Magic: The Gathering is an interesting case. On the one hand, there's a crazy amount of data in terms of win rates on Arena and MTGO, as well as popular cards i

Magic: The Gathering is an interesting case. On the one hand, there's a crazy amount of data in terms of win rates on Arena and MTGO, as well as popular cards in a given format (especially EDH with EDHREC). What a computer cannot do is win games, or analyze a deck based on its oracle text and tell you what can win. One of the more interesting facets of this (to me) is watching the prices change from spoiler season to a few weeks/months after release. Fable of the Mirror-Breaker was less than a dollar when Kamigawa: Neon Dynasty came out, and it's over $20 now. Tiamat, from Adventures in the Forgotten Realms, was over $25 after the set was released, now you can get it for around $10-$15. (Price has a lot more information in it than just game utility, but its desirability for gameplay is still pretty high in the list.) For as uniform as, say, the standard meta is, it's still based on player consensus and behavior, not computer analysis of game mechanics or simulations of games. Turns out player consensus is still pretty good. I don't know if you could moneyball Magic.