Home › Guides › Review data

What 100,000 Amazon Ratings Actually Tell You About PS5 and Switch Games

By Jamie Chen · 2026-10-11 · 8 min read

Most people buying a console game look at the star rating and stop. That is the weakest signal on the page. A 4.8 from forty ratings and a 4.8 from thirty-seven thousand are not the same claim, and Amazon displays them identically. The number that separates a game that holds up from one that disappointed people is hiding in the parentheses.

The short version: across 57 console products carrying 426,735 ratings between them, the average score is 4.67 and the median listing has 2,849 ratings. Rating volume is wildly uneven, which is the whole point: 13 of these products have under a thousand ratings while 15 have more than ten thousand. The score tells you little. The count tells you how much the score is worth.

Rating volume is not a quality signal on its own

It is easy to treat a high rating count as proof of quality. It is not, and the data here demonstrates why. The most-rated item in this collection is a streaming microphone with 57,120 ratings. The least-rated is a Switch title with 505. Both are legitimate products; the difference in rating count has far more to do with how long a product has been on sale, how broadly it was marketed, and how many people buy that category of thing at all than with how good either one is.

What a large sample does give you is reliability. A rating that has been tested 57,120 times is close to a population mean. A rating tested 505 times is closer to a handful of early opinions. The number tells you how much confidence to place in the stars, not whether the product is good.

What the data actually shows: no drift

There is a piece of received wisdom in review analysis that early ratings skew positive — that launch buyers are self-selected enthusiasts and the average falls as ordinary buyers arrive. It sounds right. It is not visible in this data.

Splitting these 57 products into a thin-sample group (under 3,000 ratings, n=29) and a deep-sample group (3,000 or more, n=28), the average scores are essentially identical: 4.66 against 4.67. A gap of 0.01 stars across a 426,735-rating sample is nothing.

So the honest conclusion is narrower than the folk wisdom: on this data, a deep sample does not make a product look worse. It makes the score trustworthy. If you were hoping the numbers would reveal a hidden quality gap between well-reviewed and thinly-reviewed products, they do not. They reveal that you can rely on one number and not the other.

The PlayStation picture

PS5 hardware carries real rating depth, and console bundles dominate because that is how most people now buy the machine:

Notice the shape of that table. The consoles sit at 4.77 on thousands of ratings, which is strong. Individual PS5 games carry far thinner histories. That is not a verdict on the games; it is an artefact of how they are sold. PS5 titles are frequently bundled, heavily discounted, and often rated against the console listing rather than their own. A game can be excellent and still have a thin paper trail.

Where the Switch catalogue is different

Nintendo's catalogue behaves the opposite way. Its titles accumulate enormous rating counts because the console has been on sale for years and first-party games never leave the catalogue:

The Switch shelf has a median of 3,034 ratings, against 1,296 for PlayStation. Minecraft alone carries 37,352 ratings at 4.8 stars. When a product has been rated that many times, you are reading a population average rather than a launch-week impression.

Peripherals: high volume, low scrutiny

Gear is where rating depth matters most and where buyers look least carefully. A headset is bought once and judged over months; the aggregate rating is genuinely informative in a way the star score of a game — which people often rate within days of purchase — is not.

Streaming equipment has the deepest history of any shelf here, with a median of 7,420 ratings. That makes sense: microphones and capture cards are tools people research, and they are bought by a market that keeps buying them for years.

Does price track rating depth?

The intuitive answer is yes: expensive products attract more research before purchase, and more research should mean more ratings. The data here does not support that either.

These 57 products run from $8.98 to $849.99, a range wide enough that any correlation should be obvious. It is not. Several of the cheapest items carry six-figure rating counts because they are commodities people replace regularly, while some of the most expensive carry only a few thousand because the pool of buyers is inherently small.

What rating depth actually tracks is how many people buy this kind of thing at all. A microphone is a commodity with a large, continuous market. A console bundle is a large one-off purchase that most buyers make once. That difference — not quality, and not price — is what produces a 57,120-rating listing next to a 505-rating one.

So when you compare two products, compare them against their own category's baseline first. A capture card with 10,000 ratings is unremarkable in a shelf where the median is 7,420. A chair with 10,000 ratings is exceptional against a shelf median of 1,282. The absolute number tells you less than the number relative to its peers.

How to use this when you buy

1. Ignore any rating under 500. It is a first impression, not a verdict.
2. Read the 2-star reviews first. They tend to be specific and operational — battery life, fit, durability — where 5-star reviews are often just enthusiasm.
3. Check the most recent reviews. A product can coast on an old average long after quality slips, and the aggregate will not tell you.
4. Compare rating volume between options. If one has 20,000 and the other 400, the difference in confidence is larger than any gap in stars.
5. Do not read a high count as quality. It means many people bought it. That is not the same thing, and this dataset proves it.

What this data cannot tell you

We deliberately do not quote individual reviewers here. Aggregate ratings are what we scraped and can verify; pulling individual quotes would mean presenting a handful of opinions as though they were the pattern. That would be a worse product than the honest numbers.

The numbers are also incomplete in a more basic way. They say nothing about how a game feels to play, whether a chair suits your back, or whether a headset clamps. Those are the things you actually care about, and no aggregate resolves them. What the data does do is tell you which products have been tested by enough people that their score means something — and that is a real, if limited, advantage over buying blind.

Every shelf on this site is ordered so the most-rated products come first. That is the closest thing to a durable signal available before you press buy.

Where to look next

Browse the PlayStation shelf, the Nintendo Switch shelf, or the gear shelf — all ordered by rating depth rather than commission.