The original bopmtg was known for its mana tables, and for good reason: mana problems lose more games of Magic than bad matchups do. This article rebuilds that foundation as a practical reference. The numbers below are starting points backed by probability math, plus the adjustments that actually matter. (The classic mana tables themselves are back too, recomputed with exact odds.)

The core question

Every mana base answers two separate questions:

1. How many lands do I need to hit my land drops on time?
2. How many colored sources do I need to cast my spells on curve?

Mixing these up is the most common deckbuilding mistake. You can have plenty of lands and still fail to cast a double-cost spell, and you can have perfect colors but stall on three lands.

Question 1: total land count

The intuition comes from a simple expected value. With 24 lands in a 60-card deck, your opening seven contains on average 7 × (24/60) = 2.8 lands. Every draw adds another 0.4 of a land. That is roughly the profile a deck full of two-, three-, and four-cost spells wants: three lands on turn three most games, four by turn five.

Starting points that competitive play has converged on:

Deck profile	60-card deck	40-card Limited deck
Aggressive, curve tops at 2-3	20-22 lands	16 lands
Midrange, curve tops at 4-5	24-25 lands	17 lands
Control or ramp, wants every drop into the late game	26-27 lands	17-18 lands

For Commander (99 cards), the equivalent baseline is roughly 36-38 lands plus 8-12 pieces of ramp or mana acceleration, adjusted the same directions as below.

The adjustments

These starting points move for specific, quantifiable reasons:

• Cheap card selection lets you cut lands. Each pair of one-mana cantrips or cheap filtering spells (think Consider, Ponder, Sleight of Hand) is worth roughly one land, because they dig you to land drops early and shuffle away flood late. This is the mana base half of the Turbo Xerox principle, which gets its own article.
• Modal double-faced cards count partially. A card that is a land on one face and a spell on the other counts as roughly two-thirds of a land. Four MDFCs let you shave about two or three real lands.
• Curve matters more than average mana value. A deck whose plan is to cast a five-drop every game on time needs to build for the fifth land drop (26+ lands or ramp). A deck that merely tops out at five as a luxury does not.
• Flood insurance changes the calculus. Lands that do something late (creature lands, channel lands, utility lands) let you play the higher land count without the cost, because excess lands stay useful.

Question 2: colored sources

This is where most homebrew mana bases actually fail. The question is not “how many Islands,” it is “how many blue sources do I need to cast this specific card on curve, reliably?”

The benchmark research here is Frank Karsten’s colored-source work, which computed, for each cost and turn, how many sources you need for roughly 90 percent castability on curve. The headline numbers for a 60-card deck are worth memorizing:

You want to cast	On turn	Sources needed (approx.)
C (single color, one-drop)	1	14
1C (two-drop, one colored)	2	13
CC (two-drop, double colored)	2	20
2C (three-drop, one colored)	3	12
1CC (three-drop, double colored)	3	18
CCC (triple colored)	3	23

Two things jump out of that table:

1. Double costs are expensive. Going from 1C to CC on turn two jumps the requirement from 13 sources to 20. This is why two-color aggro decks with double-costed one- and two-drops need nearly mono-colored mana bases, and why a “splash” can never support a CC card.
2. Later is easier. The same double cost on turn three needs 18 sources instead of 20, and every turn after relaxes it further. If your CC card is fine arriving a turn late, you can cheat these numbers a little.

For Limited, the practical translation: 8-9 sources support a main color, a splash wants 3-4 sources plus fixing, and double-costed splash cards are traps.

A worked example

Suppose your 60-card midrange deck wants to cast a CC two-drop and a 1GG three-drop reliably, in white and green. You need about 20 white sources and 18 green sources. With 25 land slots (midrange baseline), that requires 13 dual lands that produce both colors: 12 mono lands can carry one color while duals fill both requirements at once. That is exactly why modern two-color mana bases run 12+ duals, and why the quality of a format’s dual lands quietly decides which two-color decks are playable.

The short version

• Pick your land count from the curve table, then adjust for cantrips, MDFCs, and utility lands.
• Count colored sources per card you actually want to cast on curve, using the benchmarks above.
• Respect double costs. They are the single biggest source of “unlucky” games that were actually deckbuilding errors.

Flood and screw never fully go away. The goal is to make sure that when you lose to your mana, it was variance, and not a decision you made a week earlier.