlity but holding steady at ~3% total meta population. Meanwhile, Secret Mage will continue to decline as its prey deck, Control Warlock, loses meta dominance.
5. Spell Hunter isn't even viable. People will realize this.
I don't think your extrapolations are going to pan out considering there are only 3 weeks left of this meta, the trends you see on VS are going to change substancially with the release of the Witchwood. Also, playrate data of decks is not always monotonic. You need to take into account that people switch decks to counter the most-played deck during a certain week, making the representation of certains deck fluctuate considerably. If Priest rises, people are going to play the counter deck, so less people are going to go with priest as an option because the meta is heavily teched against it. That's how countering works.
I couldn't model archetypes where I didn't have data, and due to insufficient games recorded there were many archetypes Vicious Syndicate doesn't release matchup data on. I was only able to model 11 archetypes: Big Priest, Combo Priest, Control Priest, Spiteful Priest, Dude Paladin, Murlock Paladin, Control Warlock (this includes Cubelock by Vicious Syndicate definitions), Zoo Warlock, Big Spell Mage, Secret Mage and Spell Hunter. Vicious Syndicate published partial matchup data on several other archetypes, but my modeler would not work if I added a 12th archetype that had win% for Control Warlock but didn't have one for Big Priest.
While this is admittedly only 75% of the overall meta population, there simply isn't good data I can access to attempt to model that remaining 25%. They were treated by the modeler as if they didn't exist.
I couldn't model archetypes where I didn't have data, and due to insufficient games recorded there were many archetypes Vicious Syndicate doesn't release matchup data on. I was only able to model 11 archetypes: Big Priest, Combo Priest, Control Priest, Spiteful Priest, Dude Paladin, Murlock Paladin, Control Warlock (this includes Cubelock by Vicious Syndicate definitions), Zoo Warlock, Big Spell Mage, Secret Mage and Spell Hunter. While this is admittedly only 75% of the overall meta population, there simply isn't good data I can access to attempt to model that remaining 25%.
This is very interesting tho the new exp will shake things up. Im already seeing less Spiteful priest and more combo steal priest with cabal etc just to counter Warlock
1) I don't think your extrapolations are going to pan out considering there are only 3 weeks left of this meta, the trends you see on VS are going to change substancially with the release of the Witchwood. 2) Also, playrate data of decks is not always monotonic. You need to take into account that people switch decks to counter the most-played deck during a certain week, making the representation of certains deck fluctuate considerably. If Priest rises, people are going to play the counter deck, so less people are going to go with priest as an option because the meta is heavily teched against it. That's how countering works.
Numbers mine.
1) True enough. But I just created the modeler three days ago and spent a while testing it with known asymmetric metagames to verify that it works properly.
2) That's EXACTLY how the modeler works. The "top decks" in terms of population become a target for decks with high winrate but low population, so the modeler moves population from decks with <50% winrate to the counter decks.
To recap: Control Priest gets a large population gain, Combo Priest and Dude Paladin get medium population gains, Big Priest gets a small population gain, Big Spell Mage neither gains nor loses population, Control Warlock loses some but not all of its population, and five archetypes are eventually phased out of the meta completely: Spiteful Priest, Murlock Paladin, Secret Mage, Spell Hunter and Zoo Warlock. Of the two classes with at least one gaining and one losing archetype, Priest gains massive amounts of population while Paladin numbers remain constant(ish).
Big Priest from 2.63% to 2.82%. I predicted a small increase, and it happened.
Combo Priest from 6.1% to 7.63%. I predicted a medium increase, and it happened.
Control Priest from 4.34% to 3.94%. I predicted a large increase and got a small decrease instead.
Spiteful Priest from 7.54% to 7.28%. I predicted a large decrease and got a small decrease instead.
Overall Priest population increased, as predicted.
Dude Paladin from 8.49% to 8.55%. I predicted a medium increase and got a very small increase (basically no change) instead.
Murlock Paladin from 8.05% to 8.00%. I predicted a large decrease and got a very small decrease (basically no change) instead.
Overall Paladin population remained nearly constant, as predicted.
Control Warlock from 16.06% to 15.75%. I predicted a medium decrease and got a very small decrease instead.
Zoo Warlock from 2.99% to 2.98%. I predicted a large decrease and got the tiniest possible decrease (basically no change) instead.
Secret Mage from 10.99% to 11.04%. I predicted a large decrease and got a tiny increase (basically no change) instead.
Big Mage from 2.97% to 3.19%. I predicted no change and got a small increase reasonably consistent with that prediction.
Spell Hunter from 4.14% to 4.41%. I predicted a large decrease and got a small increase instead.
——————————————————
If we account for players with losing winrates being less likely to want to craft new archetypes this late before a new expansion is released, the two main predictions I got wrong were Control Priest and Spell Hunter.
I still contend that Control Priest is the meta breaker for the current meta and should be played much more than it currently is, but it hasn't gotten the coverage it deserves in commentary from sites like Vicious Syndicate and Tempo Storm.
I can't understand why people are going to Spell Hunter instead of away from it.
Big Priest from 2.63% to 2.82%. I predicted a small increase, and it happened.
Combo Priest from 6.1% to 7.63%. I predicted a medium increase, and it happened.
Control Priest from 4.34% to 3.94%. I predicted a large increase and got a small decrease instead.
Spiteful Priest from 7.54% to 7.28%. I predicted a large decrease and got a small decrease instead.
Overall Priest population increased, as predicted.
Dude Paladin from 8.49% to 8.55%. I predicted a medium increase and got a very small increase (basically no change) instead.
Murlock Paladin from 8.05% to 8.00%. I predicted a large decrease and got a very small decrease (basically no change) instead.
Overall Paladin population remained nearly constant, as predicted.
Control Warlock from 16.06% to 15.75%. I predicted a medium decrease and got a very small decrease instead.
Zoo Warlock from 2.99% to 2.98%. I predicted a large decrease and got the tiniest possible decrease (basically no change) instead.
Secret Mage from 10.99% to 11.04%. I predicted a large decrease and got a tiny increase (basically no change) instead.
Big Mage from 2.97% to 3.19%. I predicted no change and got a small increase reasonably consistent with that prediction.
Spell Hunter from 4.14% to 4.41%. I predicted a large decrease and got a small increase instead.
——————————————————
If we account for players with losing winrates being less likely to want to craft new archetypes this late before a new expansion is released, the two main predictions I got wrong were Control Priest and Spell Hunter.
I still contend that Control Priest is the meta breaker for the current meta and should be played much more than it currently is, but it hasn't gotten the coverage it deserves in commentary from sites like Vicious Syndicate and Tempo Storm.
I can't understand why people are going to Spell Hunter instead of away from it.
I see now about 50/50 control and spiteful priest and VERY rarely see a Big Priest. Spiteful still has the same problems it's always had. If you don't draw Dr 6 or Duskbreakers or the dragons to pop them the deck doesn't work. I believe that's why Control priest has gained in popularity to match it now.
I don't get the attraction to Spell Huntard either. It's really not that good. Although I did have two games in a row that I played against them the other day where they coined out barnes on 3 to get y'shaarj. Maybe it's because of those cheesy fast wins that people like it?
Big Priest from 2.63% to 2.82%. I predicted a small increase, and it happened.
Combo Priest from 6.1% to 7.63%. I predicted a medium increase, and it happened.
Control Priest from 4.34% to 3.94%. I predicted a large increase and got a small decrease instead.
Spiteful Priest from 7.54% to 7.28%. I predicted a large decrease and got a small decrease instead.
Overall Priest population increased, as predicted.
Dude Paladin from 8.49% to 8.55%. I predicted a medium increase and got a very small increase (basically no change) instead.
Murlock Paladin from 8.05% to 8.00%. I predicted a large decrease and got a very small decrease (basically no change) instead.
Overall Paladin population remained nearly constant, as predicted.
Control Warlock from 16.06% to 15.75%. I predicted a medium decrease and got a very small decrease instead.
Zoo Warlock from 2.99% to 2.98%. I predicted a large decrease and got the tiniest possible decrease (basically no change) instead.
Secret Mage from 10.99% to 11.04%. I predicted a large decrease and got a tiny increase (basically no change) instead.
Big Mage from 2.97% to 3.19%. I predicted no change and got a small increase reasonably consistent with that prediction.
Spell Hunter from 4.14% to 4.41%. I predicted a large decrease and got a small increase instead.
——————————————————
If we account for players with losing winrates being less likely to want to craft new archetypes this late before a new expansion is released, the two main predictions I got wrong were Control Priest and Spell Hunter.
I still contend that Control Priest is the meta breaker for the current meta and should be played much more than it currently is, but it hasn't gotten the coverage it deserves in commentary from sites like Vicious Syndicate and Tempo Storm.
I can't understand why people are going to Spell Hunter instead of away from it.
I see now about 50/50 control and spiteful priest and VERY rarely see a Big Priest. Spiteful still has the same problems it's always had. If you don't draw Dr 6 or Duskbreakers or the dragons to pop them the deck doesn't work. I believe that's why Control priest has gained in popularity to match it now.
I don't get the attraction to Spell Huntard either. It's really not that good. Although I did have two games in a row that I played against them the other day where they coined out barnes on 3 to get y'shaarj. Maybe it's because of those cheesy fast wins that people like it?
The data shows Control Priest lost population in the past week or so, so it hasn't gained in popularlity, at least not on ladder (I'm aware that some tournament players have ran it). That's what I don't get, it has all the right good matchups — it's positive against Dude Paladin and Control Warlock, plus also Murlock Paladin, Zoo Warlock, Secret Mage and Spell Hunter. When your only bad matchups are other Priests and Big Spell Mage, you're in a very good place.
As far as Spell Hunter goes, Priest does the Barnes thing better in Big Priest. However, Big Mage does the Spiteful thing better and people still play Spiteful Priest so whatever.
EDIT: I just noticed Vicious Syndicate posted enough matchup data in #85 for me to model Jade Druid. I'll add that deck to the mix and see what happens.
Spell hunter : There is a lot of Paladin and Mages in the meta, and Spell hunter has a positive winrate against both (every variation of those by the way), as well as combo priest and zoo warlock. Its bad matchups are not horrible (40-45%), and it crushes rogue (even if it is off-meta). I think you put the wrong matchup data in your simulator.
Control priest : It is a higher-skill deck (compared to most meta decks), thus has a harder time gaining popularity (and if it does, it would decrease in winrate).
Finally, since rotation happens in less than a month, players do not want to craft cards in order to pick a better deck, thus more staleness.
You're simulation seems to be more simple than it should. If control warlock goes up, dude paladin goes down.
There has to be the same people winning than losing, so it's not just increase people playing base on the deck their playing. If you were to keep simulating long enough, all people would be playing +50% winrate decks, which is stupid.
Interresting, however, there is a considerable human factor as well. Kingsbane rogue for example has had a stable and high playrate despite an overall winrate around 45%, and some absolutely terrible matchups.
Also, game length can be almost as important as winrate. For climbing, sacrificing a couple of % in winrate can be worth it if the games are twice as fast.
If you were to keep simulating long enough, all people would be playing +50% winrate decks, which is stupid.
This is clearly impossible. The global average winrate of Ranked play is always exactly 50%, so unless every archetype has a 50% winrate there will always be winners and losers.
game length can be almost as important as winrate. For climbing, sacrificing a couple of % in winrate can be worth it if the games are twice as fast.
From what I can tell, no. Sacrificing a fraction of a single percent, maybe, but I don't think there was any point at which any deck achieved a global 54% or higher winrate against the field. (This may have been because I didn't bother modeling the archetypes that make up the bottom 25% of the meta.) Sacrificing 2% winrate from 53% isn't a winning proposition.
If you were to keep simulating long enough, all people would be playing +50% winrate decks, which is stupid.
This is clearly impossible. The global average winrate of Ranked play is always exactly 50%, so unless every archetype has a 50% winrate there will always be winners and losers.
did you read what I wrote? just before that quote I stated the same...
I love to analyze stats as much as the next guy. And this is interesting. But humans can be very tricky to predict sometimes haha.
As for the reason people play spell hunter, wouldn't it simply be because... They enjoy playing it? I still play lots of miracle rogue (I think with a 57% win rate this season in 40 ish games). I'm not sure where miracle rogue fits in your simulation (obviously one of the bottom 25% or whatever). I still switch in murloc paladin and secret mage for quest reasons.
Obviously I'm just one guy lol, but I simply have more fun playing the "off-meta" miracle deck, compared to playing a faster deck. People like to win, but people like to also optimize their "fun factor", if you will, while winning. I feel like this whole prediction kinda ignores that.
Rollback Post to RevisionRollBack
Twitch name: Anatak15 NA Legend Season 48, 49, 52, 53, 54, 74
To post a comment, please login or register a new account.
So I fed all Mar 15 Vicious Syndicate Standard data into an Excel spreadsheet that models future meta shifts — for each simulation cycle, decks with winrates below 50% lose players at a rate proportional to how much below 50% the winrate is, while decks with winrates above 50% gain those players proportional to how much above 50% their winrate is.
Given that, here are my predictions:
1. Warlock, Hunter and Mage numbers will continue to decline, while Priest numbers grow. Eventually Priest would grow to 40% of the meta.
2. Spiteful Priest is in decline and would eventually be forced out of the meta completely. Priest players will abandon it for other Priest archetypes.
3. Overall Paladin numbers will remain constant but in 2-3 weeks virtually all Murlock Paladins will become Dude Paladins, continuing the trend of the last few weeks.
4. Big Spell Mage will hold onto its niche, not really growing in popularlity but holding steady at ~3% total meta population. Meanwhile, Secret Mage will continue to decline as its prey deck, Control Warlock, loses meta dominance.
5. Spell Hunter isn't even viable. People will realize this.
How about Warrior?
I don't think your extrapolations are going to pan out considering there are only 3 weeks left of this meta, the trends you see on VS are going to change substancially with the release of the Witchwood. Also, playrate data of decks is not always monotonic. You need to take into account that people switch decks to counter the most-played deck during a certain week, making the representation of certains deck fluctuate considerably. If Priest rises, people are going to play the counter deck, so less people are going to go with priest as an option because the meta is heavily teched against it. That's how countering works.
So according to your "numbers", pretty much everything is gonna suck. lol
Bad things happen when you use Excel to do math.
Vicious Syndicate Data Reaper #85 is out.
Big Priest from 2.63% to 2.82%. I predicted a small increase, and it happened.
Combo Priest from 6.1% to 7.63%. I predicted a medium increase, and it happened.
Control Priest from 4.34% to 3.94%. I predicted a large increase and got a small decrease instead.
Spiteful Priest from 7.54% to 7.28%. I predicted a large decrease and got a small decrease instead.
Overall Priest population increased, as predicted.
Dude Paladin from 8.49% to 8.55%. I predicted a medium increase and got a very small increase (basically no change) instead.
Murlock Paladin from 8.05% to 8.00%. I predicted a large decrease and got a very small decrease (basically no change) instead.
Overall Paladin population remained nearly constant, as predicted.
Control Warlock from 16.06% to 15.75%. I predicted a medium decrease and got a very small decrease instead.
Zoo Warlock from 2.99% to 2.98%. I predicted a large decrease and got the tiniest possible decrease (basically no change) instead.
Secret Mage from 10.99% to 11.04%. I predicted a large decrease and got a tiny increase (basically no change) instead.
Big Mage from 2.97% to 3.19%. I predicted no change and got a small increase reasonably consistent with that prediction.
Spell Hunter from 4.14% to 4.41%. I predicted a large decrease and got a small increase instead.
——————————————————
If we account for players with losing winrates being less likely to want to craft new archetypes this late before a new expansion is released, the two main predictions I got wrong were Control Priest and Spell Hunter.
I still contend that Control Priest is the meta breaker for the current meta and should be played much more than it currently is, but it hasn't gotten the coverage it deserves in commentary from sites like Vicious Syndicate and Tempo Storm.
I can't understand why people are going to Spell Hunter instead of away from it.
Spell hunter : There is a lot of Paladin and Mages in the meta, and Spell hunter has a positive winrate against both (every variation of those by the way), as well as combo priest and zoo warlock. Its bad matchups are not horrible (40-45%), and it crushes rogue (even if it is off-meta). I think you put the wrong matchup data in your simulator.
Control priest : It is a higher-skill deck (compared to most meta decks), thus has a harder time gaining popularity (and if it does, it would decrease in winrate).
Finally, since rotation happens in less than a month, players do not want to craft cards in order to pick a better deck, thus more staleness.
You're simulation seems to be more simple than it should. If control warlock goes up, dude paladin goes down.
There has to be the same people winning than losing, so it's not just increase people playing base on the deck their playing. If you were to keep simulating long enough, all people would be playing +50% winrate decks, which is stupid.
Interresting, however, there is a considerable human factor as well. Kingsbane rogue for example has had a stable and high playrate despite an overall winrate around 45%, and some absolutely terrible matchups.
Also, game length can be almost as important as winrate. For climbing, sacrificing a couple of % in winrate can be worth it if the games are twice as fast.
Editor of the Heartpwn Legendary Crafting Guide:
https://www.hearthpwn.com/forums/hearthstone-general/card-discussion/205920-legendary-tier-list-crafting-guide
I love to analyze stats as much as the next guy. And this is interesting. But humans can be very tricky to predict sometimes haha.
As for the reason people play spell hunter, wouldn't it simply be because... They enjoy playing it? I still play lots of miracle rogue (I think with a 57% win rate this season in 40 ish games). I'm not sure where miracle rogue fits in your simulation (obviously one of the bottom 25% or whatever). I still switch in murloc paladin and secret mage for quest reasons.
Obviously I'm just one guy lol, but I simply have more fun playing the "off-meta" miracle deck, compared to playing a faster deck. People like to win, but people like to also optimize their "fun factor", if you will, while winning. I feel like this whole prediction kinda ignores that.
Twitch name: Anatak15
NA Legend Season 48, 49, 52, 53, 54, 74