Dead Card Bias

[AnnanFay]

I'm curious about: "Popularity: Number of times the card has been played.
Win Rate: Number of times the card was planed [played?] in winning games divided by Popularity."

I am thinking about dead card bias. Fair examples would be Frostwolf Warlord and Pyroblast. In most games where I play FW I already have board control and am more likely to win. Pyroblast even more so because it's a natural finisher, there won't be many games where it's played by a losing player. I also find it hard to believe it's really the second best card!? 😕

Does the 'played' include dead cards?

Another thing I just though of. It is impossible to account for opponents dead cards if they never play them. It would be useful if we could filter the stats so it only takes data from first party users.

(Edit: Additional data, looking at the top cards by Win Rate the top 7-8 of 10 are all finishers.)