Quote from pizzap »This is why I stopped collecting meta data. It is a thankless job and usually it is discarded because n<1,000,000.
You can get a pretty narrow confidence interval with n=100. That's in the +/- 10% range, which helps us narrow a matchup in the favorable/unfavorable ranges. A 60/40 match at n=100, assuming an unbiased sample, is in that 50-70 range, so at least even and likely favorable. That's not to say it isn't thankless work, which it totally is. Just to say that one can do this work without n being 1000000. Although it sometimes feels that way when talking to people online haha.
EDIT: Also, some great design insights from Maro's article today -
See the Infernal Judgment entry. Notable quotes:
"One of the tasks of Core Set 2019 was to make some cards relevant in Modern but not disruptive to Standard. "
"First, how do we make it relevant in Modern? Find a problem Modern has and try to solve it. How about the Eldrazi, especially the large legendary ones?"
"This means that the spell has some functionality in Standard but has qualities that address a Modern issue."
We can quibble about whether or not this card accomplishes its stated goals, and/or whether or not Wizards is addressing the right issues. But the fact that they are deliberately addressing such issues and publicizing those efforts is a huge step in the right direction, even if there's more work to be done.