You Suck at Math: How Many Lands to Run in Commander

The state of writing content for Magic is dire. As Cal Jones wrote for Cascade Cascade, articles are written for a pittance of dollars, if there's even any payment involved. Competition and capitalism demand that we race to be first to market, to report on the newest leaks as fast as possible, to speculate on what the newest mechanics do, even if that means that we get anonymously dragged online for getting it only 95% correct. I can live with this environment — I subscribe to the theory that writing for yourself first is one of the best ways to write. But if there’s one thing, one singular concept I hope shatters through the iron curtain of the gratification-disorder-feeding algorithm dopamine desert that is the internet in 2026, it’s this article.

Put more goddamn lands in your Commander decks, guys.

We are Not as Good as We Think

As the senior Magic player in my budding playgroup, I take it upon myself to help my friends get better at the game by providing advice and guidance when asked. But I’m just a guy who’s played a lot of card games. I have no competitive Magic experience, no pro tours or GPs to my name, and certainly no advanced math degree. That being said, I’ve built a ton of Commander decks, and I’ve been playing Commander for long enough that I have a good grasp on how the format works.

Recently a player in my group made an inquiry regarding land count for an Oloro, Ageless Ascetic lifegain deck. It’s a slow, grindy build meant for a lower bracket in a color combination that’s pretty famous for struggling with ramp (besides white). I wasn’t able to respond immediately, but when I checked back later that morning, another of the players in the group advised that 33-34 lands (or a third of the deck) should be the standard. In Esper colors. Ladies, gents, theys and gays, I have crapped myself several times as an adult man for a variety of reasons, but that instance may have been the angriest. 

When I pulled myself out of a very smelly stupor 20 minutes later, I asked Mr. 34 Lands why his land count was so low. The popular opinion when I started playing Commander was that 36 lands was the baseline. This player — who started within the last three years — told me that the people who taught him how to play all said that one third of your deck as lands was the gold standard. I want to find the guys who taught my dear friend how to play this game and throttle them. 34 lands?! Why not just skew your curve to an average mana value of 5 while you’re at it, since you clearly hate actually participating in the game?

It’s More Common than You Think

It must have been fate, some grand cosmic joke, or just sheer bad luck that later, a gentleman in one of my LGS Discord servers asked for help with his edits of the recent Silverquill Influence precon. Always willing to help people refine their Commander decks, I told him that I would have a look. The sheer force of opening the link knocked all of the Mountain Dew cans and Cheetos bags off my desk, and I had to get a third pair of underwear for the day. Can you guess what assaulted my eyeballs?

I mean seriously, am I being punked? Is someone out there paying people to do this to me? The gentleman in question cut seven lands from a precon. A precon! They’re literally designed by the people who create this game to be a template for how to construct a Commander deck. Precons used to be pretty shaky from a design perspective, but these days Wizards of the Coast has stepped up their game, and modern out-of-the-box decks usually include a healthy 37-39 lands and a decent ramp package. How does someone see this and decide, “nah, I know better, my friends told me to run 31 lands in a deck!” 

I politely asked this guy why he cut so many lands, and he said his friends told him to run 31 lands. They play cEDH, where lower land counts are extremely common if not mandatory. I asked him if these friends knew he wasn’t building a cEDH deck, and he said yes. Now I had a new thought to contend with. People playing the top-level competitive version of Commander, supposedly some of the more mechanically-skilled and intelligent players, are telling people to cut huge swaths of lands out of their mana base in Bracket 2. I am rapidly running out of family-friendly ways to express my extremely angry confusion.

These two incidents, albeit anecdotal, are not at all out of the norm for Commander players. I browse lists on various deckbuilding websites on occasion when I need inspiration for a deck-in-progress, and I see lists at all power levels running 35, 33, even 30 lands without a single MDFC (which shouldn’t always immediately replace lands to begin with) in sight.

The problem of land count in Commander is apparently a widespread one. I am realizing more and more why so many Commander players seem to have such a tough time getting their decks to function properly. As it turns out, when you cut yourself off from one of the vital resource mechanics in Magic, it tends to be hard to play the game properly.

Here are some of the fallacious arguments that lead Commander players down the dark path of cutting lands, and my refutations of those arguments.

“But I Run a Lot of Ramp!”

This is an argument I hear frequently from the filthy land-cutters, sometimes smugly delivered with the oblivious satisfaction of someone who discovers MLMs for the first time. "I run twelve pieces of ramp," they might say. "I can run fewer lands if I run more ramp."

Here’s the thing: ramp requires mana to cast. If your opening hand is two lands and a bunch of Cultivates and Kodama’s Reaches, you’re banking on drawing a third land just to play the game for real. You are not ramping. You are a moron with a grip of cards that requires a third land you do not have. Ramp accelerates you from a single additional mana per turn into an increasing advantage over your opponents, but it is not a substitute for having a functional mana base in the first place. Sol Ring into Arcane Signet is a fantastic start. Sol Ring into staring at your hand because you only have one land and are hoping to draw into something, anything at all, is not.

This is especially offensive in the case of mana rocks, particularly at lower-power brackets. While mana rocks absolutely serve a purpose, especially in non-green decks without land ramp, they are significantly less durable and consistent than land ramp. In a format where artifact removal is ubiquitous and board wipes are a fact of life, a Cultivate puts a land into play and prepares your next land drop. No amount of Farewells or Austere Commands will take them off the board, and until you break into MLD-accepted tiers of play, that isn’t a problem you’ll ever have to worry about.

“But My Curve is Low!”

At a lower average mana value, a low curve can be a real consideration. A low-to-the-ground Elfball deck with a laundry list of mana dorks often wants to see fewer lands, as they want to draw into Craterhoof Behemoth and related win conditions quickly and efficiently and are vomiting their hands out anyway.

The thing is, those Elfball lists still need an amount of lands that gets them off the ground. Your deck needs a critical mass of lands to guarantee that you actually hit the threshold of handling all of your mana needs via the mass of mana dorks sitting on your board. Even later in the game, you still benefit from at least playing one land a turn — there’s literally no downside to increasing your available mana every turn. You just don't want to see multiple lands sitting dead in your hand.

Considering all of that, the decision to cut lands down is not one to make lightly, even in specialized decks like this. Through the use of some crunchy math, though, the decision can be made more easily.

The Math is Not in Your Favor

As I wrap up this written elegy borne of rage and an amount of cocaine that would make 1980s Stephen King blush, I want to leave you with something to recontextualize your deckbuilding. I start my decks at 37 lands at the floor, though I’ve gone as high as 61 in my meme budget deck. This is not a hard-and-fast minimum, as I've cut to 36 for some of my decks that mathematically supported a slightly lower land count, and in years previous I went as low as 33 lands in an aggressive Elfball deck with over a dozen low-drop mana dorks.

The complex math dealing with this issue is well outside my range of knowledge and experience. Fortunately for me, the local idiot, several Magic professional/competitive players with a significantly better pedigree than myself have created a lot of content that covers this exact topic. Sam Black posted a video a few years back after a particularly controversial Twitter thread about mana rocks, and Frank Karsten has written several articles on this exact topic. Start with How Many Lands Do You Need in Your Deck? An Updated Analysis, then read the followup articles here and here for some extremely thorough mathematical analysis on the subject of how many lands to run in your Commander deck. 

If that’s too much reading for you, here’s a TL;DR:

Lastly, if you want a helpful and dynamic tool that can help you determine exact probabilities for different stages of the game, try AetherHub’s hypergeometric calculator. Put in a deck size, adjust your number of successes in the deck (lands, in this case), and run the calculation. For example, using my baseline of 37 lands and testing for an opening hand of seven cards with a desired 3-land start, here’s a breakdown:

In summation, here are some tips: run more lands, use math, and if you ask me to review your decks, include more than 37 lands unless you like getting roasted.

Have opinions about land counts? Want to defend your 32-land deck to my face? Come find me on the EDHLAB Discord at @shadedfall and we can discuss it. Politely. I will be very polite about it. Promise.

The Best Way to Play Commander Online

For the easiest way to play Commander online with your friends or meet new opponents on the field of battle, visit EDHLAB.gg and check out our Play Online feature!

EDHLAB does not support the use of generative AI as a means to produce content. Any articles you read on our website will never incorporate generative AI for written or visual materials.