Four friends kill a boss. One sword drops. Everybody wants it.
That sentence has ended more co-op sessions than any difficulty spike ever managed, and every multiplayer game ships an answer to it whether the designers thought carefully about the problem or not.
It is worth taking seriously, because fair division is a genuinely hard problem with a century of mathematics behind it, and games keep rediscovering the same failure modes.
The simplest implementation puts loot on the ground and lets whoever reaches it first take it.
This is trivially easy to build and it converts a cooperative game into a footrace at exactly the moment the cooperation paid off. It also rewards the wrong thing: the player with the fastest character or the best connection, rather than the one who did the work.
The maths makes the frustration concrete. In a four-player party with genuinely random pickups, the chance that one specific player takes all three drops from a run is 1 in 64. Across five drops it is 1 in 1,024. Rare enough to feel like an outrage, common enough that it happens to somebody in your group most weeks.
Streaks like that are what independent randomness does, and they are precisely why almost every considered design moves away from it.
The obvious correction is to rotate. Player one takes the first drop, player two the second, and so on.
Across eight drops in a four-player party, everybody gets exactly two. Perfectly equal, completely predictable, and no arguments about who got lucky.
It also hands the plate armour to the mage and the wand to the barbarian, because the rotation has no idea what anyone can use. You have solved distributional fairness and created an allocation problem, which is the same trade every fair-division system makes.
Most groups then invent an informal trading layer on top, which works until somebody forgets who owes what.
The MMO answer is to let players declare intent. Anyone who can use the item rolls for it, everyone else defers.
Better on fit, and it introduces a new complaint. With four contenders on an item, your chance of winning a given roll is 25%, which means your chance of winning nothing across four separate items is 31.6%. Nearly a third of the time, a player sits through four contested drops and receives none of them.
That is the experience people describe as the loot system hating them. It is independence doing exactly what independence does, and it feels like malice because humans expect outcomes to even out over short samples.
Which is the same reason so many systems that distribute value under uncertainty end up adding guarantees rather than trusting the arithmetic. Anyone comparing how shared-stakes platforms handle payout timing and settlement will find the same tension documented in detail (source: https://www.silvergames.com/en/instant-withdrawal-casino), where the differentiator is not the odds but how quickly and reliably a result converts into something you actually hold.
The parallel is closer than it sounds. A raid group and a casino cashier are both answering the same question: the outcome is decided, so how long until the person who won it has it, and can they rely on that. Games that answer it badly produce the same complaint the finance side does, which is that the waiting is worse than the losing.
The solution that stuck is the one that refuses the premise.
Instanced drops give every player their own roll on their own loot. Nobody competes, nobody trades, nobody negotiates. What you see on the ground is yours and invisible to everyone else.
Diablo III moved to it after the auction house demonstrated what happens when shared loot meets an open market. Destiny built it in from the start. Borderlands offers it as an option and the option is what most groups use.
The cost is real and worth naming. Instanced loot removes an entire social layer. The trading, the negotiating, the group deciding collectively that the tank needs the shield more than the DPS wants the stats. That was genuine interaction, and the games that removed it removed something people valued alongside the thing they hated.
Which is the honest summary of the whole problem: every solution trades a social mechanic for a reduction in conflict, and no design has found a way to keep both. More on how co-op design choices land in practice across the archive.
Fair division has been studied formally since the 1940s, and the concepts translate directly.
Proportionality means everyone receives at least their fair share by their own valuation. Round-robin achieves it. Free-for-all does not.
Envy-freeness means nobody prefers somebody else's allocation to their own. This is much harder, and it is what instanced loot achieves by making comparison impossible rather than by making allocations equal.
Both sit inside a much older argument about distributive justice, which established long ago that you cannot generally satisfy every desirable property at once. Game designers rediscovered that empirically, patch by patch, usually after a community revolt.
Which is the useful thing to know when a loot system annoys you. It is not that the developers failed to think about it. It is that the problem has no clean answer, and they picked which complaint they preferred to receive.
Big Walk sold more than one million copies in its first six days after launching on 4 August 2026, and it did so while PlayStation Plus members could play it for free that same month, according to Game Developer. Nobody was buying a file. They were paying to spend an evening getting lost in the bush with friends.
The best shared setups are built around problems you can actually solve: clearer screens, reliable controllers and enough performance to keep the game running. Spending more does not guarantee a better night, so the useful question is simple: what is getting in the way of everyone having a good time?
Multiplayer games are no longer unusual. Friends form teams, complete missions together and spend entire evenings working toward the same objective. In the U.S., the Entertainment Software A