Slay the Spire and Randomness Tolerance

Over the last few weeks I’ve returned to Slay the Spire as my digital relaxation game of choice. It’s as brilliant and brutal as ever, and I’m bumping up against ascension levels that test the limits of my skills. If you haven’t yet played Slay the Spire, you should. It’s a game created with such a high level of thought and care that you can be forgiven for not noticing just how good it is. 

After a string of quick losses led to a rare triumph (thanks in large part to the ice cream cone–if you know, you know), I realized that Slay the Spire is a deckbuilding game with randomized card buy choices. I mean, I’ve known that since the first time I played years ago. It’s the central conceit of the game. But I never really understood the implications.

Let’s back up a bit. If you’ve been following The Thoughtful Gamer for a while you probably know that we’re something of Dominion apologists over here. I just placed it as my 10th favorite game of all time over on the podcast. It’s one of our Masterpiece Games. I’ve stated multiple times, both in print and on the podcast, that I think deckbuilding games that don’t utilize a static set of cards to buy are at an inherent disadvantage.

The one exception I can think of in the board game space is Mage Knight, which kind of sits in between the two extremes. It has a randomized card display, but it only refreshes each major phase of the game, giving you multiple turns to plan and scheme with knowledge of what’s available.

Two of the most popular deckbuilding games, Clank and Star Realms, suffer from their randomized selection, at least to my taste. Compared to Dominion’s strategic gait, where plans, formed at the beginning, violently shoot out into a race with only slim margins for mid-game maneuvering, Clank and Star Realms feel plodding and arbitrary. Plans almost necessarily have to shift and morph based on the whims of the draw, and one rarely feels the joy of constructing a slim, efficient killer of a deck. The gravitational force of variability pulls it all towards the middle.

Slay the Spire has many layers of randomness. The very first choice in the game, before you even start ascending the titular Spire, is generated from a few semi-randomly determined effects. The enemies you face are randomly selected, as are the winding paths that guide you up the tower. In battles you, of course, draw from a shuffled deck. The enemies’ attack patterns often involve some randomness (others repeat predetermined patterns). When you defeat an enemy you get to choose from a random set of cards. After beating an elite enemy you get a random relic. In both cases you may or may not receive a random potion. When you enter a shop you get a random display of cards, relics, and potions.

On occasion I feel like all of this luck conspires against me to create a near-impossible run. Other times I get the right relic at the right time and it spirals into easy dominance. But the former never bothers me very much (it typically steels my resolve) and the latter is great fun. Why is this the case when I tend to have much less tolerance for randomness in other games? I’ve identified a few reasons.

First, I’ve got to recognize how fine-tuned the random elements are in Slay the Spire. As far as I’m aware, hardly anything is purely random from a broad set. Card choices involve rarities, and the odds of getting certain rarities changes as the game progresses. Different sets of relics are drawn from depending on the source of the relic (elite, boss, etc). I believe there are even nested bits to help balance the unpredictability a bit, so getting an unlucky result early might cause the odds of getting a lucky result in that same category later (feel free to correct me on this). When you watch excellent Slay the Spire players you’ll notice how good they are at understanding where the risks are and what assets are likely to appear at different stages of a run.

Second, notice how judiciously Slay the Spire uses output randomness. There are a few cards that deal multiple bursts of damage to random enemies, and a couple of powerful relic effects that involve output randomness. The wonderful Snecko Eye relic is a fine example. You will always choose this relic from a set, so you’re always opting into the chaos at your own risk. It’s also quite powerful, because while it randomizes the costs of all of your cards each time you draw them, it also gives you +2 draw every turn. If that ends up destroying you, well, you chose the Snecko life.

Finally we get to the real topic of this piece: I think I have a larger tolerance for randomness in solo and cooperative games compared to competitive games.

As I type that thesis a number of questions and rebuttals spring to mind. My first thought is that I am perhaps thinking about uncertainty too granularly. Costikyan’s Uncertainty in Games demonstrates that uncertainty is at the core of game-playing, but I don’t think the discussion stops there. Certainly there are different objective and psychological effects of different types of uncertainty. Consider the difference between rolling a die and flipping a card off a deck that we discussed on last week’s podcast. Rolling a die feels more chaotic and random (perhaps stemming from the movement of the object) than flipping from a deck (which otherwise remains static). Assuming they both have the same odds parameters and are handled with sufficient randomness there’s no real difference between the two, but they generate slightly different responses from us. Of course, as soon as you continue flipping cards the deck does become objectively different–the card already drawn is excluded, changing the odds.

So, then, I believe there is a psychological difference in how we perceive uncertainty drawing from opponent’s play versus from a mechanism of the game. Perhaps I’m alone in this regard. How we react to games is a combination of individual elements and general human psychology. (I recall Omari Akil writing about how his experience as a Black American makes him more comfortable with output randomness in The Manifold). But there is an actual difference between these two forms of uncertainty. Even though we can’t necessarily predict what other people are going to do, we do know that their decisions are the product of intelligent thought rather than cold luck. Even if we were to reduce their decisions to probabilities we would only be making estimates. If they do something we didn’t expect we can analyze that decision and predict better next time, or gain an insight into the particular preferences and quirks of that person. If we get an unlucky die roll we got an unlucky die roll.

In that sense opponent uncertainty is superior in an educational sense, but we’re comparing randomness to randomness. In a competitive game, randomness exists alongside opponent uncertainty. Perhaps my lower tolerance is due to this additive effect in competitive games. I think there’s more.

Any competitive game is pitched as a test of person vs. person (vs person, etc), and outside randomizing elements distract from that core fundamental. A perfect information game tests this most purely, and games that introduce more and more randomness distract from it. I’m not making any value judgements here–it’s perfectly fine to enjoy either extreme and everything in between.

But when the core fundamental of the game is that you’re trying to beat another person, the mechanisms of the game are your allies in that quest. They’re the medium in which you’re trying to beat the other person. In a cooperative or solo game the mechanisms are still the medium of play, but more importantly they’re the enemy. Because our minds love to construct narratives so much, we perceive the game mechanisms as having more agency when they’re also our opponents. When one of the spire’s minions decides to initiate its big attack on the turn I draw no block cards, I think of it (to a degree) as the spire working against me. When the sandstorm takes a nasty route in Forbidden Desert we curse the winds. When I whiff all of my shots at a key moment in Memoir ‘44 I rolled poorly.

I fully submit that this analysis is entirely a result of my very limited knowledge in relevant psychological matters and my own introspection. I would love to hear from you if you too think you have a greater tolerance for randomness in cooperative and solo games. Now I’m going to think about how I could devise a way to scientifically test this thesis. Maybe one more go at Slay the Spire first.

Share this post

Share on facebook
Share on google
Share on twitter
Share on linkedin
Share on pinterest
Share on print
Share on email

Join The Discussion

This site uses Akismet to reduce spam. Learn how your comment data is processed.