Execution in strategy games should be considered “randomness”

I have long considered execution to be basically a form of randomness, similar to dice rolls or card draws. I have also long recognized that not everyone agrees with me, and some people find the idea deeply offensive. I can understand both of these responses, but today I would like to address them directly and see if I can demonstrate to anyone that it’s helpful to classify execution in this way.

The short version of my claim is this: an individual instance of a performance of something that requires execution skill in a game is random. A good example would be shooting in basketball. A player has a certain amount of skill for this task, but all that does is increase the odds of success. Odds as in, there’s a statistical, probabilistic distribution of outcomes—or, more colloquially,  randomness.


Randomness vs. Unpredictability

The most important component of randomness is its unpredictability. Are all things “unpredictable” random? Well, if we want to get technical, nothing is really random. We use the word random in strategy games to refer to elements that are designed to be unpredictable; that being unpredictable is something that we want. We add RNG to-hit to our combat systems to make the combat less predictable.

There are some things which are unpredictable, but which we probably wouldn’t use the word “random” to refer to. For instance, knowing where your current Go match will be in 20 turns is highly unpredictable. I would not use the word “random” to explain why that is, though. One reason it is unpredictable is that Go has an extremely huge possibility space and looking ahead very far is extremely difficult for humans. But really what you’re seeing unfold there is the game of Go itself happening, so to call it “random” would be a strange and not-useful application of this term. All games that are functional are technically “random” in that their outcomes are unpredictable. This would be extremely diluting to the term.


Another component of randomness

As I talked about in the recent episode of 3 Minute Game Design, “Systemic Meaning“, strategy games build meaning over time through player input. If you haven’t watched it, it’s only 3 minutes long so you could check it out below.

“Generally speaking, most things that players do in most games are deterministic. Within a sequence of deterministic actions, you build meaning. First you moved here, then you moved there. Randomness, on the other hand, can almost be defined as an “anti-meaningful” event. In a typical turn-based tactical game, if someone asks you why you moved to a particular tile, you can explain it to them. There is meaning behind it. If you attack and miss, you can’t give them any insight as to why you missed other than “that’s just how the dice rolled”.”

This element of “disconnectedness from the greater system” is also an important component of what makes something random. A dice roll’s outcome is totally independent of any other conditions in a strategy game – it is this technically meaningless piece of information that has not had any time to build any meaning before it delivers its permanent mark on the match.

The broader point is that strategy games have a skill to them, which is in having some kind of mental picture of the system and how it works when you do certain things in it. This is distinct from the question of your physical body’s ability to perform certain tasks. Strategy games ask the question of “should you do X?”, fundamentally. Sports, contests and other systems may ask this should question, but if they have execution, they’re also asking “can you do X?”. It is not as though placing some of those “can” questions in a strategy game completely ruins it or anything, but it creates a situation quite similar to adding a dice roll into a strategy game (which I should remind everyone that not everyone considers a problem either!)



The most common response to my claim that execution is a form of randomness is something along the lines of this: “execution isn’t randomness, it’s skill”. The argument is that because you can get better at execution, it’s therefore not random. Of course you can get better at shooting baskets, but on each individual shot, there is still variance and unpredictability; all you can do is increase your odds with training. But from a strategy game design perspective, that’s no different than the fact that in X-Com, some of your soldiers have a better chance to-hit because of their RPG stats (which could be framed as their “training”).

I think a lot of the disagreement on whether execution should be considered randomness has to do with the particular aesthetics of different kinds of games; particularly strategy games vs. sports. For example, in an FPS with strategy game elements, something like Overwatch, it feels like the execution is pretty central to the game, so to call it “randomness” feels wrong.

In the context of an obvious strategy game, what I’m saying becomes more clear, I think. If we said that you have to toss a chess piece in the air and catch it behind your back anytime you made a capture for it to “succeed”, I think we would agree that this is adding randomness to the game. Yes, you can get good at it, but there would always be some percent chance of failure.



If you’re talking about some sport, or a fighting game, or an FPS; anything where execution is a major, arguably fundamental aspect of play, then maybe it doesn’t make sense to call the execution randomness. But in strategy game design, and strategy game design theory, execution is randomness, because it cannot possibly be anything else in the context of a strategy game. Execution is always a matter of personally training some physical ability, whether it be muscle memory, strength, reflexes, etc. There is no way that a particular player’s ability here can be implemented into a strategy game in a way that’s strategically meaningful. You could balance around it to get it back to something like a 50% win ratio but it still needs to basically be factored out and then treated as a kind of output randomness.

An interesting side note: it is technically possible that we could use the randomness of execution as a source of randomness, but build around it in such a way that the execution is input randomness. In just about every situation that I can think of right now, the execution is in an “output randomness” form, but this doesn’t have to be true, I don’t think. With that said, it’s still probably inferior as a source of randomness compared to RNG, since:

  • As the player gets better at the execution the distribution changes, which is weird because again you’re bringing this external stat into the strategy game, gaining extra resources that weren’t earned in the context of a match
  • It is variable, not uniform randomness

Overall, I think there’s a place for games in general that have execution, but strategy games should, as they move away from variable output randomness in general, also move away from execution for precisely the same reasons.