Check this out, Mike Zuurman at XeNTaX have done statistical analysis of MG. Here's the abstract, where you can leave comments, and full article in PDF. Enjoy :)

Looks like I'm one-upped!

Ain't that a good time to raise the quality plank? ;) Also, what do you think about decline in number of games after 2008? GameFly incident surely contributed towards that trend, but there are surely more factors than this. For example Giant Bomb was launched in 2008, could some users migrate there?

I think one reason might be that most of the top game contributors focus on the past. Here is a look at the Top 10 of each year:
2007: 1254 games
2008: 1857 games
2009: 2339 games
2010: 1908 games
2011: 1390 games
2012: 2317 games

That's what I wanted to say: such analysis must take into account that we're not paid researchers but hobbyists, who add what they please. Also we're not a news site, so we tend to lag behind when it comes to latest games (this was true even before 2010 member loss).

Analysis based solely on MobyGames data can be only about... MobyGames itself (just what vedder always did). Not using other sources not only is a critical flaw in research method (if you want to be serious and not just-another-kid-with-internet), even untrained eye will see that C64, Atari 8-bit or mobile games are way underrepresented in our catalog, so even if you skip minor platforms (as he did), you still have holes in your data.

Certain mentioned trends are true (e.g. 8-bit era being multi-plaform, Windows-only releases), but I wouldn't go as far as quoting the numbers for truth.

Here's an interesting message buried in that report: "Of the 59936 entries, 59912 (99.96%) had a Year of Release."

That means that we have 24 entries with no year of release! I have only ever found one, through the Random Game button, and quickly corrected it with an accurate date submission, but I am curious what the other ones are.

(I also like to use these opportunities to indicate that the world does not explode when a dateless entry is approved, it just makes game browser lists look wonky with a dateless entry on top. There ought to be many more dateless entries for games whose definite date will never be known.)

Next chart also reveals some interesting stats about how many games have been ported to how many platforms -- most games are on one platform only, but there is one game that exists on 22! 1 on 21, 2 on 20, 2 on 19... we have long debated which games those would be -- Lemmings, Prince of Persia, what do you think

Well, Pac-Man has 25...

Huh, that detail slipped their net. Of course, Tetris should be up there also.

Not really, it should be noted that not all platforms have been taken into consideration in their study. (Lemmings has 27, btw)

But that's because they didn't include all of out platforms in their research. They seem to focus on platforms of US/UK origin and no mobile platforms.

I did some scripting, Lemmings is indeed the game with the most platforms (not counting different versions as one). Here's the list.

Hey, that's very interesting, though perhaps nothing there is too surprising 8) (Was the HHG game extraordinary among Infocom games to getting ported to the most platforms, released at a unique moment in time to be sold for the older microcomputers as well as the newer ones?)

This is an incredibly awesome thing. I have thought about the possibility of conducting some serious statistical research suing the data from MobyGames. Is there any way to quickly extract the data (release date, platform, etc., etc.) from the site in some already suable format like Microsoft Excel? That would certainly simplify such searches.

Stuff I have thought would be interesting to look at are words that are most frequently used in game titles (perhaps arranged by time periods to uncover possible game titling tendencies) and also the ratio of original games and ports from other platforms per platform on a yearly basis.