How good are we at figuring out which cards win?
Drafting Magic: The Gathering (or anything, really) requires you to predict in advance the best picks for what you need. It’s not easy to know in advance if you’ve made the right pick or not. Thanks to the team over at 17lands, we can use the aggregated data of everyone who has installed their tracker to rate MTG cards. One thing that’s nice about this data is that there is data for pick priority, as well as in-game performance for each card. The simplest ways of thinking about this are using the two metrics of “average taken at” (ATA), and “game-in-hand win-rate” (GIH WR). These metrics describe when a card is picked on average, and how often a card wins on average when it’s actually in hand at some point in the game. Here’s a plot of ATA against GIH WR, using the latest available draft set, Dominaria United:
You can mouse-over the plot and see the card names. Cards above the line are better than you’d expect from their pick rate. This makes sense with excellent cards (like Archangel of Wrath). Cards below the line are worse than you’d expect (sorry Urza Assembles the Titans).
For attribution, please cite this work as
andrewstiegler (2022, Sept. 7). the data doctoR: Evaluating Cards in MTG Drafts. Retrieved from https://andrewstiegler.github.io/posts/2022-09-01-my-second-post/
BibTeX citation
@misc{andrewstiegler2022evaluating,
author = {andrewstiegler, },
title = {the data doctoR: Evaluating Cards in MTG Drafts},
url = {https://andrewstiegler.github.io/posts/2022-09-01-my-second-post/},
year = {2022}
}