MTG Manabase Guide — Frank Karsten Math 2026

Frank Karsten’s probability math for every format — from Standard to Commander

I have lost count of the games I have thrown away because my manabase was wrong. Not close—wrong. Twelve white sources in a deck with four copies of a WW card on turn two. Three tap lands in an aggro deck where being a turn behind meant losing. One game at a Pioneer RCQ, I kept a two-lander with a Temples of Silence and Temple of Malice, curved out perfectly in my head, and then watched my opponent kill me on turn five because I was always one mana behind. That loss stuck with me, and it is the reason I eventually sat down with the math and figured out how manabases actually work.

This guide covers everything I have learned about building manabases across every format. The numbers here come from Frank Karsten's hypergeometric probability research, cross-referenced with tournament data and thousands of games on our Monte Carlo simulator. If you want the short answer for your specific deck, use the calculator below. If you want to understand why the numbers are what they are, read on.

Skip the math — get your answer now. Paste a Moxfield or Archidekt link and get exact land counts with Monte Carlo simulation.
Calculate My Lands →

Frank Karsten's Mana Math

The gold standard for manabase construction comes from Frank Karsten, a Pro Tour Hall of Fame member and trained mathematician who published groundbreaking research on colored mana requirements. Karsten used hypergeometric probability—the same math that governs any drawing-without-replacement problem—to calculate exactly how many colored sources you need to cast your spells on curve with roughly 90% consistency.

His core findings for a 60-card deck:

These numbers assume you are on the draw (seeing one extra card). On the play, you need roughly one additional source for the same consistency. We have run our own Monte Carlo simulations across 3.75 million sample games and Karsten's numbers hold up almost exactly—the 90% threshold is real, and dropping even two sources below it tanks your consistency noticeably.

Try our Manabase Calculator to apply Karsten's math to your own deck automatically. It runs 50,000 simulated games against your specific decklist.

Land Count Guidelines by Format

Every format has different constraints on how many lands you should run. Fetch lands, cantrips, MDFCs, and mana dorks all shift the math. Here is what I have found works across thousands of games in each format.

Standard (60 cards): 24–26 lands

Standard lacks fetch lands, so you cannot thin your deck. This means you need more raw lands than in Modern. Aggro decks with curves topping out at 3 mana can go as low as 22, but I would not go below 23 unless every spell in your deck costs 2 or less. Midrange decks want 24–25, and control decks running expensive planeswalkers or board wipes should run 26–27. If your deck plays modal double-faced cards (MDFCs) like Shatterskull Smashing, count the spell-side as roughly 0.3 of a land—they are not free.

Modern (60 cards): 20–24 lands

Modern is the most flexible format for land counts because of fetch lands. Scalding Tarn and Misty Rainforest effectively thin your deck and fix multiple colors simultaneously. Combined with cantrips like Consider and Serum Visions that smooth draws, many top Modern decks run 20–22 lands. Izzet Murktide famously runs 19–20 lands because the entire deck costs 0–2 mana. But do not blindly copy low land counts—if your curve includes 3+ mana spells, you still need 22–24.

Pioneer (60 cards): 23–26 lands

Pioneer plays like Standard in terms of mana requirements. Without fetch lands for thinning, you need similar raw land counts. The key difference is the available dual land cycles: Pioneer has shock lands, fast lands, and pain lands, but no fetches. This means your fixing comes with a real cost (life loss or entering tapped), and you need to be more deliberate about your color ratios. I have found that three-color Pioneer decks need at minimum 10–12 sources of each color to function.

Commander (99+1 cards): 35–38 lands

Commander is where manabase math gets interesting because of the 99-card singleton deck. I ran Monte Carlo simulations across five Commander archetypes and found that 36 lands is the sweet spot for most decks. cEDH combo lists can go as low as 28–30 because they run heavy fast mana (Mana Crypt, Chrome Mox, Mox Diamond) and have extremely low curves. Battlecruiser decks casting 6+ mana commanders want 37–38 lands plus 10+ ramp spells. See our Commander land count research for the full Monte Carlo data across 3.75 million games.

Limited (40 cards): 17 lands

The default is 17 lands in a 40-card draft or sealed deck. Aggressive decks with very low curves (all 1–3 drops) can trim to 16. Control decks or decks with expensive bombs should go to 18. In my experience, the most common draft mistake is cutting to 16 lands in a deck that has two 5-drops and a 6-drop. Those expensive cards are your best cards—you need to cast them.

Colored Source Requirements

Raw land count is only half the equation. You also need the right colors in the right quantities. Here is Karsten's breakdown for 60-card decks targeting ~90% reliability:

The "untapped" distinction matters enormously. A Temple of Silence counts as a white source but not an untapped white source for turn-one plays. If your deck needs to cast Thoughtseize on turn one AND Wrath of God on turn four, you need 14 untapped black sources AND 14 total white sources (tapped is fine for turn four). This is where most players get tripped up.

Multi-Color Decks: Overlapping Sources

In a two-color deck, dual lands count toward both colors. This is why shock lands and fetch lands are so powerful—a single Godless Shrine is both a white source and a black source. In a WB deck that needs 14 white sources and 14 black sources from 24 total lands, you could run: 4 Godless Shrine, 4 Brightclimb Pathway, 4 Concealed Courtyard (12 dual sources) plus 6 Plains and 6 Swamps. The 12 duals count toward both colors, giving you 18 white sources and 18 black sources total.

Three-color decks are where the math gets tight. In a Mardu (WBR) deck, each land can only help at most two of your three colors. You need a critical mass of triomes, fetch lands, or other three-color fixers to make the numbers work. This is exactly why our calculator exists—manual math for three or more colors is error-prone.

For Commander, multiply Karsten's 60-card numbers by roughly 1.6x due to the larger deck size. A single-pip spell needs about 22 sources; double-pip needs 29. But Commander also has universal fixers like Command Tower, Arcane Signet, and the Signets that make hitting color requirements easier than the raw math suggests.

Types of Mana Fixing

Not all mana sources are equal. Understanding the tiers of mana fixing helps you prioritize which lands to include, especially on a budget.

Fetch Lands

Fetch lands like Flooded Strand, Polluted Delta, and Scalding Tarn are the best mana fixers ever printed. They find any land with a basic land type, meaning a single Flooded Strand can find a Hallowed Fountain (white or blue), a Godless Shrine (white or black), or a basic Plains. Fetches thin your deck, shuffle away bad topdecks, fill your graveyard for delve, and trigger landfall. The only downside is 1 life per fetch. In formats where they are legal (Modern, Legacy, Vintage, Commander), always start your manabase with fetch lands.

Shock Lands

The ten Ravnica shock lands (Hallowed Fountain, Watery Grave, Blood Crypt, Stomping Ground, etc.) are the second-best dual lands in the game. They have basic land types so fetch lands can find them, and they give you the choice: pay 2 life to enter untapped, or enter tapped for free. In aggressive decks, you almost always pay the life. In control decks, you can afford to take the tap on early turns. Shock lands are legal in Pioneer, Modern, and Commander.

Fast Lands

Kaladesh fast lands (Concealed Courtyard, Spirebluff Canal, Blooming Marsh, Inspiring Vantage, Botanical Sanctum) enter untapped if you control two or fewer other lands. They are excellent for aggressive decks that care most about turns 1–3. After turn three, they enter tapped, which is irrelevant for low-curve decks but painful for midrange. I run 4 fast lands in every Pioneer aggro deck I build.

Check Lands

Check lands (Glacial Fortress, Drowned Catacomb, Dragonskull Summit, etc.) enter untapped if you control a land with the right basic type. They pair well with shock lands and basic lands. In Standard, where shock lands and basics are your main sources, check lands are very reliable. In Commander, they get worse because you have fewer basics.

Pain Lands

Pain lands like Adarkar Wastes, Caves of Koilos, and Battlefield Forge always enter untapped. Tap for colorless for free; tap for a color and take 1 damage. They are underrated in my experience. The damage adds up in long games, but having guaranteed untapped mana on every turn is worth it, especially in aggro decks.

Tap Lands and Budget Options

Tap lands (temples, gain lands, bounce lands) are the cheapest dual lands but the weakest. Entering tapped means you are effectively a turn behind every time you play one. I try to limit tap lands to 0–2 in competitive 60-card decks. In Commander, where games are longer and slower, 4–6 tap lands are acceptable. If you are on a strict budget, prioritize untapped sources for your early plays and accept tapped lands for your later turns.

Mana Rocks and Dorks

Sol Ring, Arcane Signet, Signets, and Talismans are the backbone of Commander ramp. When counting sources, I treat a 2-mana rock as roughly 0.5 of a land—it does not help you hit your land drops, but it does fix colors and accelerate you. Mana dorks like Llanowar Elves and Birds of Paradise count similarly: they fix colors and ramp, but they die to removal. In 60-card formats, Noble Hierarch and Ignoble Hierarch do double duty as fixing and early pressure.

Tap Land Math: The Hidden Cost

This is the section most manabase guides skip, and it is the one that changed how I build decks. A tap land does not just delay you by one mana on the turn you play it. It delays you by one mana for every turn after, because you are always behind where you should be.

Consider two scenarios on the play. In Scenario A, you play an untapped land every turn: turn 1 you have 1 mana, turn 2 you have 2, turn 3 you have 3. In Scenario B, your turn-1 land enters tapped: turn 1 you have 0 mana, turn 2 you have 2 (but you missed a turn-1 play), turn 3 you have 3. That missed turn-1 play could have been a Thoughtseize, a Fatal Push, or a Ragavan. Against aggro decks, being a turn behind on tempo is often the difference between stabilizing and dying.

My rule: in competitive 60-card decks, every tap land needs to earn its slot with a significant upside. Temples give scry 1, which is worth it in some control decks. Triomes have three colors and basic land types (fetchable), which is worth it in three-color decks. A Dimir Guildgate with no upside? Never.

Building a Manabase Step by Step

Here is the process I follow for every new deck I build:

  1. Determine your color requirements. List every card in your deck and count the colored pips. A deck with 16 white pips and 8 black pips needs significantly more white sources than black.
  2. Set your land count. Use the format guidelines above as a starting point, then adjust for your curve. If your average mana value is below 2.5, you can trim a land. If it is above 3.5, add one.
  3. Start with the best duals. Fill in fetch lands and shock lands (or the best available cycle for your format) first. These do the most work per slot.
  4. Check your colored source counts. Use Karsten's numbers (14 for single pip, 18 for double) to verify you have enough sources of each color. Adjust with additional duals, basics, or fixing spells.
  5. Minimize tap lands. Every tap land should have a specific reason to be there. If you are over your tap land budget, cut the weakest one for a basic or a pain land.
  6. Add utility lands last. Lands like Castle Locthwain, Mutavault, or Field of Ruin are powerful but produce only one color (or none). Add them only after your color math checks out.

Or, paste your decklist into our Manabase Calculator and let the Monte Carlo simulation handle all of this for you.

Common Manabase Mistakes

Mana Base Adjustments by Archetype

Different deck archetypes have different mana base needs that go beyond raw land count and color ratios. Here is what I adjust for each archetype:

Aggro

Aggro decks need every land to enter untapped. I run zero tap lands in competitive aggro builds. Prioritize pain lands and fast lands over check lands and temples. Accept life loss for tempo — you plan to win before the life loss matters. If your aggro deck has 3+ tap lands, you will lose games to sequencing.

Midrange

Midrange is the most flexible archetype for mana bases. You can afford 1–2 tap lands in the early game because your curve usually starts at 2–3 mana. Prioritize color consistency over speed. A midrange deck that stumbles on colors loses harder than one that plays a tapped land on turn one.

Control

Control decks can run more tap lands because they do not need mana on turns 1–2 (you are holding up counterspells, not deploying threats). Temples with scry are excellent in control. Run 26–27 lands in Standard control, 24–25 in Modern. Control decks flood less painfully than other archetypes because they use excess mana for card draw and instant-speed interaction.

Combo

Combo decks need to hit a specific mana threshold to go off. If your combo costs 5 mana across two cards, you need 5 mana of the right colors by turn 5. Run more lands than you think — a combo deck that misses a land drop often just loses because the combo turn gets pushed back and opponents find interaction. I add one land above the format average for every combo deck I build.

Use the Calculator

Our MTG Manabase Calculator implements Frank Karsten's mana math for every format with Monte Carlo simulation. Paste a Moxfield or Archidekt link, and it will analyze your decklist, identify your color requirements, recommend exact land counts and colored sources, and show you the probability of hitting each color on curve. It recognizes 48 dual land cycles and 10 types of ramp, and it runs 50,000 simulated games to give you real probabilities—not estimates.

Further Reading

Related Guides & Tools

Explore every dual land cycle in our Dual Land Cycles Guide or use the interactive Dual Lands Reference Tool to compare options by format. Building a Commander deck? Our Commander Deck Building Guide covers the 10-10-10 framework, power brackets, and budget tips. Want the raw Monte Carlo data? Read our Commander land count research covering 3.75 million simulated games. Check draw probabilities with our Hypergeometric Calculator.

Frequently Asked Questions

How many lands should I put in my MTG deck?

For a 60-card deck, 23–26 lands is standard. Aggro decks run 22–23, midrange 24–25, and control 25–26. Commander decks (99 cards) need 35–38 lands plus mana rocks. Use Frank Karsten’s mana math or a manabase calculator for precise counts based on your specific mana curve and color requirements.

What is Frank Karsten’s mana math for MTG?

Frank Karsten’s mana math uses hypergeometric probability to calculate the exact number of colored mana sources needed to cast spells on curve with ~90% consistency. Key numbers for 60-card decks: 14 sources for a single-pip spell, 18 for double-pip, 21+ for triple-pip. This research is the gold standard used by competitive Magic players to build optimal mana bases.

How do I build a manabase for Commander?

Start with 36–38 lands in your 99-card deck. Include fetch lands and shock lands for reliable fixing, add check lands and fast lands for budget options, and finish with utility lands like Command Tower and Exotic Orchard. Add 10+ mana rocks (Sol Ring, Arcane Signet, Signets) for ramp. Use a manabase calculator to determine exact colored source counts based on your casting costs.

How many colored mana sources do I need?

For a 60-card deck: 14 sources to reliably cast a single-pip spell on curve, 18 for double-pip, and 21+ for triple-pip. Dual lands count toward both colors. For Commander (99 cards), multiply by roughly 1.6x: you need about 22 sources for a single-pip cost and 29 for double-pip. Mana rocks and dorks also count as colored sources.

Are tap lands bad in MTG?

Tap lands are significantly worse than untapped alternatives because they delay your mana by a full turn. In competitive 60-card decks, limit tap lands to 0–2 and only include them if they provide substantial upside (scry from temples, three colors from triomes). In Commander, where games are slower, 4–6 tap lands are acceptable, especially on a budget.