Difficulty ratings on arcade games mean something different from difficulty settings in longer games, and the distinction is worth understanding before you pick something labelled Expert and conclude you are bad at games.
What the ratings describe
In an arcade context a rating describes how demanding a game is, not how much content it contains or how long it takes to finish.
Easy does not mean shallow. A one-input game can be rated Easy and still take a very long time to master, because the rating is about the barrier to a good first experience rather than about the ceiling.
Expert is the one that misleads people. It is not a reward for finishing the easier games; it is a different kind of experience, and going in expecting to win immediately produces frustration the design did not intend.
| Rating | What it means |
|---|---|
| Easy | A beginner can have a satisfying run within a few attempts |
| Medium | The game expects one specific skill; you will be poor at it until you build it |
| Hard | It expects several skills at once, and weakness in any one ends runs |
| Expert | Losing repeatedly while learning a fixed pattern is part of the design |
Curves beat selectors
In an arcade game, difficulty should rise inside the run rather than being chosen at the start.
A selector hands the tuning problem to the player, who has less information than the designer. It also fragments the experience: your score on Easy is not comparable to anyone's score on Hard, so the number stops meaning anything.
A curve solves both. Everyone plays the same game, the difficulty finds its own level as the run progresses, and a score is a score. That is why a difficulty menu in an arcade game is usually a sign the curve does not work.
Ad space reserved. See README.md to connect an ad unit.
The shape of a good curve
Two properties matter and both are observable within a couple of runs.
It should start shallow. The first twenty seconds exist to teach, so anything that kills a competent player that early is badly tuned.
It should flatten rather than rising forever. A curve that climbs without limit guarantees every run ends because the game exceeded human capability, which makes the ending about arithmetic rather than about you. A curve that approaches a ceiling means a long run ends because of a lapse, which is the only ending worth trying to avoid.
Stepped curves, where difficulty rises in visible tiers, add a third benefit: you know when the game changed, so you can attribute a failure correctly rather than wondering whether you got worse.
Hard versus unclear
These feel similar from inside a run and are completely different problems, and telling them apart is the most useful diagnostic a player has.
A hard game is one where you know what killed you and could not do it. An unclear game is one where you do not know what killed you. The first is a skill problem you can work on; the second is a design problem you cannot.
The test is simply to ask, after each death, what you should have done differently. If you can answer most of the time, the game is hard and worth practising. If you regularly cannot, it is not signalling properly, and no amount of practice will fix that.
Why ratings differ between sites
There is no standard, so a Medium on one site may be a Hard on another. Ratings are only comparable within a library.
What makes them useful is consistency of the person assigning them. A site where one person has played everything and rated it against the rest produces internally consistent ratings; a site aggregating games from elsewhere is usually repeating whatever the source claimed.
The practical advice is to calibrate against one game. Play something rated Medium on a site, and you then know what that site's Hard is likely to feel like.
What makes a difficulty rating honest
Ratings are only useful if they were assigned by someone who played the game, and there are a few signals that distinguish the two cases.
Variation is the first. A library where ratings spread across the range has somebody making judgements; one where everything is Medium has a default value in a template.
Correlation with the controls is the second. A game listing six inputs and rated Easy is suspicious, because input count and difficulty are not independent. Not impossible - a forgiving game can have many controls - but worth a second look.
And the presence of an Expert rating at all is the third. Sites reluctant to label anything as genuinely hard usually have nothing genuinely hard, and a library with no difficult game in it has a ceiling you will reach quickly.
The honest version of a rating also acknowledges what it does not cover. Difficulty here describes how demanding a game is, not how long it takes, how deep it is, or how much practice it rewards - and those are different questions that a single label cannot answer.
Why one number cannot describe difficulty
A rating compresses several independent things, and knowing which ones helps you interpret it.
Execution demand: how precisely you must act. Knowledge demand: how much you must have learned. Attention demand: how many things you must track. Consistency demand: how long you must sustain it.
A boss fight is high on knowledge and low on attention. A survival arena is the reverse. Both might be rated Hard, and they will feel completely different - and a player strong in one may be genuinely poor at the other without either rating being wrong.
This is why a rating should be read alongside the category rather than on its own, and why 'I found the Hard game easier than the Medium one' is a common and entirely coherent experience rather than a sign that something is mis-rated.
Key takeaways
- Ratings describe how demanding a game is, not how long or how deep it is.
- Expert means repeated learning losses are part of the design, not that you chose badly.
- A difficulty curve beats a selector: everyone plays the same game and scores stay comparable.
- Good curves start shallow and flatten, so runs end because of a lapse rather than arithmetic.
- Hard means you know what killed you; unclear means you do not. Only the first is worth practising.
- Ratings are only comparable within one library - calibrate by playing something rated Medium.
- Ratings that never vary were not assigned by anyone who played the games.
- Input count and difficulty correlate, so a six-control game rated Easy is worth a second look.
- A library with no genuinely hard game has a ceiling you will reach quickly.
- A rating compresses execution, knowledge, attention and consistency demands into one number.
- Two games rated the same can feel completely different, which is why ratings should be read alongside the category.
Frequently asked questions
What does Expert difficulty mean?
That losing repeatedly while you learn a fixed pattern is part of the intended experience. It is a different kind of game, not a reward for finishing easier ones.
Does Easy mean shallow?
No. The rating describes the barrier to a good first run, not the skill ceiling. One-input games can be rated Easy and take a long time to master.
Why do arcade games not have difficulty settings?
Because the curve should do that job inside the run. A selector fragments scores and usually indicates the curve itself does not work.
How do I tell if a game is hard or badly made?
Ask what you should have done differently after each death. If you can usually answer, it is hard. If you usually cannot, it is not signalling properly.
Are difficulty ratings comparable between sites?
No. There is no standard, so calibrate by playing one Medium game on a site to learn what that site's ratings mean.
How can I tell if difficulty ratings are real?
Check whether they vary across the library and whether they correlate with control counts. Uniform ratings are template defaults rather than judgements.
Why was a Hard game easier for me than a Medium one?
Because a rating compresses four different demands. A game high in knowledge demand and low in attention demand suits a different player from the reverse.