There are a lot of ways to distribute loot in video games. From the loot drops of MMORPGs, to the Pokemon Card Game's booster packs, to just plain gachapon-based systems, players love to spin the wheel and see the multicolored loot fly out.
But it turns out when balancing a loot system, it's not just enough to say "common, uncommon, rare, epic," seed the loot by power level, and call it a day. The experience of making finely-tuned loot experiences---known in the professional tabletop world as "collating," is a fascinating mix of art, science, and data. At GDC Summer, former Heroclix game designer Eric Engelhard took the time to break down a series of principles that can help developers make better, more fair-feeling experiences for players.
To help break down some of Engelhard's math-laden examples, we've selected a few key lessons you can take away from his talk. (But be sure to watch the whole thing once it hits the GDC Vault).
Build a randomization strategy that lets you communicate EXACTLY how items should be distributed.
The core component of Engelhard's talk was the notion of a "Fixed" method of item collation. In fixed collation, a designer creates a randomization formula that first divides the items into rarity tiers, then calculates how items should be distributed by precisely specifying which cards of which rarities should be in which booster packs, which are then arranged into sets of boosters (referred to here as "bricks"), and then boxed again as "cases."
This creates specific rules for how items should be arranged in what is encapsulated as a "packout." If loot is distributed only along the lines of a simple randomization chart, it creates a system that's not only difficult to fulfill at a physical factory or in programming terms, but also creates long gaps where players won't have the great loot experiences that make buying booster pack-type items so rewarding.
Engelhard's example lets you control, say, what Heroclix figures appear in what boxes, but you can apply this to an MMORPG's loot table too. If a boss just follows a randomized loot formula, there's a good chance the rarest loot, though it's statistically likely to appear, will never drop for some players, even if they defeat said boss at every possible opportunity. (Source, me, having flashbacks to my days raiding Zul'Gurub in the summer of '07).
Ensure your players can always get a rare item of some kind.
Depending on the sliding scale of your game's rarity and item-total, you may wind up in a scenario where your distribution system doesn't guarantee the rarer, high-value items that make your loot system feel good. For instance, if you're using a 4-tier rarity system (common, uncommon, rare, super-rare), you may wind up with a scenario where players have a 64 percent chance of receiving a rare item, and a 32 percent chance of receiving a super-rare item.
Applied to each respective item, those values make sense. But combined, they create a 2 percent chance that a player won't receive either item in a pack. Engelhard demonstrated this with an example from an anonymous online RPG, where he opened 88 booster packs bought in one bulk package. According to the game's publicly released values, higher-value items were supposed to appear in every 1/8 packs, which should average out to 11 of those items.
He got zero. "A lot of players would have quit the game right then and there. That kind of thing is the rough edge of randomness that I try to minimize with what I do."
Consider the cost to collect, and make sure your math doesn't skew it
When distributing loot through randomized systems, there's still a fixed cost of what it will take to collect items at each reward tier. When figuring out the odds of your individual loot items, this can create some wonky math where even though your loot is prioritized appropriately, once the cost of an individual pack is considered, it creates disparities in an item's relative value.
For instance, in one example Engelhard drew up, it was possible for the cost of acquiring common items to be GREATER than acquiring uncommon items. This was because even though the distribution of common to uncommon cards felt 'fair' in an isolated instance, it created assymetrical values and drove up the cost of common cards across a larger item set.
Engelhard's solution in this example was to increase the chance of common items being found, and decreasing the chance for uncommon items
Don't let the costs to collect balloon from rarity tier to rarity tier
Once your loot system's cost-to-collect values are consistent, you may encounter another problem that Engelhard discussed; the cost differences between your different tiers may create an un-fun experience.
For instance, in a loot set, it may cost $50 to acquire all of a set's common items, and $80 to collect all its uncommon items. Then it's $100 to collect all of the rare items. In this case, the delta from common to uncommon is greater than that from uncommon to rare.
Even if your loot system doesn't have a monetary cost, but measures 'cost to collect' in hours played, Engelhard argued it still can create a disjointed feeling for players.
Re-examine your loot system to make sure it has fun flavor
Once you have your loot table spread out, Engelhard argued that it's worth diving back into what makes opening packs so much fun, and not just counting on your players being satisfied with another cookie-cutter randomized loot system. While a strong balance in a packout can help ensure your game's power curve is managed fairly and profitably, he argued that players are still more attached to the surprise factor of opening these packs than the power itself.
In digital games, this can be done by creating specific rules that help players guarantee valuable loot after a certain amount of time. (Think how Apex Legends guarantees its 1 percent-rate 'Heirloom' drops within 500 packs).
In collectible games broadly, you can think of bonuses that add value to your lower-rarity items. This can happen on the game's design side, where specific lower-value items have synergy with higher-value items, or on the flavor side, with alternate art schemes or unique aesthetic characteristics that make unspecial items feel more special.
One of Engelhard's pleas to developers watching his GDC talk was that they not treat item collation as an afterthought. Engelhard acknowledged that much of the reason that loot distribution systems can feel bad is that designers often don't have the knowledge on best practices in the world of collation, since the random nature of the business means collation formulas have to be kept a tightly-guarded secret.
By sharing the tips above (and others that you can see in his full talk), Engelhard hopes that developers can finally outgrow the short-term memory syndrome this practice creates, and establish long-lasting loot principles that can be rewarding for players and developers alike.