Schrodinger's Axe

Hello and welcome to my first blog post. About a small (theoretical, if any) design flaw within the Fire Emblem game series.

I want to preface this post by saying this is exclusively about pure strategy games, and that I consider a pure strategy game to be a game with no elements of “Execution”. Real-time or reaction-based elements are what I consider to be elements of execution.

In the game, your units have a certain percent chance to hit, and a certain percent chance to miss. Most units in most situations have a >85% Hit chance. But sometimes, bad decisions can result in your units having a 50% chance. You are able to see the hit chance before you start the attack, meaning you can back out if you don’t like your odds.

39% is not a good chance. (Wikipedia.com)

39% is not a good chance. (Wikipedia.com)

Let’s say I am in a position where my Unit has an 80% chance to hit. If my attack lands, I win the game. If my attack misses, the enemy retaliates and I lose the game. My options are to either attack or to run away. Most people would agree that attacking is a good idea, as the Risk-Reward is in your favour. Fire Emblem is a game where many such decisions are made, and so over the course of the game, someone who consistently attacked in that situation would be more often than not, successful. Picking the statistically best option is the optimal way of playing Fire Emblem.

Now, onto Schrodinger’s Axe. The next situation is exactly the same, but now I have only a 50% Chance to hit. I have the same two options as before. The problem here, is that choosing to attack can only be severely rewarded, or severely punished. Despite being an average decision. Additionally, there is no statistically better option, which means that there is no best way to proceed. Since there is no best way to proceed, you are basically flipping a coin, and that is not healthy for its design.

This of course makes perfect sense within the scope of Risk-Reward. But within the context of Fire Emblem, a game designed from the ground up to reward statistically advantageous play, it falls flat. It is worth mentioning that over time, a player who consistently makes that decision will be rewarded with equal victories and defeats, and as such, the player will find their average result. But this is little comfort for the player faced with the situation.

Choosing to attack is a decision that is only proved intelligent or dumb after the event. This is similar to Schrodinger’s cat, in the sense you only know the result after you open the box (or execute the attack), and since it cannot be an average of the two, it cannot be rewarded as such.

And so what can Fire Emblem do to fix this? Should they do something?

Truth be told, the situation of Schrodinger’s Axe only rarely occurs, and is more a conceptual problem than a realistic one. If I was working on Fire Emblem and a fellow game designer explained the above to me, I would laugh and tell the programmers to change 50% accuracy to 51% (And thus restoring a statistically better option).

And so we reach an interesting conclusion. A game designed to reward good decisions must enable the player to make good decisions. If a player is capable of making good decisions that means they must also be able to make bad decisions. If a game allows good and bad decisions, that must mean there is an optimal way to play it (Picking 100% Good decisions). And if a game is optimised, that means there is a dominant strategy and the game is inherently unbalanced.

Strategy games are all unbalanced, and the essence of the game is finding out how?

This was honestly a bit of an experiment for me. I’ll write more about real design problems in the future, I promise.