Relationship of lands played with game winrate, MKM

I’m playing in the MKM Arena Open tomorrow; the day 2 event is best-of-three (Bo3) draft. A classic question for a given limited format in Magic is how many lands a player should be running in their list. This isn’t a simple question to answer, and depends on the kind of deck that’s been drafted. However, I wanted to see if there were any patterns that have emerged over the lifetime of the set since it was released last month.

For this analysis, I’ve used game data from 17lands.com, and tallied the number of basic lands and nonbasic lands that a player has included in their list (by game). I’ve excluded any games where a player submitted a deck with more than 18 lands; the decision to run 19 is atypical for any limited format. This exclusion is about 400 games in total.

First, we have the overall performance, without regard for other factors (such as the colors in a player’s deck). While the differences between our groups are pretty small, running a list with 17 lands has the highest rate of success. This is also the most common choice by far (approx. 79% of games are running a list with 17 lands).

# Lands	Losses	Wins
16 (or less)	6,485 (42.6%)	8,752 (57.4%)
17	41,708 (41.8%)	58,057 (58.2%)
18	4,909 (42.6%)	6,625 (57.4%)
Overall	53,102 (42.0%)	73,434 (58.0%)

Breaking apart the decks by colorpair, we see some hints that green decks might benefit from running 18. (Note, I’m excluding monocolor and 3+ color decks, but I’m not breaking out splashes.) Interestingly, it looks like only WB benefits from running 16 lands. Every red deck in the format is giving up a lot of consistency for playing fewer lands.

It seems notable that WR doesn’t seem to improve (on average) by cutting lands. You might expect that it could afford to do so, given its access to the best 1-drops in the format, such as Novice Inspector. However, even a closer look at Novice Inspector suggests why this isn’t the case. The benefit this card grants you for casting it (a clue token) requires additional mana to leverage. Given the deck’s popularity, I haven’t been able to draft it frequently, but the result we’re seeing for WR makes sense. In modern limited formats, it’s difficult to overstate the importance of impacting the board early. However, not playing a land on turn 3 appears to be particularly punishing in MKM due to the disguise mechanic. Even in aggressive decks, you have to keep pace with your opponent along the mana curve. In terms of what’s realistic for a given draft, there simply aren’t enough cards that cost 2 mana or less to form enough of a core for one’s deck.

For MKM, the choice of how many lands to play is complex, but running 17 remains a solid default. It seems that green decks with 18 lands might outperform other builds in their colors, but you’ll need some good intuition on whether your pool leans more towards the deck’s theoretical ceiling to justify cutting an additional spell.