If you mean the class winrate percentages on the homepage, those are worth little to nothing. Pretty sure those include all decks of each class, including shit-tier meme decks.
Specific deck winrate percentages can be useful though.
Also remember when you are looking at decks that can be complicated/hard to play or take time to learn that will effect the win rate over all. So a lot of times you are better off filtering out lower levels.
In addition when a new release happens, many players that are not having luck with a deck will switch to the new flavor of the week, leaving the players that know how to use a deck better using it. Another filter to use that can help is to limit the time frame of your search. This will also help to know what is "working" for people currently.
Don't forget to check your sample size as well to see how popular the deck is. This is a double edged sword. As deck gets more popular people will be drawn to it so you will have less talented players and people learning the deck using it. This can actually lower the win rate of the deck.
Yeah I understand that. I don't have a payed account on HS replay, so I won't be able to see those stats. I'm just very curious about the winrates of each class between rank 1 - 5 and the legendary ranks.
They are useful but do not account for micro-metas of specific ranks and servers. These can change quickly too! If you play the a popular deck witha a high winrate expect to be hardcountered...
They are useful but do not account for micro-metas of specific ranks and servers. These can change quickly too! If you play the a popular deck with a high winrate expect to be hardcountered...
Even without a payed account it's quite useful, just look at specific decks with a large sample (high number of games played). It's interesting to see, when a deck has few games, how daily winrate jumps sometimes from 40 to 90% in different days, because numbers are so small, that the luck of the few players playing that exact deck have a big impact on the overall statistics. But for popular decks with many played games, statistics are rock solid.
The winrate is a usefull number but like some said before it's the meta you are in that matters the most
eg; if you are playing a 60% winrate deck but the the most popular deck in the meta is a bad matchup for you, you'll prolly won't make the 60%.
On the other hand, you can have a 50% winrate deck and get mostly good matchups and you might get to 65%.
I have a 78% winrate with my big bad shaman because I played it the first 2 days of release, people experimenting and expecting murloc shaman makes me a better player. But if i'd retry now i'd be lucky to hit ~50%
At all ranks : Classes with popular meme archetypes or overrated bad archetypes will be undervalued
Say there is a tempo rogue that's really OP, and a Tess rogue that's really popular (more popular than it is now). Tess rogue is going to drag the entire class winrate down, making the class look worse than it really is
Alternatively, a streamer came up with a new pogo rogue deck and convinced everyone it's really good when it actually isn't. This will have the same effect
If you are at a high rank : Classes with difficult to pilot archetypes will be undervalued; easy to pilot classes will be overvalued
Unless you pay for premium, HSReplay doesn't discriminate between ranks. At high ranks, pilot error is less common than the average, therefore decks that are difficult to pilot correctly and decks that are impacted harder by errors are better than they appear in the stats
Conversely, simple to pilot decks are as "powerful" as they are at lower ranks but will lose more often as they face opponents that are more powerful than their lower ranked counterparts
If you are at a low rank : Classes with easy to pilot archetypes will be undervalued; difficult to pilot classes will be undervalued
You need to balance the winrate with the games played too. Some decks have high winrate 60% but only 400 games played. I use the stats but try to balance the games with how often they're seen. If I was picking a deck to play I'd rather play a 55% winrate deck with 20k games played than a 60% winrate but minimal games to prove it.
At all ranks : Classes with popular meme archetypes or overrated bad archetypes will be undervalued
Say there is a tempo rogue that's really OP, and a Tess rogue that's really popular (more popular than it is now). Tess rogue is going to drag the entire class winrate down, making the class look worse than it really is
Alternatively, a streamer came up with a new pogo rogue deck and convinced everyone it's really good when it actually isn't. This will have the same effect
If you are at a high rank : Classes with difficult to pilot archetypes will be undervalued; easy to pilot classes will be overvalued
Unless you pay for premium, HSReplay doesn't discriminate between ranks. At high ranks, pilot error is less common than the average, therefore decks that are difficult to pilot correctly and decks that are impacted harder by errors are better than they appear in the stats
Conversely, simple to pilot decks are as "powerful" as they are at lower ranks but will lose more often as they face opponents that are more powerful than their lower ranked counterparts
If you are at a low rank : Classes with easy to pilot archetypes will be undervalued; difficult to pilot classes will be undervalued
The inverse of the above
These are all valid points, and there are more to be had as well. All it takes is for a Twitch streamer to play a deck and you will see it pop up everywhere, even if the deck has a terrible win rate. All of these factors skew the results.
What's worse is that if a streamer is at high ranks, then decides to play 1 trick pony meme decks like Disguised Toast, those are going to really affect the stats across the board as he is often at legend rank. And once you are at legend, many people stop trying and then play 'fun' decks. It's really hard to measure everything. There is enough data out there and enough serious players that never play meme decks and only play competitive stuff, but how do we know those numbers?
I say the numbers are good guidelines and starting points, but never use them to decide how to spend your dust and craft cards. You should be crafting cards for decks that YOU like to play, not what is popular or has a good win rate. And just because a deck has a good win rate, it doesn't mean that you will, especially at non legend ranks.
I think it's always close to 50/50. Those numbers can vary so wildly (and for many reasons, like even the time you play on or your region can affect win rates and thus those numbers) that it's not worth even looking at them.
Play what you like and what feels lucky during that specific moment. Over a very large sample size, decks should get close to 50/50.
Over a small sample size, it matters more to have fun, i.e. if you win 3 out of 4 matches with a deck while playing 1-2 hours in total, then don't change it until you stop playing (or, of course, until you become unlucky). If, instead, you win 1 in 4 matches after 10-15 games, then maybe try another archetype/etc, which might be luckier at that specific moment in time.
Nevertheless, to have a real chance at having fun you need to use the best decks available in the meta.
FWIW - Iksar has often cited HSReplays data when posting on the subreddit. Presumably, the data is pretty legit.
I'd advise not paying too much attention to the HSReplays homepage - overall class win-rates are essentially meaningless. Mike Donais was a guest on the Omni podcast about a year ago - he explained why class win-rates don't reflect game balance very well, since lots of people play good decks poorly, or simply enjoy playing shitty decks (his example was Mill Rogue - at the time, about 10% of the Rogue player-base was tanking the overall class win-rate by 2%, simply by playing the deck.) From the point of view of game balance, the devs don't look at class win-rates, and focus on whether or not each class and archetype are viable.
As far as Priest goes - if you look deeper at the numbers, about half of all Priest decks are Resurrect Priest. It has a win-rate just below 49%. All the rest of the Priest decks are as shitty as the OP seems to think, with win-rates bottoming out well below 40% - the net result is a class win-rate of 44%.
Rollback Post to RevisionRollBack
To post a comment, please login or register a new account.
So... How trustworthy are these?
I mean, before RoS priest had a winrate of like 44%? Priest was seen very often at rank 5 upwards.
I myself usually play Priest.
After RoS released, Priest is pretty much dead. However, the win% only dropped 1%.
I just can't believe those stats personally...
If you mean the class winrate percentages on the homepage, those are worth little to nothing. Pretty sure those include all decks of each class, including shit-tier meme decks.
Specific deck winrate percentages can be useful though.
Also remember when you are looking at decks that can be complicated/hard to play or take time to learn that will effect the win rate over all. So a lot of times you are better off filtering out lower levels.
In addition when a new release happens, many players that are not having luck with a deck will switch to the new flavor of the week, leaving the players that know how to use a deck better using it. Another filter to use that can help is to limit the time frame of your search. This will also help to know what is "working" for people currently.
Don't forget to check your sample size as well to see how popular the deck is. This is a double edged sword. As deck gets more popular people will be drawn to it so you will have less talented players and people learning the deck using it. This can actually lower the win rate of the deck.
Yeah I understand that. I don't have a payed account on HS replay, so I won't be able to see those stats.
I'm just very curious about the winrates of each class between rank 1 - 5 and the legendary ranks.
They are useful but do not account for micro-metas of specific ranks and servers. These can change quickly too! If you play the a popular deck witha a high winrate expect to be hardcountered...
Editor of the Heartpwn Legendary Crafting Guide:
https://www.hearthpwn.com/forums/hearthstone-general/card-discussion/205920-legendary-tier-list-crafting-guide
They are useful but do not account for micro-metas of specific ranks and servers. These can change quickly too! If you play the a popular deck with a high winrate expect to be hardcountered...
Editor of the Heartpwn Legendary Crafting Guide:
https://www.hearthpwn.com/forums/hearthstone-general/card-discussion/205920-legendary-tier-list-crafting-guide
U have to look at the time frame, old lists that arent played will still show the old winrates
Even without a payed account it's quite useful, just look at specific decks with a large sample (high number of games played). It's interesting to see, when a deck has few games, how daily winrate jumps sometimes from 40 to 90% in different days, because numbers are so small, that the luck of the few players playing that exact deck have a big impact on the overall statistics. But for popular decks with many played games, statistics are rock solid.
The pleasure is mine.
All you need yo know is the HS development team looks at those numbers
The winrate is a usefull number but like some said before it's the meta you are in that matters the most
eg; if you are playing a 60% winrate deck but the the most popular deck in the meta is a bad matchup for you, you'll prolly won't make the 60%.
On the other hand, you can have a 50% winrate deck and get mostly good matchups and you might get to 65%.
I have a 78% winrate with my big bad shaman because I played it the first 2 days of release, people experimenting and expecting murloc shaman makes me a better player. But if i'd retry now i'd be lucky to hit ~50%
At all ranks : Classes with popular meme archetypes or overrated bad archetypes will be undervalued
Say there is a tempo rogue that's really OP, and a Tess rogue that's really popular (more popular than it is now). Tess rogue is going to drag the entire class winrate down, making the class look worse than it really is
Alternatively, a streamer came up with a new pogo rogue deck and convinced everyone it's really good when it actually isn't. This will have the same effect
If you are at a high rank : Classes with difficult to pilot archetypes will be undervalued; easy to pilot classes will be overvalued
Unless you pay for premium, HSReplay doesn't discriminate between ranks. At high ranks, pilot error is less common than the average, therefore decks that are difficult to pilot correctly and decks that are impacted harder by errors are better than they appear in the stats
Conversely, simple to pilot decks are as "powerful" as they are at lower ranks but will lose more often as they face opponents that are more powerful than their lower ranked counterparts
If you are at a low rank : Classes with easy to pilot archetypes will be undervalued; difficult to pilot classes will be undervalued
The inverse of the above
Legend with : S65 Freeze Mage, S57 Maly Gonk Druid, S57 "Okay" Shaman, S53 Boom-zooka Hunter, S53 Maly Tog Druid, S52 Wild Tog Druid ft.Blingtron, S50 Quest Rogue, S49 Dead Man's Warrior, S41 Wild Clown Fiesta Druid, S41 Hadronox Jade Druid, S40 Wild OTK Dragon Druid, S35 SMOrc Shaman, S33 Jade Druid, S22 Control Priest, S19 Control Priest
You need to balance the winrate with the games played too. Some decks have high winrate 60% but only 400 games played. I use the stats but try to balance the games with how often they're seen. If I was picking a deck to play I'd rather play a 55% winrate deck with 20k games played than a 60% winrate but minimal games to prove it.
These are all valid points, and there are more to be had as well. All it takes is for a Twitch streamer to play a deck and you will see it pop up everywhere, even if the deck has a terrible win rate. All of these factors skew the results.
What's worse is that if a streamer is at high ranks, then decides to play 1 trick pony meme decks like Disguised Toast, those are going to really affect the stats across the board as he is often at legend rank. And once you are at legend, many people stop trying and then play 'fun' decks. It's really hard to measure everything. There is enough data out there and enough serious players that never play meme decks and only play competitive stuff, but how do we know those numbers?
I say the numbers are good guidelines and starting points, but never use them to decide how to spend your dust and craft cards. You should be crafting cards for decks that YOU like to play, not what is popular or has a good win rate. And just because a deck has a good win rate, it doesn't mean that you will, especially at non legend ranks.
look at VS meta report as a general direction giver - then use your brain to adapt according to your current meta pocket
I think it's always close to 50/50. Those numbers can vary so wildly (and for many reasons, like even the time you play on or your region can affect win rates and thus those numbers) that it's not worth even looking at them.
Play what you like and what feels lucky during that specific moment. Over a very large sample size, decks should get close to 50/50.
Over a small sample size, it matters more to have fun, i.e. if you win 3 out of 4 matches with a deck while playing 1-2 hours in total, then don't change it until you stop playing (or, of course, until you become unlucky). If, instead, you win 1 in 4 matches after 10-15 games, then maybe try another archetype/etc, which might be luckier at that specific moment in time.
Nevertheless, to have a real chance at having fun you need to use the best decks available in the meta.
FWIW - Iksar has often cited HSReplays data when posting on the subreddit. Presumably, the data is pretty legit.
I'd advise not paying too much attention to the HSReplays homepage - overall class win-rates are essentially meaningless. Mike Donais was a guest on the Omni podcast about a year ago - he explained why class win-rates don't reflect game balance very well, since lots of people play good decks poorly, or simply enjoy playing shitty decks (his example was Mill Rogue - at the time, about 10% of the Rogue player-base was tanking the overall class win-rate by 2%, simply by playing the deck.) From the point of view of game balance, the devs don't look at class win-rates, and focus on whether or not each class and archetype are viable.
As far as Priest goes - if you look deeper at the numbers, about half of all Priest decks are Resurrect Priest. It has a win-rate just below 49%. All the rest of the Priest decks are as shitty as the OP seems to think, with win-rates bottoming out well below 40% - the net result is a class win-rate of 44%.