The Real Problem With Output Randomness Is Unrewarded Effort

In game design circles, you will often hear "use input randomness, not output randomness." Input randomness sets choices up while output randomness resolves them; think of drawing a tile in Carcassonne versus rolling for combat in Risk. The argument against output randomness is that it negates player agency by making outcomes arbitrary. Yet it still plays a role in many well-regarded games. The real problem with output randomness is not that it removes choice, but rather that it often leaves player effort unrewarded.

Our brains seek behaviors that produce the highest reward for the least effort. In games, this means greater engagement when our choices lead to better-than-expected outcomes. When it comes to output randomness, spending time and energy setting something up only to have it nullified by a bad roll leads to an unsatisfying outcome and a desire to do something else instead.

Output randomness is not inherently unrewarding; randomized rewards can actually form stronger habits than static ones. The trick is to make sure something always happens, even on failure. Players want to see progress, and failures that change nothing make the game feel like it is stalling.

One approach is to give players something when they fail that offers a new, different opportunity. This reframes failure as forward momentum in a new direction. In Cosmic Encounter, if you reveal a negotiate against an attack, you lose the battle but get to take cards from your opponent's hand as compensation. These new cards open up options that may lead to new strategies.

I do this in my game Junkernauts, where players roll dice to deal damage in combat. Each die can deal 0, 1, or 2 damage. However, if the player rolls a 0, they get to "recoil," allowing their bot to move for free. This gives them an opportunity to take advantage of a bad roll by moving and potentially dodging a counterattack.

You can also use output randomness to create more input randomness by making all outcomes good but in different ways. In Catan, buying a development card gives you a random card from the deck. While Knight cards are more likely than any other card type, players do not have strong expectations about the outcome. Whatever they draw gives them direction for what to do next.

We should not fear output randomness. It is a powerful tool for creating moments of fun surprise for players. What we should avoid instead is effort without reward, which is the true killer of fun. Whatever your players choose to do, make sure it accomplishes something, even if it fails.