When language locks a genre
Our genre scan scores five things: how validated a mechanic is at home, how crowded it is where we would ship, how hard it is for one person to build, how much regulatory trouble it carries, and how dependent its appeal is on a specific language. That last number is the one I underestimated for a year, and it is the one that explains why some genres I adored were never really available to us.
Two questions that are not the same
The easy confusion is between "could we build it" and "could we reach an audience with it." The first is a difficulty score, the kind we write after estimating how much content and netcode a genre demands. The second is something else. A genre can be trivial to build and completely closed to an English only studio, because the fun was never in the building. It was in the references. The jokes. The idioms. The IP that everyone in one market knows and no one in another has met.
We give that second question its own column so it cannot hide inside the others. If it rode along inside difficulty, I would keep reading a high difficulty score as the only reason to skip a genre, and I would miss the genres that are easy to build and impossible to sell. Those are the more dangerous ones, because they look free until you try to localize the humor and find there is no humor left after translation. None at all.
The score that catches the trap
In our current snapshot of fifteen categories, the language dependency score runs from 1 to 9. A 1 means the mechanic travels with no change. A 9 means the appeal is welded to a specific culture. The spread matters more than any single value, because the high end is where the trap lives, and the high end is not where I used to look. I used to scan for the genres with the best home validation and the lowest difficulty, which is exactly the set most likely to be locked by language, and I called that set opportunity. It was usually a wall wearing an opportunity's clothes.
The cruelest combination in the whole table is a genre with a high home validation score and a high language score at once. The genre is proven. The audience demonstrably exists. The references simply do not cross a border. You can see the players on the other side. You cannot hand them the thing they came for, because the thing they came for is a set of shared jokes, and you do not share them. No amount of build quality repairs that. I have watched myself want to argue with this, and the argument is always the same and always wrong: surely the core loop works without the jokes. Sometimes it does. Often it does not, because the jokes were the loop's personality.
A genre that dies at the border
One category in the scan sits at language dependency 9. It is an idle game built on Three Kingdoms and Journey to the West memes, the kind of thing that dominates its home market, and the scan note on it says plainly that those references die the moment they leave that market. The home validation score on that row is 8, which is high, and the difficulty is 6, which is middling, so on the two numbers I used to care about it looks like a strong, buildable, proven genre. On language it is a 9, which means for a studio that ships in English it is effectively closed. Not because we lack the skill. Because the product is the references, and we do not own them and cannot fake them.
I want to be careful here. This is not a claim about quality. The genre is brilliant at home. The point is narrower and it is the only point: a reference locked genre is not reachable by a team that does not share the references, and reachability is a real constraint, as real as difficulty, even though it shows up in a different column. Treating it as a soft factor is how you spend three months building something the target audience experiences as a translation.
The gentler version of the same wall
A softer case sits at language dependency 4. A garden themed cozy genre scores well at home, but the scan note records that a specific national art style already failed in North America, a validated failure rather than a guess. The score is lower, so the wall is lower, and the same note proposes a fix: change the visual language to a globally readable cozy style and the genre opens up. That is the useful kind of high language score, the kind with a known workaround, because a 4 is a styling problem and a 9 is an identity problem, and those are not the same distance to travel.
Another row at language 4 is an office themed satire built on workplace recognition. The scan notes the emotional resonance crosses cultures and the supply is thin, which sounds like an opening, but it also flags that the home validation score is only 5, meaning the genre was never proven at home either. So you have a cross cultural idea with no validation behind it, which is a different risk entirely, the risk of guessing at demand in two markets at once. I keep these two 4s separate in my head, because one is a localization fix and the other is a validation gap wearing a localization fix's costume.
What the language score is not
It is not a verdict on a market. A high language score on a genre says nothing about whether that genre is good, only about whether its appeal travels. A genre can be culturally specific and enormously successful, and the score is just a flag to a team like ours that success is not portable. I have to resist the urge to read a 9 as "niche" or a 1 as "universal," because both readings sneak a value judgement into a reachability number, and the number was meant to stop me from doing exactly that.
The practice I have now is to read the language column before the difficulty column, which is the reverse of my old habit. If a genre is locked by language, the difficulty score is irrelevant to whether we should touch it, and reading difficulty first is how I used to talk myself into prototype after prototype for things I could build and could not sell. The order is the whole correction. A wall you check first is a wall you stop running at. A wall you check last is a wall you decorate.
We do not ship reference locked genres, and not because we decided they are bad. Because the references are not ours to ship, and a translated personality is not a personality. The language score is how we admit that early, on the same sheet as the build estimate, so the honest answer shows up before the prototype does, which is the only time an honest answer can still save the weeks. It is a small discipline, written into a single spreadsheet column, but it has saved us more wasted months than any of the build estimates ever managed, because a wall you name on page one is a wall you never quietly build a whole game against.