How Many Lands in MTG? Land Count Guide
Karsten’s formula, colored source tables, and land counts for 60-card, Limited, and Commander decks
You just built a sweet deck. Now you’re staring at the land pile wondering — how many do I actually need? Cut too many and you’re stuck on two lands while your opponent curves out. Leave too many in and you’re topdecking Plains on turn nine. We’ve all been there.
Luckily, someone already solved this. Frank Karsten — Hall of Famer, mathematician, and the guy who’s been doing Magic probability math since before most of us knew what a hypergeometric distribution was — published the numbers everyone in competitive Magic references. If you just want the answer for your deck right now, plug it into the Mana Base Calculator. If you want to understand why, keep reading.
The Formula
Karsten ran regression analysis on thousands of winning tournament decklists and found a surprisingly clean formula for 60-card decks:
Lands = 19.59 + (1.90 × average mana value)
That’s your baseline. Then you adjust:
- −0.28 per cheap ramp or draw spell (Llanowar Elves, Consider, Rampant Growth)
- −1.0 per fast mana piece (Sol Ring, Mana Crypt, Chrome Mox)
- −0.74 per untapped MDFC (Emeria’s Call, Shatterskull Smashing)
- −0.38 per tapped MDFC (Bala Ged Recovery, Hagra Mauling)
So a midrange deck averaging 2.8 MV with 4 cantrips and 2 rocks: 19.59 + 5.32 − 1.68 = about 23 lands. An aggro deck averaging 1.8 MV with no ramp: 19.59 + 3.42 = about 23 lands. Yes, aggro still needs lands. The curve is just lower, not nonexistent.
Land Counts by Format
| Format | Deck Size | Typical Lands | Notes |
|---|---|---|---|
| Standard | 60 | 24–26 | No fetches to thin; the formula fits cleanly |
| Pioneer | 60 | 23–26 | Same deal — no fetches, honest mana bases |
| Modern | 60 | 20–24 | Fetches + cantrips let you cheat on count |
| Legacy | 60 | 18–22 | Brainstorm, free spells, and Daze do heavy lifting |
| Pauper | 60 | 22–24 | All basics and gainlands; no shortcuts |
| Commander | 99 | 36–38 | Plus 8–12 ramp pieces; details below |
| Draft | 40 | 17 | 16 for aggro, 18 for bombs — the eternal default |
| Sealed | 40 | 17–18 | Pools run higher curves than draft, lean toward 18 |
The Karsten Table: How Many Sources of Each Color?
Total land count is only half the battle. The harder question is: how many of those lands need to produce blue? Or black? Or both? Karsten’s colored source table tells you how many sources of a single color you need to reliably cast a spell with that color’s pips. These numbers assume about 90% consistency after realistic mulligans.
Sources Needed in a 60-Card Deck
| Cost | Example Card | On Turn | Sources |
|---|---|---|---|
| C | Ragavan (R) | 1 | 14 |
| 1C | Thalia (1W) | 2 | 13 |
| 2C | Fable of the Mirror-Breaker (2R) | 3 | 12 |
| 3C | Sheoldred, the Apocalypse (3B) | 4 | 10 |
| CC | Counterspell (UU) | 2 | 21 |
| 1CC | Liliana of the Veil (1BB) | 3 | 18 |
| 2CC | Wrath of God (2WW) | 4 | 16 |
| CCC | Steel Leaf Champion (GGG) | 3 | 23 |
| CCCC | Phyrexian Obliterator (BBBB) | 4 | 21 |
Look at Counterspell. UU on turn 2 means you need 21 blue sources. In a 24-land deck, that’s nearly every single land producing blue. If you’re playing three colors and jamming Counterspell, the math is screaming at you — you either need a pile of blue duals or you’re going to brick on UU more often than you think. There’s a reason competitive players treat double-pip cards as a real cost during deckbuilding.
Other Deck Sizes
For 40-card Limited, multiply the 60-card number by 0.67. A turn-1 one-drop needs about 9–10 sources. For 99-card Commander, multiply by 1.65 — a single-pip spell you want by turn 3 needs about 20 sources. The calculator does this math for you.
Commander: 99 Cards, 99 Problems
Commander mana bases are a different beast. Your deck is 65% bigger, but you’ve got some things going for you: guaranteed access to your commander, a free mulligan, and 40 life to absorb a slow start. Karsten accounts for these with a −1.35 land offset in his Commander formula.
In practice:
- Mono-color: 36–38 lands. Sol Ring, a few signets, and you can dip to 35.
- Two-color: 36–37, with 8–10 duals. Command Tower and your guild’s full cycle do the heavy lifting.
- Three-color: 37–38. You really want fetch+shock+triome if budget allows. On a budget, check lands and pain lands get you there.
- Four or five-color: 37–39. Command Tower, Mana Confluence, City of Brass, and every rainbow land you can get. Green ramp is your best friend here.
For ramp adjustments: each 2-mana rock is worth about a third of a land, so every 2–3 rocks let you cut one land. Sol Ring is worth about one land by itself. Don’t go below 33 unless you have an absurdly low curve and 15+ ramp pieces — and even then, think twice.
Limited: 17 Lands and When to Deviate
17 lands in 40 cards. Everyone knows this one, and the math confirms it — 17/40 (42.5%) closely matches 25/60 (41.7%), which is where most balanced constructed decks land.
- 16 lands: Your curve caps at 4, your average MV is under 2.5, and you’d rather draw gas than lands. Think mono-red aggro in a fast draft format.
- 17 lands: The default. Right for the vast majority of decks.
- 18 lands: You opened two 6-drops that win the game, you’ve got card draw to avoid flooding, or your sealed pool pushed you into a grindy three-color build. Sealed pools need 18 more often than draft.
Why These Numbers Work: The Actual Math
Karsten’s table isn’t guesswork or “it feels right.” It comes from the hypergeometric distribution — the probability of drawing specific cards from a finite deck without replacement. If you have S sources of blue in a 60-card deck and you’ve seen n cards (hand + draws), the chance of having at least k blue sources is a straightforward combinatorial calculation. We have a calculator for that too.
But raw hypergeometric math doesn’t account for mulligans. Karsten goes a step further and simulates actual mulligan decisions — shipping hands with 0, 1, 6, or 7 lands the way real players do. That’s what makes his numbers more accurate than pure math alone: they reflect how the game is actually played, including the London Mulligan rule.
Splashing? You Don’t Need Full Sources
Here’s something a lot of players miss: not every color in your deck needs the full Karsten treatment. If you’re base white-blue and splashing red for a single Fable, you don’t need 12 red sources. You need maybe 7 or 8.
A splash color — less than about 12% of your colored pips — only needs around 70% of the full Karsten target. A secondary color (12–40% of pips) needs about 85%. That’s how three-color decks carve out room for utility lands and avoid playing 30 duals with no basics.
Mistakes That Cost Games
- Treating tapped lands as untapped. Your Dismal Backwater doesn’t cast Counterspell on turn 2. Karsten’s numbers assume untapped sources for on-curve plays. Tapped lands are fine for later turns, but don’t kid yourself about early-game consistency.
- Underestimating double-pip costs. You’ve got 12 white sources and figure that’s enough for white cards. For Thalia? Sure. For Wrath of God at 2WW? You need 16. The jump from C to CC is 8 additional sources. That gap is where mana bases fall apart.
- Cutting 3 lands “because I added draw.” Each cantrip is worth 0.28 lands. Three cantrips save you less than one land. You’d need 10+ to justify cutting 3 lands. People do this all the time and wonder why they mull to five.
- Loading up on colorless utility lands. If you’ve got 24 lands but 6 are colorless (Field of Ruin, Mutavault, etc.), you only have 18 colored sources. That’s often not enough for double-pip cards. Utility lands are great — until they cost you the game by not making the right mana.
- Forgetting that fetches count double. Scalding Tarn finding Steam Vents counts as a blue source and a red source. Fetches are the strongest mana fixers in the game because they multiply your effective colored source count. Factor them in.
Run the Numbers
If you don’t want to do this by hand, the Mana Base Calculator does everything above automatically. Import a decklist or punch in your colors and curve. It runs Karsten’s formula, figures out your colored source targets with splash detection, suggests specific dual land cycles for your budget, and runs Monte Carlo sims to validate cast rates. Works for any format, any deck size from 40 to 100.
More Guides
- MTG Manabase Guide — full breakdown of Karsten’s mana math
- Dual Land Cycles Guide — every fetch, shock, fast, and check land with format legality
- Commander Deck Building Guide — ramp ratios, removal counts, and power brackets
- Hypergeometric Calculator — exact draw probabilities for any card
- Dual Lands Reference — compare duals across every color pair and format
- Commander Land Count Data — 3.75 million Monte Carlo simulations across 5 archetypes
Frequently Asked Questions
How many lands should I put in my 60-card MTG deck?
Frank Karsten’s regression formula gives the answer: lands = 19.59 + (1.90 × average mana value). For a typical midrange deck with a 2.8 average MV, that’s about 25 lands. Aggro decks with a 2.0 average run 22–23, while control decks with a 3.5+ average need 26–27. Adjust downward by 0.28 for each cheap draw or ramp spell.
How many lands do I need in a Commander deck?
Start at 37–38 lands for a typical Commander deck. Karsten’s research scales the 60-card formula to 99 cards and subtracts 1.35 for the guaranteed commander access and higher starting life total. Add 8–12 mana rocks for ramp. Cut 1 land per 2–3 cheap ramp spells and 1 per 3–4 cheap cantrips.
How many lands should I run in Limited (Draft/Sealed)?
17 lands is the default for 40-card Limited decks. Aggressive decks with a low mana curve (average MV under 2.5) can go to 16. Slow, mana-hungry decks with bombs at 5+ mana may want 18.
How many colored mana sources do I need to cast my spells on curve?
At 90% consistency in a 60-card deck: 14 sources for a single-pip turn-1 spell, 13 for turn-2, 12 for turn-3. For double-pip costs like WW on turn 2, you need 21 sources. Triple-pip costs like BBB on turn 3 need 23 sources. These numbers come from Frank Karsten’s hypergeometric probability research.
What is the formula for calculating MTG land count?
Frank Karsten’s regression formula: lands = 19.59 + (1.90 × avgMV) − (0.28 × cheap ramp) − (0.28 × cheap draw) − fast mana − (0.74 × untapped MDFCs) − (0.38 × tapped MDFCs). This was derived from winning tournament decklists using least-squares regression.