How Card Advantage Works, Part 2: Picture of Consistency

Literally a picture of consistency… or in this case, inconsistency:

I have gotten a lot of comments about this screen shot.

What the heck is going on?

Why did I put that in the blog post?


Imagine if you will that this was your hand and board, instead:

  1. You are attacking with a Firebrand Ranger.
  2. You have three basic Mountains in play.
  3. Your hand is two basic Mountains and an, um… Tribal Flames (there is no burn spell bad enough to use a stand-in).
  4. Your opponent is at 19.


How bad is it?

You would be hard pressed to win against a well-built sealed deck with those tools only.

And while “cards in hand” have a non-zero value if for no other reason than you can bluff, that is what was depicted in the previous post’s screen shot.

So why am I bringing this up? Surely there is something productive to talk about beyond “Blood Moon is good against Zoo.” And there is! Zoo is just a perfect example.

I told Josh today that I took out one of the Sacred Foundries from the Zoo deck I listed last time around and replaced it with an Overgrown Tomb. This was night and day better than the previous version, just one card different. Why would this be, and why would I have previously played two Sacred Foundries?

The bonus to this deck from having an overgrown Tomb is simply that it can run another set of “opposites” to cast spells. Zoo is a much more challenging deck to play that it seems at first glance, to a novice. The reason is that in most Zoo games you will have access to 10-13 mana total… That is over the course of the whole game. So you have to make sure you have lands that can cast your spells.

A two land combination that functions pretty well together is Sacred Foundry plus Overgrown Tomb; not the best, but pretty well. Ironically, Sacred Foundry ha[d] no natural partner. If you only have two lands, you are likely to have an extra source of Red, but be unable to play half your hand.

To wit:

Godless Shrine goes with Stomping Ground.

Temple Garden goes with Blood Crypt.

And now… finally…. You have the option to play Overgrown Tomb and Sacred Foundry. Of the three combinations, this is the worst. You can’t play Lightning Helix and you can’t pump Viashino Slaughtermaster. But the reason I didn’t have it is that I was phoning in my mana base reminiscing about pre-Berlin testing, when Overgrown Tomb couldn’t pump Figure of Destiny. It’s the worst land in the deck, but not applicable to this deck.

Anyway, point being, if you get the wrong lands early, even though you have access to ten to twelve taps… You don’t actually get to cast a lot of your spells.

And when you can’t cast a spell… It’s almost like it wasn’t there at all.

  • Why don’t we play Nicol Bolas in Zoo?
  • Why are fast decks with low mana costs consitently better performers than ponderous mid-range decks that play a lot of weird and expensive stuff?
  • Why did I cut some Incinerates from Naya Burn, replacing them with Tarfires?
  • Why does Cruel Ultimatum leave a bad taste in GerryT’s mouth?

The answers to these questions aren’t all exactly the same, but they are pretty closely related. In essence, the value of a card is very closely correlated with our ability to utilize it. We don’t play Nicol Bolas in Zoo because it is essentially impossible to cast; ever drawing it would be a mulligan.

Why are fast decks with low mana costs consistently better performers than ponderous mid-range decks yadda yadda yadda? Because if the slower deck stumbles, it loses more than a potential land drop: It loses the efficacy of the cards in its hands. On balance the fast deck “stuck” at two lands will usually be able to knock over an entire city, let alone slow mid-range opponents. Time is also an issue. It doesn’t matter if the slower deck is packing Future Sight or just Future Sight cards.

So card utility — an element on which card advantage inextricably relies — has to do with one three-letter word: Now.

We don’t play Nicol Bolas in Zoo because we can’t cast it now (well… in this case, ever).

Fast decks with low mana costs consistently out-perform ponderous mid-range decks because the fast decks can typically use their spells now whereas the ponderous decks often have their hands clogged, doing nothing, for many turns. When they are mana screwed, fast decks can usually still play a lot of their spells; on balance, the slower decks go from being unable to play their spells this turn to being unable to play their spells ever. Why? Because the game is over and they are dead.

Why did I cut some Incinerates from Naya Burn, replacing them with Tarfires? This one is subtler. I could usually play the Incinerates… But because the Naya Burn deck would often have to operate with only two or three lands in play, I wouldn’t necessarily be able to play the Incinerate and something else. That took immediate utility away from me (even a little bit)… But with three mana I might be able to play a Tarmogoyf and get a blocker out of the way with a Tarfire so I could get in for three the same turn. In a format like Extended — which is essentially all about racing, in so many matchups — not being able to play the right spell this turn might as well be like being unable to play that spell ever.

The thing that I like best about this line of thinking is that there are immediate practical applications.

The most obvious one is how it might affect your mulligan algorithm.

Now a lot of us play with general rules like “you have to keep every hand with two lands” … Let’s see how those theories work out when we think about how many cards we functionally have, based on our ability to cast our spells right now.

(These screen shots are all courtesy of my Domain Zoo deck. Deck list at the bottom.)

This opening hand is about all you could ask for. You go and get Overgrown Tomb with the Bloodstained Mire and lay out a 1/1. You are striking for three on turn two thanks to the Sacred Foundry. This hand can’t fundamentally pump the Viashino and can’t cast Lightning Helix… But you don’t have Lightning Helix.

This second hand isn’t too bad. I know that your instinct is to not Mulligan it. That instinct would be correct. But there is a very concrete reason why you wouldn’t mulligan it. When you mulligan, you are trading this hand for a six-card hand of unknown quantity. This is a six-card hand that’s actually pretty good (you have about a one-in-four chance of speculating it into a full-on seven-card hand with the right topdeck).

Six card hand? Huh?

To wit:

The “seven” card hand only has the immediate utility of a six-card hand. With an Overgrown Tomb the natural land to go and get is a Sacred Foundry. You can’t play Lightning Helix with those lands.

This next “six-card” seven-card hand is a little bit worse than the previous one. I’ve already grayed out the Might of Alara. You actually get to play Lightning Helix for the first time, but with no Green mana you can’t play the Might of Alara. I would not mulligan this hand because you actually have some action… But unless you get to Green, you not only can’t play the Might, you can’t pump the Viashinos (so you basically have some Firebrand Rangers). It goes without saying that the Tribal Flames is a mere Volcanic Hammer… But there’s nothing wrong with that, expressly.

This hand is obviously poop. Clear mulligan – no action due to having no lands. Let’s examine the mulligan hand:

You would have to deeply consider going to five cards on this one. Of the six cards present, two are completely grayed out and I did a little half-thingie on the Wild Nacatls (no one is excited about a 2/2). I would be more inclined to keep this hand on the draw. 

Just something to think about:

  • You have a little bit better than a one-in-three to draw a land on your first pull.
  • Unless that land is a Blood Crypt or a Blood Crypt proxy such as a Bloodstained Mire or Wooded Foothills — NOT a Windswept Heath — the marginal utility nothing to write home about (by the way a Windswept Heath can’t get a Steam Vents, either). So now you are down to 16% to pull the land “you want” on your next pull.
  • If your hand doesn’t improve quickly, you are certain to lose to any competitive Extended deck.
When we talk about “doing the math” … This is what we are talking about. Are you better off with an unknown five-card hand or one of the above percentages? I would be hesitant to mulligan… It would depend on more than a screen shot in the abstract.

So when we talk about consistencyI think that these black-and-white images are what we are talking about. The decks we think of as less consistent play with functionally less card advantage, at least from an opening hand perspective. Now usually they are paired with greater power… There is no doubt that a Zoo deck that can attack for lethal damage on the third turn is “more powerful” than a ho hum Naya Burn deck that needs the stars to align very nicely in order to get a fourth turn concession (not necessarily kill), but at the same time, its ability to set up those kinds of draws with so many five-card openers means that it might have certain disincentives for play.

Think of the decks we complain about most in terms of “more mulligans” … A lot of those decks only play 20 lands (or not even 20 lands). Unless they are Elves (a deck of all one drops) these decks often have problems getting past Stage One (“basically manascrewed”)… And even a modest Stage Two deck will habitually beat a manascrewed opponent.

One more time for the road:

2 Umezawa’s Jitte

4 Dark Confidant

4 Lightning Helix

4 Might of Alara
4 Tarmogoyf
4 Wild Nacatl

4 Kird Ape
4 Mogg Fanatic
1 Seal of Fire
1 Tarfire
4 Tribal Flames
4 Viashino Slaughtermaster

1 Blood Crypt
4 Bloodstained Mire
1 Godless Shrine
1 Mountain
1 Overgrown Tomb
1 Sacred Foundry
1 Steam Vents
1 Stomping Ground
1 Temple Garden
4 Windswept Heath
4 Wooded Foothills

2 Umezawa’s Jitte
4 Duergar Hedge-Mage
4 Ancient Grudge
1 Volcanic Fallout
4 Ethersworn Canonist


How Card Advantage Works, Part 3: How All-in Red Works

That, or my deck for this weekend… One of the two!


facebook comments:


#1 StaplerGuy on 03.12.09 at 8:31 pm

So was it Swans?

You never said 🙁

#2 admin on 03.12.09 at 8:37 pm

Yes it was Swans obviously! 🙂

#3 DavePetterson on 03.13.09 at 2:07 am

Actually, he did say it was Swans in the first post.

Very interesting series so far. Mulligan theory is obviously important, but it doesn’t get a whole lot of attention. I guess a big ‘thank you’ is owed to GerryT for getting you onto this kick. I think this and your response to the Zack Hill article on card advantage are going to be similar to Sligh/mana curve in a few years, with people saying “Did we really used to think about Magic in any other way?”

#4 mpace on 03.13.09 at 8:15 am

i really like this. immediate use has always been a big card evaluation for me. my question would be, how does this transfer over to standard and even more importantly (for me) limited. it seems to apply much less considering with the higher mana costs (isn’t that the same as not having the appropriate colors?).

#5 wrongwaygoback on 03.13.09 at 3:06 pm

Surely there’s got to be some sort of cost/benefit payoff – that’s the discussion I’m looking for.
If you had a deck that was 40 blood pets and 20 swamps, sure it would be (a) consistent (b) entirely interactive (ie each blood pet interacts with each other), easy to cast, immediately relevant – but would it win?

If ease of casting and immediate relevance is the trump card, why doesn’t affinity win every matchup?

Where’s the model for what matters and what doesn’t?

#6 Five With Flores » Part Two, Who Wants a Loxodon Hierarch? on 05.01.09 at 10:37 pm

[…] This, my friends, sounds notoriously like a question of How Card Advantage Works. […]

You must log in to post a comment.