Build Commander
Roughly 2.5 to 3.3, counting your spells and not your lands. It's a useful single number to check a decklist against, but it's a summary, not the whole picture.
2.5 to 3.3 average mana value, measured over your nonland cards only, is a reasonable range for most Commander decks with no specific speed agenda. It's a fast way to sanity-check a decklist: add up every spell's mana value, divide by the number of spells, and see where you land.
This number is a summary, not a target to hit exactly. Two decks can share an average of 3.0 and play completely differently, one with a tight cluster of 3-drops and the other split between a pile of 1-drops and a handful of 6-drops. Average mana value tells you roughly how expensive your deck is; it doesn't tell you the shape. See what makes a curve for the shape question.
Comprehensive Rules · Rule 202.3
"The mana value of an object is a number equal to the total amount of mana in its mana cost, regardless of color." An object with no mana cost, which includes every basic land, has a mana value of 0.
With roughly 37 lands in a hundred-card deck, folding them into the average pulls every deck's number down toward the same meaningless middle. Measuring average mana value over nonland cards only is what makes the number actually describe your deck's spells, which is the thing anybody asking this question is actually trying to learn.
A deck's average should follow its plan, not the other way around. Picking a target average and then forcing cards to fit it backwards is how a deck ends up full of filler; picking a plan first and checking the average afterward is how you catch a curve that's drifted higher than the deck can actually support.
Too high for your ramp count. If your average is climbing past 3.3 without a real ramp package behind it, see how much ramp to run before cutting cards, since the fix is often adding acceleration rather than lowering the curve itself.
Too low for the deck's power ambitions. A curve that's very cheap across the board can mean the deck is missing its own payoffs, worth checking against how many win conditions you're running.
Right on average, wrong in shape. A deck can land in the healthy range and still play badly if its cards cluster at the extremes rather than filling the middle. That's a curve problem, not an average problem, and the fix is different.
An average is a sum divided by a count, so it helps to see what's actually on either end of it.
Sol Ring ↗
1 mana value. Cards this cheap are what let a deck run genuinely expensive payoffs and still land in a healthy average.
Eternal Witness ↗
3 mana value. Sits right in the middle of the typical range, doing real work for a cost that barely moves the average either direction.
Craterhoof Behemoth ↗
8 mana value. One card like this costs almost nothing against a hundred-card average, but a deck can't run many before the average climbs past what its ramp can support.
Roughly 2.5 to 3.3, excluding lands, covers most decks with no specific speed agenda. That's a starting range rather than a target to hit exactly: an aggressive deck wants to be lower, a ramp or control deck can comfortably run higher, and both are correctly built.
No, and this matters a lot. Lands have a mana value of 0, and roughly a third of a Commander deck is lands, so including them drags every deck's average down to a number that doesn't describe the spells at all. Average mana value should be measured over your nonland cards only.
No. A lower average means faster, more consistent turns, but it also means less raw power per card. The right average depends on your deck's plan: an aggressive deck wants low, a deck built to grind out a long game can profitably run higher, as long as it has the ramp to get there on schedule.
Related, but not the same question. Average mana value is one number that summarizes the whole deck; mana curve is the shape of the distribution behind that number. Two decks can share the same average and have very different curves, one bunched in the middle and one split between very cheap and very expensive cards, and they play completely differently. See what makes a curve for the full picture.
The Deck Lab's blueprint measures your deck's average mana value against your template's target as you build, nonland cards only.
Open the Deck Builder