A hit game refunded 55,000 times shows who a rule really protects

Gaming 4 min 80 sources

Steam's two-hour refund window was built to protect buyers. For a short, well-liked indie game, it became a way to play free — and reopened a fight over who a storefront's defaults are for.

Key takeaways

A short, well-liked indie game was refunded 55,000 times because Steam's two-hour refund rule — built to protect buyers of big games — lets players finish a short game and get their money back.
A refund policy is a default that decides who carries the risk before any sale happens; change it and the advantage moves from buyers to makers, which is why both sides are fighting over it.
Single-player puzzle game Blue Prince is being turned into a $20,000 esport — proof that competition can be a layer added on top of a finished game, not something the game was built with.

The best-reviewed game a solo developer has ever made was returned 55,000 times — by people who liked it. The reason isn’t a bug or a scam. It’s a rule working exactly as written, for a game it was never designed with in mind.

A refund policy, used against the maker

This week, a 22-year-old solo developer who goes by Zoroarts said his game Paddle Paddle Paddle — a short co-op boat-paddling adventure — had been refunded 55,000 times [1][2]. The game sold about 262,000 copies in its first year and carries a “Very Positive” rating on Steam, with roughly 1,400 reviews [1][2][4]. It costs $4.99, often on sale for around $3 [2][4]. The refund rate: 21 percent, against 90 percent positive reviews [1][4].

The cause is Steam’s refund policy. Valve, the company that runs Steam, lets a player return any game for a full refund if they’ve owned it under 14 days and played under two hours, no reason required [1][2][4]. For a $60 game you hate, that’s a genuine protection — you’re not stuck with a purchase you regret. But Paddle Paddle Paddle can be finished in under two hours. The developer says he planned about four hours of content, yet speedrunners and skilled players cleared it in one to two [4]. So players could complete the whole game, leave a glowing review, and get their money back — all inside the window meant to shield them from bad purchases [1][4].

Here’s the mechanism worth holding on to. A refund policy is a default — a rule that decides who’s protected before any single transaction happens. Steam’s default is built for the standard case: a big game, a buyer who might regret it. For that case it works. But the same rule, unchanged, transfers the risk of a short game entirely onto its maker. One of Zoroarts’ own reviews reads: “GREAT GAME, finished within 1:40 hrs (refunded)” [4]. The rule didn’t fail. It did precisely what it says — and short games were simply never the buyer it was written to protect against.

Valve’s terms do include a section on refund “abuse,” stating that refunds are not “a way to get free games,” and that anyone taking unfair advantage could lose refund privileges [4]. How much that’s enforced in practice is unclear [4]. Zoroarts isn’t asking for the policy to be scrapped — he wants games to show expected playtime and price up front, so “too short” can’t be a hidden refund reason [4]. Developer reaction has been broadly sympathetic; many argue Valve needs a policy that accounts for short games [2]. Some players, who benefit most from the current rule, disagree — the top-voted review on the game calls out the developer for complaining at all [2]. That split is the whole story: change a default, and you move the advantage from one side to the other.

A puzzle game with no opponent becomes a $20,000 sport

Blue Prince — a single-player puzzle-roguelike from small studio Dogubomb, and one of 2025’s most praised games — is being turned into a competition [3]. In July, Dogubomb is running a tournament with a $20,000 prize pool, built on “bingo”-style rules: players race to reach the game’s hidden final room while ticking off side objectives [3]. Two open qualifiers run on July 12 and 19; the top four from those join eight invited players [3].

The interesting part is the mechanism. Blue Prince has no built-in multiplayer, no leaderboard, no opponent — it’s a solitary mystery you unravel over 100-plus hours [3]. Competition here isn’t a feature the studio shipped; it’s a layer the community and studio added on top of a finished game. The speedrunning scene got there first — Blue Prince raced at a Summer Games Done Quick event in 2025 [3] — and the studio is now formalising it with prize money. A game becomes a sport not because it was designed as one, but because enough people agreed to race the same way. The rules of the contest are invented after the fact, layered over a game that never needed them.

The quieter cost of a shrinking MMO

For a less visible story: The Elder Scrolls Online, the long-running online game from ZeniMax Online Studios, has reportedly lost “as much as half” of its development team, according to a report cited by IGN, as its roadmap is re-evaluated [44]. The exact number of people affected is unknown [44]. It came as part of Microsoft’s wider Xbox layoffs, with parent studio Bethesda now expected to concentrate on a small set of franchises [44].

A live online game is a promise that costs money every day it keeps running — servers, updates, a team large enough to make new content. Cut the team in half and the promise doesn’t break loudly; it thins. Fewer updates, a slower roadmap, and eventually players notice the world they live in has stopped growing. The people who lost jobs are the immediate cost. The slow erosion of a game thousands still play is the one that lands later.

02 · Lesson · why it matters

The rule that protects you was written against someone

Every default quietly decides who carries the risk — and it looks like fairness right up until it lands on the person it was never written for.

A rule doing exactly what it says

Fifty-five thousand people bought a boat-paddling game, finished it, left kind reviews, and took their money back. No one broke a rule. That’s the unsettling part. Steam’s refund policy says you can return any game played under two hours, no questions asked, and the game could be finished in about that time. The players did nothing but use the rule as written. And the developer lost roughly a fifth of his sales to it.

It’s tempting to call this a loophole, as if someone found a crack. But there’s no crack. The rule worked perfectly. The trouble is that a rule which works perfectly can still be aimed — without anyone aiming it — at a person it was never meant to touch.

Every default takes a side before you arrive

A refund policy sounds like plain fairness. It isn’t. It’s a choice about who carries risk. Before any single purchase happens, the rule has already decided: if a game disappoints, the buyer walks away whole and the seller eats the loss. That’s a defensible choice. A $60 game you hate is a real harm, and protecting the buyer from it is reasonable.

But notice what the choice assumes. It assumes the game is long. It assumes two hours is a taste, not the whole meal. Build the rule around that picture and it protects buyers from bad purchases. Point the same rule at a game that fits inside two hours, and it stops protecting anyone from a bad purchase — it just lets a good purchase be un-bought after it’s fully enjoyed. Same words. Opposite effect. The rule didn’t change. The person standing under it did.

This is what a default is: a decision about who is protected, made before the specific case shows up, and then treated as if it were simply the way things are. “You can refund under two hours” doesn’t feel like a decision. It feels like a fact about the world. It is not. Someone at Valve chose it, for a reason, imagining a certain kind of buyer and a certain kind of game.

It poses as neutral, and that’s the trick

The reason a default is powerful is that it hides. A price, a border, a deadline, a return window — each looks like the natural state of things until you ask who wrote it and who it was written for. Ask that, and the neutrality dissolves. The refund window isn’t neutral between the player who wants safety and the maker of a short game. It picks the player. It was always going to pick the player, because that’s the case it was built to handle.

You can see this the moment you try to change it. The developer isn’t asking to abolish refunds. He’s asking for something smaller — show the expected playtime and price up front, so “too short” can’t be a quiet reason to return a game already finished. But even that modest shift moves the advantage. Players who like the current rule pushed back; the top review on his own game scolds him for complaining. Of course they did. When a default protects you, its fairness is invisible. When it’s changed to protect someone else, the cost suddenly feels sharp. The rule was never neutral. It only looked neutral to whoever it happened to be shielding.

You are already standing under a hundred of them

Here is the part that includes you. You have almost certainly used a return policy to walk away from something you regretted — and been grateful the rule had your back. On that day, the default was on your side, and it felt like justice, not like a choice made at your seller’s expense. It was both. Every rule that protects you is, in the same motion, a risk handed to someone else. That’s not a scandal. It’s just how a default works: it can’t protect one party without exposing another.

The web here is wider than one game. The pattern runs through every arrangement you live inside. The late-fee grace period that helps you also lets a system run on your unpaid balance. The insurance rule that pays out for your case denies someone whose situation the drafters didn’t picture. The default in your bank, your lease, your job — each was written around a typical person, and each quietly transfers risk to whoever falls outside that picture. Usually that person is far from the room where the rule was made. Sometimes it’s you.

What seeing this is for

You don’t get to redesign Steam’s refund policy, or most of the defaults you live under. Seeing the shape doesn’t hand you a lever. What it does is stop you mistaking a rule for a fact. When something in your world looks like the plain, neutral way things are — a fee, a window, a default setting, a “standard” term — the useful question isn’t “is this fair?” It’s “who was this written for, and who is standing under it now?” Those can be different people. When they are, the rule will still do exactly what it says, and someone will still pay for it who never agreed to.

Hold that loosely. It doesn’t make you clever, and it won’t tell you which defaults are right. Plenty of them protect real people from real harm, including you. It only means that when a rule finally lands on you the wrong way — and one will — you’ll know it isn’t the world being unfair. It’s a choice someone made, for a case that wasn’t yours, working exactly as designed.

03 · Lab · your turn

Set the Refund Line

Rehearse setting one refund default and feel how any single line protects some players while exposing others — the same rule, opposite effects depending on who stands under it.

04 · Hope · carry this

A rule only stays unfair as long as no one who's hurt by it can be heard, and here a lone developer with a small game got the whole industry talking about a policy most people never questioned. Defaults get better exactly this way — slowly, out loud, once someone standing under one decides to point at it.