The second-land-drop problem: why some decks mulligan forever

TL;DR: If you mulligan three times and still lose to mana screw, your deck doesn't have a land problem. It has a ramp problem. This article walks you through counting your actual mana sources (lands plus rocks plus dorks), comparing that count to your mana sinks (everything that costs 3+ mana), and finding the ratio that stops the bleeding. Most decks that "never hit their third land drop" are actually running 15 ramp pieces when they need 22.

---

ThePattern You've Seen A Hundred Times

You keep a two-land hand. Sol Ring, two action spells, two lands, two more cards. London mulligan says this is keepable. The math says you'll draw your third land by turn 4 about 70% of the time in a 38-land deck. You keep.

Turn 3 arrives. No third land. Turn 4, still no third land. Turn 5 you draw it but you're already two turns behind the pod. You lose. Next game you mulligan to 6 looking for three lands. You find them but now you're down a card and the Simic player has ramped twice. You lose again.

This is not variance. This is deck construction.

What Mulligan Data Actually Tells You

The London mulligan (introduced 2019, formalized in Comprehensive Rules 103.4) lets you bottom cards equal to the number of times you've mulliganed. That rule change made two-land hands statistically better than they were under the Vancouver mulligan. A 38-land deck keeping 7 cards with 2 lands will see a third land by turn 3 roughly 54% of the time. By turn 4 that climbs to 71%.

Here's the problem: those percentages assume you're drawing into lands. They don't account for what happens when you draw into a Cultivate on turn 2 with no third land to cast it. They don't account for keeping two lands and a mana rock when your deck has 8 total ramp pieces and 45 spells that cost 3 or more mana.

The hypergeometric distribution governs this. The formula is:

P(drawing k successes in n draws from a deck of N cards with K successes total)
= C(K,k) × C(N-K, n-k) / C(N,n)

Where C(a,b) means "a choose b" (combinations). For our purposes: if your deck has 38 lands, 8 mana rocks, and 6 mana dorks (52 total mana sources in a 99-card deck), and you're looking at a 7-card opening hand, the probability you see exactly 2 mana sources is about 24%. The probability you see 3 or more is about 73%.

That sounds good until you realize: if you keep the 2-source hand, you now need to draw into the remaining 50 mana sources from a deck of 92 cards over the next 2-3 turns. The probability you hit land number 3 by turn 3 (drawing 3 cards total including your normal draws) is 58% with 52 mana sources. With 46 mana sources (38 lands, 8 rocks, no dorks) that drops to 54%.

Six percentage points matters when you're playing 100 games a year.

Counting What Actually Produces Mana

Most players count lands. Some count lands plus Sol Ring. Almost nobody counts correctly.

A mana source is anything that puts you ahead on mana the turn you cast it or the turn after. That includes:

• Lands. All of them. Even the ones that enter tapped.
• Mana rocks that cost 2 or less. Sol Ring, Arcane Signet, Fellwar Stone, Talisman cycle, Signet cycle. These go in the "mana source" pile.
• Mana rocks that cost 3. Chromatic Lantern, Worn Powerstone, Coalition Relic. These do NOT go in the mana source pile. They cost 3 to cast, which means they don't help you hit your third land drop. They help you hit your fifth and sixth.
• Mana dorks that cost 1. Llanowar Elves, Birds of Paradise, Arbor Elf. These count. You cast them turn 1, they tap turn 2, you're ahead.
• Ramp spells that cost 2. Nature's Lore, Three Visits, Rampant Growth. These count.
• Ramp spells that cost 3. Cultivate, Kodama's Reach, Explosive Vegetation. These do NOT count as early mana sources. They're mid-game stabilizers.

Llanowar Elves

Go through your decklist right now. Count only the items in the first four bullets. Write that number down. If it's under 45, you have a second-land-drop problem.

The 45/30 Rule

Here's the ratio that wins games in casual pods (defined as pods where the average deck costs under $500 and games go to turn 8 or later):

• 45 total mana sources (lands plus 2-CMC-or-less ramp).
• 30 mana sinks (spells that cost 5 or more mana).

The gap between 45 and 30 is your "middle curve" (3-CMC and 4-CMC spells). That's where your interaction lives, your card draw lives, your early plays that aren't ramp.

If your deck has 38 lands and 4 mana rocks, you're at 42 mana sources. You're three short. If you're also running 12 six-drops and 8 seven-drops (20 mana sinks), you have enough top end. The problem is you can't reliably cast it.

Conversely: if your deck has 38 lands, 6 mana rocks, and 8 mana dorks (52 mana sources), but you're running 40 spells that cost 5+, you'll hit your land drops and then flood out because you drew three lands in a row when you needed a win condition.

The 45/30 rule is a starting point. Faster pods need more like 48/25 (you're racing, you need to curve out). Slower pods can get away with 43/35 (games go to turn 12, you have time to durdle). But 45/30 is the center of the bell curve for the format as currently played.

Why Cultivate Doesn't Solve This

Cultivate

You kept a two-land hand with Cultivate in it. Turn 3 you're still on two lands. Cultivate is stranded in your hand. You draw your third land turn 4, cast Cultivate turn 5, and you're now on turn 5 with 5 mana. The Azorius player hit their third land on turn 3, cast a Worn Powerstone turn 4, and is now on turn 5 with 6 mana.

Cultivate is a good magic card. It's in 35% of green decks per EDHREC data as of writing. But it doesn't fix the second-land-drop problem because it costs 3 mana. You need to have already hit your third land to cast it.

The ramp pieces that fix the problem cost 1 or 2 mana. Llanowar Elves on turn 1 means you have three mana on turn 2. Sol Ring on turn 1 means you have four mana on turn 2. Nature's Lore on turn 2 means you have four mana on turn 3 (you played your land for turn, then Nature's Lore puts a Forest onto the battlefield untapped, then you have two lands plus the new Forest).

Three-mana ramp is for decks that have already solved the early game. Two-mana ramp is for decks that are trying to survive until the mid-game.

The Five-Card Swap Heuristic

You've counted your mana sources. You're at 42. You mulligan to 5 cards twice a night and you're tired of it. Here's the fix:

1. Cut two high-CMC spells. Not your win conditions. Not your bombs. Your "I included this because it's cool" 7-drops. You know which ones.
2. Add two 2-CMC ramp pieces. Nature's Lore if you're in green. Mind Stone if you're not. Arcane Signet if you have it. Fellwar Stone if your pod plays enough basics that it's reliable.
3. Add one land.

You're now at 45 mana sources. Your mana sink count dropped by 2 (you probably cut two 6+ CMC spells). Your curve is lower. Your mulligan rate will drop from "every third game" to "every sixth game."

Run it for 10 games. Track how often you mulligan and how often you get stuck on two lands. If you're still stalling, you need to go deeper. Cut two more high-CMC spells, add two more ramp pieces. The floor for a functional Commander deck in 2024 is about 45 mana sources. Competitive decks run 50+. Casual decks that consistently make turn-6 plays run 46-48.

If you swap five cards and your mulligan rate doesn't change, the problem isn't your ramp density. It's your curve. You're running too many 3-drops and not enough 2-drops, or you're in a color combination that can't support the density you need (mono-white struggles here, Sultai does not). That's a different article.

When The Data Contradicts Your Experience

You've done the count. You have 46 mana sources. You still mulligan twice a night. Now what?

Go back through your last 10 games and write down what you mulliganed away from. If you're pitching hands with 2 lands and 2 ramp spells because "I need interaction," your mulligan criteria are wrong. If you're pitching hands with 4 lands because "this is a flood hand," you're mulliganing correctly but your curve is too high.

Birds of Paradise

The London mulligan rewards you for keeping marginal hands because you get to bottom the cards you don't want. A two-land hand with Sol Ring and Birds of Paradise is not marginal. It's a turn-2 four-mana hand. Keep it. Bottom your 7-drop.

If you're keeping those hands and still losing, the problem moved. It's no longer about hitting land drops. It's about what you're doing with the mana once you have it. That's past the scope of this article, but the diagnostic is the same: count the things, compare the ratios, find where the deck doesn't match what the math says it needs.

Most players who think they have a mana base problem actually have a ramp problem. Most players who think they have a ramp problem actually have a curve problem. The second-land-drop problem sits at the intersection: you don't have enough 1-and-2-CMC mana acceleration to smooth out the gap between your opening hand and your first real turn.

Fix the ramp density first. Then fix the curve. Then worry about interaction and win conditions. You can't cast a Cyclonic Rift if you're stuck on three lands.

---

This is a Commander Pulse evergreen. For weekly EDH data digests delivered to your inbox, sign up here.