Mana Base Analysis is live under Tools → Mana Base Analysis. Paste a decklist or import from Moxfield, Archidekt, or TappedOut, and get a full read on how your manabase actually performs — not just how many lands you're running.
What's in it
Cast rate and tipping point. For each color in your deck, see the probability of holding a source of that color by turns 1, 3, 5, 7, and 10, computed with the same hypergeometric math used for combo-probability calculators — not a rough approximation. The Tipping Point stat gives you the earliest turn you're 90%+ likely to have enough lands in hand to cast your deck's average spell.
Mana curve and mana identity. A curve chart of your nonland spells, plus a side-by-side of pip demand versus total color sources (lands and rocks/dorks) so you can spot a color that's over-demanded relative to its fixing at a glance.
Full land breakdown. Every land in your deck sorted into categories — fetch, shock, check, fast, pain, gain, tri-land, battle, Gate, pathway, and more — classified directly from each card's own oracle text, so it works correctly even on brand-new sets.
Source Requirements (Karsten math, solved for your deck). For each color, we find the toughest pip demand anywhere in your list — say, a card needing WW by turn 3 — and solve for exactly how many sources of that color you need for 90% confidence, using the same hypergeometric approach behind Frank Karsten's well-known mana-math research. The difference: instead of reading numbers off a published 60-card table, this is computed live against your deck's actual size, so it's correct for 99-card Commander lists, partner commanders, or anything nonstandard. Right next to it: a recommended total land count from Karsten's 99-card regression formula, flagged if your deck is 3+ lands off.
Smart upgrades, now deficit-targeted. Pick a budget (No limit, or capped at $5 / $20 / $40) and hit Find Upgrades. The engine no longer just ranks candidates by speed and popularity — it scores every candidate land by how much of your actual color shortfall it fixes, re-scoring after each pick so five suggestions don't all patch the same color while another stays short. Cuts are chosen from a color you already have spare sources of, so a swap never quietly makes a different color worse. A before/after comparison and one-click copy (land base or full decklist) round it out.
Why quantities matter here
Manabase math lives and dies on accurate counts — "12 Island" needs to count as 12, not 1. This tool reads decklists directly (bypassing the shared analyzer) specifically to keep every basic land's quantity intact through the whole calculation, so cast rates and the land breakdown reflect your deck exactly as built.
It's also built on the same shared, rate-limited Scryfall client as the rest of the site, so browsing, filtering, and re-running upgrade searches never puts unnecessary load on Scryfall's API.