What an hour of an indie game costs, and what happens when it costs too much

The question a developer asks before setting a price is whether the game is worth it, and the only honest way to answer is against what comparable games deliver for the same money. Median hours rise with every price band, from 3.3 in Under $5 to 26 in $40 and up. That much is unsurprising, and on its own it is not useful: knowing that expensive games are longer does not tell you what your own price is promising.

The useful version is the ratio. Across the whole pool the median game records 8.2 hours at a median price that works out to 1.73 dollars an hour. Those hours are what a reviewer had played when they wrote their review, so they run high against what an average buyer would put in, and they are comparable between games because that bias applies to every game equally.

Below 10.00 dollars, cheap games are genuinely cheap per hour: 1.12 in Under $5. Above it the rate stops responding to price. Across the 5 bands from 10.00 dollars up, the median cost per hour stays between 1.90 and 2.06, a spread of 0.16 while the price itself quadruples.

The middle half of each band moves just as little, which is what makes this a norm rather than a coincidence of medians. Whatever a developer thinks they are deciding when they set a price, in aggregate they are deciding how many hours to owe. Games in $40 and up record 26 median hours against 7.2 for games in $10 to $14.99, and the two bands charge almost the same for each of those hours.

This is the number to compare yourself against, and it is not an argument for pricing high. It says that a price and a length are the same decision made twice, so choosing them independently is how a game ends up outside the norm without anyone having meant to.

Sorting games by what an hour costs instead of by what they charge produces a clean gradient in review scores. The median positive share falls from 89.5 in Under $0.50 to 67.1 in $10.00 and up, without a single reversal along the way.

The tail moves much harder than the median, and the tail is what a developer actually experiences. The share of games sitting under 70 percent positive goes from 5 percent in the cheapest group to 55 percent in the most expensive, on 107 games. In that last group, $10.00 and up an hour, being poorly reviewed stops being the unlucky outcome and becomes the common one.

Read this as a symptom and not as a lever. A game that people put down after an hour records few hours and collects bad reviews, and both come from the same place. What the ratio offers is an early, measurable sign of that happening, available long before the review score has settled.

Compared band against band, expensive games are the worst reviewed: median positive share falls from 87.3 in Under $5 to 83 in $40 and up. Taken at face value that reads as an argument for pricing low, and it is wrong.

Inside every price band the relationship runs the other way. Splitting each band by hours, the longest quarter is better reviewed than the shortest in all of them: 86.1 against 89 in Under $10, and 78 against 86.8 in $30 and up. The aggregate reverses the direction because expensive bands hold more games that fall short of what they asked for, not because charging more is punished.

The width of that gap is the finding. It grows from 2.9 points to 8.8 across the bands, which is the same statement as the rate above, seen from the buyer's side. Short and cheap is forgiven. Short and expensive is not, and the tolerance narrows with every dollar.

The pool-wide rate is an average of habits that differ sharply. RPG delivers hours at 1.20 dollars each, Horror at 2.16, and the spread between them is wider than the entire plateau. A game at 2.16 dollars an hour is unremarkable in one genre and an outlier in the other.

That the expensive-per-hour genres are not the badly reviewed ones is the clearest evidence that the rate is a norm rather than a law. Adventure charges 1.86 dollars an hour, squarely in the middle of this table rather than at its cheap end, and is the best reviewed genre here at 91.1 percent positive. Its players are not buying hours, and pricing it against a genre whose players are would be the wrong comparison.

The practical version: the number worth measuring yourself against is your genre's, not the pool's, and the further your genre sits from the middle the more misleading the pool figure becomes.

The hours here are sampled from each game's most recent reviews, which means an older game is described by whoever bought it last, often during a sale. That could manufacture the entire gradient: late buyers of an old game pay less, so the cost per hour computed at full price would be wrong in a way that correlates with age.

Splitting the pool into three release cohorts and re-measuring inside each shows the gradient holding in all of them, over 6,954 games. The most recent cohort, where the gap between the sampled reviewers and the launch audience is smallest, shows the steepest fall of the three.

That does not eliminate the bias, and no cohort split can. It does mean the finding is not an artefact of measuring old games through their bargain buyers.

It cannot tell you that adding hours will improve your reviews. The correlation is real and the causal story behind it is almost certainly the reverse one, where a game that fails to hold people generates both the short sessions and the bad scores. Padding a game to move a ratio would produce exactly the experience the ratio is detecting.

It cannot tell you what any of this earned, either. There are no sales figures in this pool, and a game with a low score at a high price may still have sold well.

What it does support is narrower and worth having. There is a going rate for an hour, it is stable above 10.00 dollars, it differs by genre, and the tolerance for falling below it shrinks as your price rises. If you are pricing above what your length supports, that is measurable before launch, against comparables, rather than after launch, in your review score.