A funnel design is a game in which the players’ varied activities feed into a single, central resolution mechanism. Funnel designs have two distinct components: the things the players do, and the resolution mechanism that aggregates their choices to determine the impact they have on the game. That can be an extremely valuable and even necessary function, but just as a bad cooking funnel can squeeze off the ingredients needed for a recipe, a bad game design funnel can choke the fun out of a game. It’s important, therefore, to make sure that funnels in games are well-implemented, with due thought given to issues of balance and with care taken to avoid unduly limiting the design.

The odds are you’ve played a funnel design. Most wargamers will be familiar with combat resolution tables, which boil relative strengths down to dice results: at 3-1 odds a 4-6 on a single die means the defending units are eliminated, while at 2-1 odds only a 5-6 eliminates the defenders, etc. Players maneuver their units, decide which supply centers to call on for logistical support, weigh the odds of the weather improving in a few days, and make many other decisions all to get an advantage on the game’s combat resolution table. Then the table produces its result, and play continues.

If you’re not a fan of wargames, you might have played Sentinels of the Multiverse, a card-driven game about superheroes that uses funnel design. Each player stands in the shoes of a particular hero, playing a deck full of unique cards representing that hero’s powers. However, most of these cards feed into a basic mechanism in which the players inflict damage on an arch-villain by making, and then modifying, attacks. Thus, a player might play a card that causes her character to punch the villain, and then modify that attack with other cards that inspire the character to greater effort and encase the character’s fist in a ball of fire; the game’s underlying math recognizes the punch as a 3 damage attack, modified by +1 for the extra effort and +1 for the fire, for a total of 5 damage. The same math engine does the same translation into numbers for gunfire, artificially induced hailstorms, and every other form of superpowered aggression.

By way of comparison, consider a game that doesn’t use funnel design—say, the popular deck-building game Dominion. The goal in Dominion is to acquire victory cards, and one normally does that by getting treasure with which to buy them. However, treasure is not an intermediary system evaluating the players’ performance; it’s a game element that players can directly manipulate. They can buy it, find it, steal it, trade it in for better treasures, and in some cases ignore it entirely. Playing well does not directly and always mean more treasure. While treasure is important, then, players’ actions are not funneled through it while playing Dominion the way they are in the examples above.

Similarly, having a score at the end does not make a game a funnel design. It’s true that a score serves as an overall measure of player performance. However, a score does not operate on a player’s choices, mediating between the player and the game, in the way that a combat resolution table or Sentinels’ underlying math does. Scoring systems can impact the designs of their games, but they don’t pose the same issues as a funnel operating during the game. To put it another way, funnels are used while cooking, while final scoring is akin to the bowl in which the meal is served. Both are important, but they’re not the same and shouldn’t be conflated.

True funnel designs are interesting because they have two very different, yet nevertheless interrelated, parts: the players’ actions, and the funnel mechanism that interprets those actions. Often the two parts do not look, or work, anything like each other; a wargame’s tactical map-based play is replaced with die-rolling when the combat resolution table comes to the fore. Neither, however, can be understood in a vacuum. The players’ actions are all taken in light of the funnel, and the funnel is unimportant without their actions.

Having two vitally interconnected parts to a game can have substantial advantages. First, it creates what Magic: the Gathering’s developers sometimes refer to as “knobs:” values that can be changed to achieve game balance. If one of the choices available to players is too strong or too weak, the designer can either manipulate the choice directly, or alter how the funnel treats the choice. That alternative means of accessing the game’s inner workings can be very useful.

Second, a funnel can summarize very different and complex interactions, allowing a variety of pieces to interact in a consistent way. Actually simulating a battle between tanks and infantry involves measuring their very different strengths: the tanks are sturdier, but the infantry can more easily take advantage of terrain. The tanks are stable firing platforms, but the infantry might be better at getting advantageous angles. Well-designed funnels translate those disparate capabilities into a single system, making it easier for players to evaluate what’s happening and speeding resolution of dense, multifaceted situations.

These advantages do not mean, however, that funnels are invariably desirable or good. Since players must interact with them constantly, their flaws are magnified. A boring funnel can color the entire experience; an unbalanced funnel unbalances everything; an overly complex funnel slows every single turn and renders the import of player choices opaque. Errors in a funnel’s design have multiplicative effects, which makes funnels something to be implemented only when necessary and only with caution.

Moreover, even correctly designed funnels can have unintended and undesirable effects on a game’s design. Some things will be easier to fit into the funnel than others. Inevitably, this encourages the design to move in the direction of including more of what the funnel readily handles, and less of what it does not.

A quick thought experiment demonstrates the impact of a bad funnel. Suppose someone designed a wargame whose combat resolution table said, for every entry, “the defenders destroy all attacking units.” That game would be unplayable; since taking any sort of initiative would be punished by swift destruction, no one would ever want to do anything!

That is, of course, an extreme example. A subtler question might focus on the impact of the die roll that is normally involved in getting a combat resolution table’s final result. Some players strongly disfavor random elements in games, feeling that they privilege luck over skill. Others feel that they simply have “bad dice luck.” (In my experience, this is a majority of people.) Is the die roll important enough to risk losing the former group as potential customers, and to risk frustrating the latter group during play?

There is no single, constant answer to that question, which reminds us that there can be no single, constant answer to the question of whether funnel design is a good idea. It is always a balancing test, a matter of deciding whether the funnel’s advantages outweigh its dangers. Just remember that, like many tools, funnels can be harmful if employed carelessly.