How Lotus Vault works

Opening hand

The odds of keeping your hand, of being mana-screwed or mana-flooded, are real calculated probabilities. A hand counts as keepable with 2 to 5 lands — the calculation only judges the number of lands, not the quality of the spells. The odds of hitting a land drop on turns 2, 3 and 4 assume an acceptable hand, no mulligan and no cantrips. They answer 'have I drawn enough lands?', nothing more.

Mana sources

For each colour, I check whether the number of your sources matches the spread of your mana costs (your pips). I leave a 15% tolerance margin before raising the alarm. That's deliberately generous, especially since I count your mana rocks and mana dorks as sources in their own right. The idea is a radar that flags a manabase that's completely broken, and to leave the hypergeometric probability maths to Frank Karsten!

Power level

The score out of 10 is calculated from eight criteria. The three that matter most are combos, interaction and card draw; next come the mana curve, fast mana, tutors and Game Changers; salt barely weighs anything. Tutors boost the combo score, because a deck that fetches its combo pieces simply has them more often.

The consequence is deliberate: with no combo and no Game Changer, a deck tops out at 7/10. That's not a gap in the model. In practice a "normal" deck reads between 3.5 and 7, and the top of the scale is reserved for combo decks. The scale was calibrated by degrading a real decklist step by step, from an optimised build down to jank — not against a reference corpus.

Bracket

Wizards' own criteria are used as-is: no more than 3 Game Changers, no mass land destruction, no early two-card combo, to stay in bracket 3. I add two house caps of my own — tutors and extra turns — and, above all, the line between 2 and 3, which the official rules don't actually draw: they say what a deck must not play, not how sharp it is. So I look at five tuning signals — a low curve, dense interaction, heavy card draw, high salt, acceleration — and it takes three of the five to move up to bracket 3. Without that test, every "clean" deck would fall back to bracket 2, and a precon would be indistinguishable from a deck built to beat it. Imported precons are analysed exactly as they're built, for what it's worth, and can land in bracket 3 too.

The + or the − marks where you sit within the bracket, and a deck that maxes out its whole tier budget gets promoted to the next tier with a "−". If Commander Spellbook ranks your deck higher than my calculation does, it wins: it knows combo speed better than I do.

Salt and the deck plan

Salt comes from EDHREC's votes ("how much does this card ruin the fun"). It's the deck total that matters, not the average: the average gets diluted by the roughly 60 neutral cards — little or no salt — that every deck plays.

The builder's Plan proposes quotas by role. The "balanced" line — 37 lands, 10 ramp, 10 draw, 8 removal, 3 wraths — takes the median values that deckbuilding guides agree on. The other archetypes aren't sourced from anywhere: they're that same baseline nudged toward the archetype. Treat it as a starting point, not a truth. You can also generate the plan from your commander's EDHREC decks — there, the proportions come from real lists.