“Nah, I don’t like that deck, it has consistency issues”
“This deck can consistently kill on turn 3!”
“You can’t play that on curve consistently”
“Consistency” is a term we use time and time again in our Magic lingo, but do we actually know what it means? Of course, you have an idea in your head of what it is supposed to mean. Something about doing the same thing every game, the opposite of variation, everything that Commander isn’t. But can you really define it?
In this article I will attempt to voice my thoughts on this subject which goes much deeper than you’d initially assume.
It’s nice to be able to cast your spells
When it comes to manabases, we have a pretty clear idea of what consistency means. Your lands have one job and that is to cast your spells. And if they don’t do that often enough, your manabase is unplayable, which means that your deck is unplayable. So for this purpose we can define “consistency” as just a number. The mathematical likelihood that you have the amount of lands and the colors you need on the turn you need them.
Ideally this would be 100%, but even a deck like Mono Red doesn’t draw the perfect amount of lands all the time, but at least it always has the right colors.
Frank Karsten (and even I as you can see here) have written a number of articles on this subject and the general baseline for what we consider consistent lingers a little above the 90 % mark. This means that in 1 out of 10 games you will not be able to cast the card you’d like on time. In reality it’s more games, because sometimes you will lack double blue for Absorb, some games it will be the double black for Kaya’s Wrath, sometimes it will be the 5th land for Teferi, Hero of Dominaria that will be missing. This adds up to a lot of games where some part of your game plan doesn’t function optimally, because you ask a lot of different things from your manabase. Maybe up to half of your games, something will be suboptimal, and that’s a pretty high price to pay for power.
This sheds a light on why mono colored decks (especially ones with mana sinks, so they get to run lots of lands), are often very attractive.
Consistency is speed
What did Einstein say about this?
I’m pretty sure that with E=MC², Einstein actually meant: Efficiency=Much Consistency².
Now let’s consider a combo deck like Modern Grishoalbrand. It attempts to reanimate Griselbrand on turn 2 with Goryo’s Vengeance and win from there. The deck is widely regarded as inconsistent. The thing to consider here, is that it’s not inconsistent regarding whether or not it achieves its combo. Rather it is inconsistent regarding how quickly it can assemble the lethal set-up.
It’s not like in the games where the deck has bad draws that it would just roll over and die to a goldfish. It would get there eventually, maybe by turn 6, but most often that means it will have lost by that point.
If we imagine that the deck is up against a literal goldfish and just trying to combo out as fast as possible, we could then figure out how likely the deck is to win on any given turn – in theory at least, there are too many variables to actually figure this out. Let’s say for the sake of argument that Grishoalbrand boasts the following numbers:
10% likely to win turn 2
25 % likely to win turn 3
35% likely to win turn 4
30% likely to win turn 5
Realistically, the numbers would be more nuanced and go up much higher than turn 5, but they’d all go up to 100% because it will have to eventually win at some point.
Now let’s imagine that our Grishoalbrand deck is put up against a Burn deck. This deck isn’t nearly as fast, as it can never win on turn 2, but it’s also full of many of the same cards which almost always leads to a turn 4 kill. Let’s say this hypothetical Burn-deck is 20% to win on turn 3 and 80% to win on turn 4. Again, it would go up much higher, but I don’t think this is too unrealistic.
The head to head
3, 2, 1, fight!
In this imaginary pit fight we say that the decks have no way of interacting with each other, so it is just a pure race. While you would think that Grishoalbrand should have the upper hand – be faster, so to speak – because it can win on turn 2 and is more likely than burn to win on turn 3. But if you add up these numbers, you’ll see what each deck’s average kill turn is.
For Grishoalbrand, its average kill turn with our hypothetical numbers is 3.85.
For Burn it is 3.80.
This means that Burn is slightly favored in the match-up, because it is realistically faster. Not by much, you’d have to pit these decks against each other thousands of times to see the difference, but Burn is ahead.
If you add up all the probabilities that we get from examining consistency, you don’t end up with an overall “consistency score”. You end up with a number for the amount of time it takes for the deck to achieve its goal. Burn achieves its goal faster on average than Grishoalbrand because it is more consistent.
This is part of why Grishoalbrand is widely regarded as unplayable, because when push comes to shove, it’s actually slower than a lot of the other decks it’s trying to race. And this is also why a deck like Izzet Phoenix, which basically can’t kill turn 3 but has many roads to a turn 4 kill and lots of cantrips to consistently find one of these roads, keeps performing and outracing decks that look they should be “faster”.
How about the interactive decks?
Now, talking about a goldfishing kill turn doesn’t really describe the majority of decks in Magic. Most of the time decks like to interact with each other, but what then does consistency mean if we are not talking about pure race situations?
Consider a deck like Modern BG Rock. It’s widely perceived as a consistent deck, but it doesn’t play any cantrips to find the right cards more often, the way that so many decks in Modern achieve their consistency. Instead it just runs duplicates of its effects while being a straight 2-color deck with a clean manabase. If you don’t draw Tarmogoyf, you’ll be satisfied with Scavenging Ooze instead. If you don’t have Thoughtseize in your opener, maybe you have Inquisition of Kozilek. You can use both Dark Confidant and Tireless Tracker for card advantage, and Fatal Push and Assassin’s Trophy both get rid of creatures.
This is why I stated consistency as how quickly on average you achieve your goal, rather than how fast you can win the game. Clearly these decks aren’t racing most linear decks, but they don’t need to because that is not their goal. Rather, the goal is to answer opposing threats and the bury the opponent in card quality and quantity once you get to a topdecking state. In order to measure the consistency of this, you need to figure out how often you have the answer to the questions that are asked of you on time.
This concept is even more expressed with a deck like Standard Esper Control, where you won’t have a problem drawing plenty of answers, but you have to have the right ones as well. It’s not every time you have Kaya’s Wrath ready turn 4 that it is the answer you need, maybe you’d rather have an Absorb for a planeswalker. This means that not only will control decks have to juggle drawing the right amount of lands and getting their 3-color mana base to fit, they also have to deal with inherent inconsistency in playing reactive cards.
After all, if your opponent’s average kill-turn is faster than your average sweeper-turn, you might end up with a problem. And such, reactive decks can still be considered under the lens of consistency is speed, with the added inconsistency provided by the possibility to have the wrong answers.