Skill explosions are all around us. Just compare Simone Biles’s 2024 gold medal vault with Larisa Latynina’s 1956 gold medal vault, set to delightful race car vrooms:
A lot of this is to be expected. Over time, as we recruit larger talent pools and improve on best practices, our practitioners grow more sophisticated. (Steph Curry couldn’t have existed in the sixties.) And today, the fruits of our most successful practices are sweeter than ever.
The same is true in the world of board games. The most striking thing about the Board Game Geek Top 100 is how recent all the games are. Most are from the last five or ten years. Only one game (El Grande, number 91) is from the twentieth century!
If I say “board game” and you think Monopoly or Risk, please think again.
This is not at all true of the Internet Movie Database Top 250, by the way—as you might have noticed, movies aren’t doing quite as hot this century, Deadpool & Wolverine be damned. Skill explosions aren’t automatic. These lists are obviously imperfect, but I think this is evidence that the upward trend in board games is largely real and not just Zoomers overrating their recent faves on the World Wide Web.
Deck-builders are less than twenty years old, which is so new you might not know what they are. But they’ve taking the board gaming world by storm.
Here’s the basic idea:
You start each game with a small deck of bad cards (usually 10).
Each turn, you draw a hand of cards (usually 5) to play.
During the game, you improve and personalize your deck by adding new cards and removing old ones.
Now when you draw a hand and it’s good, you built that! And if it’s bad, why’d you build that?!
This simple gameplay loop is really compelling and oh my God, Steam says I’ve logged 600 hours in Slay the Spire.
Sometimes deck-building is the entire game (e.g. Dominion), but lots of games are now including deck-building elements alongside other mechanics (e.g. Dune Imperium).
Deck-building’s definitely trendy right now, but the trend has been going for a while, and the idea itself is so flexible and powerful that I think we have more than a gimmick on our hands.
One reason for the tremendous success of deck-building as a gameplay mechanic is the unique relationship it gives players to randomness. They are, in some sense, responsible for the cards they draw.
Okay, at a very high level, all games include some variance. Even in chess, where we have a perfectly deterministic grid system that you can try to calculate out infinitely far ahead, there are still weird random quirks like,
“Oh I didn’t even see that move.” —famous last words
The classic way to increase variance in games for the past several thousand years has been to roll dice.
I’m sure you know that dice rolls are independent—if you got a 6 last time, and the die is fair, you’re just as likely to get a 6 this time. (And yes, the singular form of ‘dice’ is traditionally ‘die,’ it always weirds me out, but at least other people are weirded out too because now it’s starting to change.)
The math on this is stunningly recent by the way. Mathematical study of probability only really emerged in the 16th and 17th centuries as several all-time mathematicians took a break from gambling to go, wait, how do the odds work in these dice games anyway?
Anyway, playing cards are actually much newer than dice—if you think about it, you have to be pretty good at manufacturing and printing to make a bunch of uniform rectangles.
But card draw is not independent—if you drew a King of Hearts last time, and you haven’t shuffled yet, guess what? If the deck is fair, there’s not another one in there!
To summarize:
Dice rolls are statistically independent.
Card draws are statistically dependent.
Even so, we’re very good at reading stories into dice rolls.
There’s even a fancy word (apophenia) that names our tendency to look for and ‘discover’ meaningful patterns in random data. As you roll dice repeatedly in a game, things will tend to average out in the long run, but you might still get some weird streaks or extreme events, like rolling a 1 three times in a row.
We can read stories into card draws, too. But there’s also a certain sense in which card draws can actually give us stories.
Let me explain.
In games like Monopoly, Risk, and D&D, one dice roll = one action.
You wanna avoid Boardwalk, or invade Greenland, or persuade the shopkeep? Alright, roll the dice and let’s see how it shakes out.
In Risk, you roll die number 1 to see if attacker number 1 hits a defender. And you roll die number 2 for attacker number 2.
And die number 3 for attacker number 3.
Since dice rolls are independent, their maneuvers don’t affect each other, and no, dice colliding doesn’t count because it doesn’t change the odds!
But the previous history of that conflict has no influence either, except for determining how many soldiers are still standing. With dice, each action you take is statistically cordoned off from each other action you take.
The independence of dice rolls is a blessing and a curse. Dice don’t remember what came before—talk about living in the moment! And sometimes that’s exactly right given what a game is going for.
But not always.
Notice that cards do remember what came before. As you repeatedly draw through your deck, you’ll see each card the same number of times (barring hijinks). What varies is when you see it, and in what context.
The real difference with card draw is that the unit of randomness is larger.
With dice, the unit of randomness was a single roll, and there were only as many possibilities as there were sides on the die.*
But with cards, the unit of randomness is the ordering of the entire deck, which is cosmologically larger. Here’s how many orderings there are of a standard 52-card deck:
80,658,175,170,943,878,571,660,636,856,403,766,975,289,505,440,883,277,824,000,000,000,000
Even with your lousy starting 10-card deck, there are still 3,628,800 possible orderings, which is admittedly a bit of an overcount. If you’re drawing it in hands of five at a time, that’s still 252 starting hands. With duplicate cards, the functional count may be lower.
But this extra internal complexity changes your relationship to the variance that cards introduce:
1. You can learn more about the deck ordering as you draw through it.
Wow, I haven’t drawn my power card yet. Well, I know it’s in here somewhere, so it’s getting more likely.
The dependence of card draw forces players to make deeply contextual choices based on where they think they are in both their deck and the larger game. And in a deck-builder, I may not have total control, but the current variance in my hand is tied up with my own previous choices.
The story here is one of my historical choices seeding the randomness I have to manage now. I built my deck. So of course it gave me this hand.
2. You can have complex internal interactions between cards.
This is where deck-builders really shine.
The ordering of the deck determines what hands you’ll draw, as well as the context in which each card’s individuals powers can activate.
This makes even simple interactions powerful. Say I have a card that lets me draw two more cards.
If it’s the fifth card in my deck, I’ll draw it as part of my first hand. Now if I play this card, I’ll make my first hand larger. Functionally, the sixth and seventh cards are added to my first hand.
If it’s the sixth card in my deck, no luck! I don’t get to see Card 6 yet, which means I’m moving through my deck a bit slower so far. So I have fewer options on Turn 1, but I’ll get more later.
And again, in a deck-builder, the interactions in my deck are tied up with my own previous choices. Over time, my deck becomes an artifact that records my past choices (I built this deck), and shapes my present powers (I’m holding this hand).
My choices led to a deck with these interactions, which have brought me exactly this far along. If I can keep improving my deck fast enough to stay ahead of the game’s increasing difficulty, I can keep shaping my own historical choices into continual relevance and growth.
Unlike dice, decks have a memory. And deck-builders use this to record your history of choices and reliably give you the tools you have chosen to live with.
Now if my deck does well, it’s because my ongoing decisions keep it ready for everything coming its way. And if things go off the rails—well, I guess circumstances changed too quickly for my deck to handle. Whoops.
That’s an incredible relationship to variance. Good and bad luck no longer just happen to you out of the sky. You loaded the dice, so to speak.
Of course, really great deck-builders have much more fascinating interactions than a bit of extra card draw. In Slay the Spire, I’ve gotten very good at creating elaborate stance-changing decks, but if you want a more immediate sense of the game-breaking interactions that can happen between cards, there’s no better place to turn than a speedrunner:
I still gotta beat Ascension 20. Maybe I’ll try again soon on Twitch.
Many games have you roll multiple dice and add the results to make more bell-shaped distribution curves. (Same deal with modifiers—we’re just trying to shape and shift a distribution.)
If you roll 1d6, each outcome from 1 to 6 has a 1/6 chance. That distribution is perfectly flat. But if you roll 2d6, the most common outcome (7) happens 1/6 of the time, and the least common (2 & 12) happen 1/36. Wow, we’re starting to get a bell curve!
But if all we can do is add together statistically independent randomizers, that’s still only one kind of randomness. And it remains a sort of randomness that externalizes the results of randomness from my own actions. Sure, maybe bad choices have led me to have to pass a difficult dice roll. Oops, I needed a 20 and only rolled a 19.
But mechanically, I don’t think that’s as devastating or elegant as drawing into a big hand of nothing right as the boss takes a final swing. I know exactly who put those lousy cards in my deck. The bad choices are coming from inside the house, and now they’re staring me in the face.
Loved this article. One of your best for sure.