via Spong

With all the next-gen news of late (and flurries of name calling based on leaked stats, manipulated figures and plain nonsense) the phrase “There are three kinds of lies: lies, damned lies and statistics”†, appears to be more on the mark than ever.

But sometimes statistics can throw up some truly intriguing information. The ‘Many Eyes’ project’s findings regarding average game review scores for last-gen systems including Xbox, PS2, GameCube and Dreamcast, present one of these moments.

According to its ‘visualisation’ (pictured right), which is based on data from the Metacritic review-score site, both the Xbox and the PlayStation 2 are beaten by GameCube and Dreamcast in terms of games receiving the highest-scores.

Very little other information, aside from the raw data is given. So, for example, we don’t know over what period the information derives from (although Metacritic began business in 1999). However, what it does tell us (or in this case show us) is the following:

Of games rated between 97-100 points (from 100), the Dreamcast caught three while the Xbox and PS2 attracted two and one respectively. At the lower end of the scale, games that rated at less than 29 points appeared five times on the PS2, six times on the Xbox, twice on the GameCube; no Dreamcast games rated that low-down.

The number of review ratings used by the data in total is 2,913. This breaks down as follows:

PS2 – 1,417
Xbox – 854
GameCube – 500
Dreamcast - 142

What is also made clear by the stats is that the media likes to be 'nice', with the bulk of the review scores (2,077) appearing between the 57 to 84 point range.

The Many Eyes project is part of the part of IBM's Collaborative User Experience research group – and there’s a vague chance that you could take it seriously. However, the more you drill into the stats, the more questions need to be asked, including: with more than 2,500 PS2 games released in Europe, why only 1,147 reviews in this sample? Why so few Dreamcast games - 142? Which sites did Metacritic use?

So, once again, while looking over the lovely graph, bear in mind that stats are only as good as the information supporting them.