See also: lifetime head-to-head grid, breakdown of each player's season, elo-based metrics.

This table was made by simulating the tournament 2.5 million times. I assumed that two people couldn't play each other in constructed twice and that pairings are random among people with the same number of match points except for the final round of Swiss. (This is the same as what I did in 2017, so if there were issues last year I didn't fix them.)

There are of course many limitations, and this shouldn't be taken to be much more than a conversation starter. Although the Elo system seems to give about the correct win percentages on a very large scale (see the data in this blog post), Magic is a game with a lot of variance and it takes a long time for the percentages to converge to the theoretical values. There's a reason I had to simulate the tournament millions of times instead of thousands. This is more of a thought experiment than any kind of accurate science. We're imagining that each player will continue posting wins at the same rate against players of the same pedigree as they had been in the recent past.

Also it's not clear whether the current Elo rating of each player is the right seed for their win expectancy. That scheme almost certainly undervalues the lower few people and overvalues the highest couple. I reran the simulation 2,5 million times using each person's average Elo from the previous year as the seed. To return to the original view, refresh the page or click here.