┻┳| •.•)
┳┻|⊂ﾉ
Time's running out to add the Desert Bloom card back to your collection! Win 5 ranked games to snag this succulent… https://t.co/fRYCRe8wtc
"Greetings fellow cuties."
#TBT to when Dalaran got a new set of rockets, courtesy of The League of E.V.I.L. 🚀 https://t.co/FX3bwrreOE
Same, as soon as I saw this I thought Divine Beasts.:) Super cool OP.
@Goddasaurus @sailing_sheep @Warcraft @Blizzard_Ent A toast to your amazing talent! https://t.co/CA2hzZu8IP
RT @Goddasaurus: Latest commission. Beautiful, charming and deadly Mornil for @sailing_sheep This took a while :D @Warcraft @PlayHearthst…
TFW you play Galakrond for the win 😈 https://t.co/19iJE66X77
You, sir, are an inspiration!
The latter approach is exactly how we would approach designing dual-tribe minions. I'm not sure where the narrative for the former approach comes from. There are a bunch of unanswered questions when it comes to having minions with more than one tribe/type, but it sounds like an interesting design challenge. If there was ever a set that made sense thematically for it, it sounds like something we would investigate.
Usually we track data to help inform whether or not a trend in the meta is likely to continue. Whether to actually make or not make a balance change because of a trend pretty rarely comes from data itself. So, it’s a tool that can inform not a tool that decides. Also no, the percentage population of individual archetypes across ranks are not identical. What I mean to say is that they are so similar that if you were to stack rank the ‘best decks’ in both metrics it’s unlikely those decks would change much, if at all, at least from rank 3-L.
Balance decisions don't come from balance data. If that were the case, we would just do game balance through automation. The reason we use data at all is because it's a small piece of informing what player perception of power level might be. Perception of power is really the piece that matters, and that is difficult to track through real metrics. The purpose of posting data is because some people find it interesting to look at and discuss, not to inform anyone on how balance decisions are made because they aren't made by looking at numbers in a spreadsheet.
The only objective is to share data on what happened over the last few days within a specific rank range, not to inform anyone on what deck is strongest for any individual player. We track a variety of data internally to help inform balance decisions, I did not share all of those here because it can be hard to follow along with a giant post about complex calculations. The metrics I chose to share were chosen because they are easily digestible and get the general point across. There are many much more complex and specific to each individual metrics that would be better for determining the best deck for any one person. Expected win rate (generally I see this referred to as weighted win rate) is a reasonable way to look at what is most powerful, though you would find actual and weighted are nearly identical in most circumstances, especially in the case of rank 3-L. The extreme circumstances listed in this post could happen, but rarely occur in practice. The population of decks at rank 3 and L could be much different, but are generally very similar.
Besides providing explosive backup, Blastmaster Boom's also great at setting the stage for a grand master plan reve… https://t.co/mfk748qM3O
You can say the League of E.V.I.L. are partners in crime. https://t.co/5YARYaMHzi
Behold, mini Rafaam, ADORABLE SUPREME ARCHAEOLOGIST!!! https://t.co/oociWqiafn
It's no adventure without villains. Show some love for the League of E.V.I.L with these truly outrageous wallpapers! https://t.co/49kg2YWokt
Sure, I think expected winrate is just not the terminology I’m used to hearing to describe this. Corbett and I spent some time in DM last night and I think we both have a better understanding of the merits of either type of analysis. Thanks for the thread.
I think I am confused then, I don't understand what you mean when you say actual winrates aren't the best measure of looking back. Aren't they absolutely the best way of looking back because by definition they are actually what happened? If your goal was to say "players that played this deck had an X% win rate on X/X date" wouldn't you rather use actual win rates?