A Visual History of Genres and Platforms

Only the other day we were discussing the possibility of creating a history of game genres by parsing data.

And here, NcikVGG has posted such a history on Reddit. This is scraped from Videogamegeek. The top part concerns genres, and the bottom part concerns platforms.

Looking at it, the truth is that the platform part is more useful since the data it concerns is less prone to changing interpretations: a SNES is still a SNES after all these years. But the genre chart really shows how genre labels change over time: “arcade” is no longer a useful category, “action” and “scrolling” even less so.

 

As a supplementary chart, NcikVGG has posted a platform release history with an absolute vertical axis (counting # of releases). This one shows just how many games are being released these days compared to earlier days. It also shows us how important mobile platforms have become (though Android seems to be missing).

 

16 thoughts on “A Visual History of Genres and Platforms”

  1. Wow, it’s hard to imagine that we release only about 2.5 times as many games now as we did in 1985. I take it that this is a bit skewed for consoles and mobile platforms vs. pc, as there may be many more games being made for the PC (through those online game portals), but they probably don’t count as real releases. In contrast, probably anything that gets published on the app or android store is considered a real release.

    I do agree that the genre picture is a bit useless. Many of those genres overlap (I do like the concept of scrolling games ;))

  2. @Nick Yes, there is a question of how the count is made. I wonder if old platforms like the C64 and Amstrad appear to have so many games because there are communities that spend time tracking every single game made, whereas there are no similar communities for Flash games?

    At the same time, the dip from 1990 to 2005 could reflect the fact that this period was dominated by consoles and that very few smaller-scale games were being distributed during that time period.

  3. It might be due to the differences in scope of video games. The old-arcade types of games are generally quite short and feature only one form of game play. The new flashy games of these times are filled to the brim with everything one can think of (and thus have ridiculously long development times).

    I think you’ve got a point with the tracking of the old games. Quite typical btw how there is so little data readily available on the amount of games released for each platform in a certain year. It’s a good thing to have these charts at least!

  4. Really interesting visualisation, it’s just a shame it’s based on such a flawed dataset. The platform one is at least somewhat acceptable and useful, the genre chart only shows one thing: Their classification is not up to the task. Just look at the adventure category. The heyday of that genre with Sierra and Lucas Arts fighting it out shows a huge decrease in relative importance. That makes no sense.

  5. @Nick, Jan It would be great to have a chart like this based on dataset for which we all have confidence.

    As for the adventure genre, nothing says that popularity leads to many games being made in that genre. Think only of the Sims.

  6. It could be that the chart is based on a tiny selection of games from the http://videogamegeek.com/ database. The problem with charts is that they are only as good as the data provided.

    The chart seems to show that there were more Amstrad CPC games then Sinclaur Spectrum games, aboud 1200 to 950 games. However a glance at http://worldofspectrum.org would seem to suggest that there are over 12000 games on Sinclair Spectrum, removing compliations, budget releases and oversea releases it it still around 7000 odd games (as a note that is more then the combined number of game released on the all the Nintendo consoles worldwide ever).

    Also where do you draw the line in counting games? Do you just count commericial releases and if so, how would you define a commericial game, a lot of early games were sold via mail order, literally by the author placing an advert in a magazine and making up copies as he got orders. I remember selling a Dragon 32 game that the author said he sold about 30 copies back to the author as he lost his original copy.

    Ditto Windows and MS-DOS game releases, the large number of freeware and public domain games worldwide does mean that they are probably somewhere in the region of a quarter of a million games on the platform over the last 30 years. Granted a lot of them are going to be utter trash but if you did use this then the graph would be basically a single colour…

    Other machines do seem to have either near (but not quite) the correct amount of games or are way way short (FM Towns has a page of games, just another 1400 to go…), so although it is nice eye candy it does not really show anything useful. It would be a bit like me trying to show gaming trends by asking the 150 or so people in this building.

    Also the Genres are fairly useless, scrolling is a description and not a genre. You can have scrolling platforms, scrolling beat ’em ups, scrolling shooters and even scrolling action games, but there is no such thing as a scrolling game… Ditto a text adventure to most people is basically the computer equivilent of a paperback novel, however a digital novel is more a graphical adventure rather then a text adventure game.

    There are basically too many genres, it could be whittled down to about 10 which would give more useful information.

  7. I agree with Barry H. There are ~12000 Flash games at MochiAds, 122,127 games on the Apple App Store, etc. I think many platforms are heavily underrepresented.

  8. @Barry, Christian
    Definitely: the Videogamegeek database is surely inadequate. It would be great if we had a more reliable collection.

    I think one criteria for inclusion would be public distribution.

    The genres are completely messed up of course.

  9. I agree with many here that the dataset appears to have various issues. I also think that, in addition to counting releases (which, it appears is problematic in this visualization), we also should think about the weight of those releases.

    At GDC, they continuously discussed how there are so many gaming apps released for iOS and Android, but some of these releases may have only a handful of buyers. This doesn’t necessarily translate to the quality of those releases but it does speak to the high amount of visibility game releases today may have over our memory of game releases of the past and in what form (mail-order releases, Flash games, etc.).

    Also, at what point do we start considering the games (and genres) that have a bigger popular impact? I’m not saying these games are more or less important, but they do have significantly different kinds of impact on popular perceptions of importance. For example, according to one of the charts, there are way more releases for iPad than for the major platforms or PC, yet many of the most discussed and popularized games are on traditional gaming platforms or PC.

  10. @Gabriela Agreed. I was thinking that there are at least six obvious ways to weigh the games of each platform in order to give a different kind of platform graph:

    -Count releases as was done here
    -Count units sold
    -Count income (adjusted for inflation)
    -Count hours spent playing
    -Count mind share (perhaps media citations)
    -Weigh how influential a game was on other games

  11. @Jesper Juul – One issue with collating data is finding useful data, especially for anything pre 1990s.

    Count releases as was done here – It is a pity that it took a small and incomplete database and generated basically a pretty graph with no useful information. It could be possible with several months to work out the number of releases on nearly every major machine from the US, Japan and the UK (as these three countries did make most of the computers and consoles sold in the world, even semi large companies like Radofin (who did make a number of machines) had a UK origin, despite being based in Hong Kong) and then to produce a more interesting graph.

    Genres on the other hand are more subjective, newer games tend to be jack of all trades and master of none. You could whittle it down to Action, Thought and Reflection and Simulation and cover nearly everything.

    Count units sold – On things like Steam, Xbox Live, PS Network, etc. It is quite hard to find out how many titles a unit has sold, except for when press releases are made and even then it is impossible to really verify. It is the same for anything pre late 90s, although big games you have an idea of what a game might have sold there is no indepedent numbers for sales.

    Also the further you go back the fewer computers and consoles were sold. So games sold against user base, technically a games that sold 2000 copies on a Sinclair ZX81 / Timex 1000 would probably have the same ratio as an Apple Store App that has sold about 500,000 copies. So you would have to factor in a lot of weighting to get anything useful.

    Count income (adjusted for inflation) – If you did this for one specific country the maybe, doing it worldwide with the 6 largest markets over the last three decades (USA. Canada. Japan, UK, France and Germany) then it would be a lot of fun trying to work out what the income was, especially for older games. It would probably take one person several lifetimes to do this…

    Count hours spent playing – How would you do this? Pong was a generic game that was cloned to death by countless arcade companies like Atari, Taito, Williams, Bally, Sega, etc and by countless chip makers like General Instruments, Texas Instruments, etc so was effectively available from 1972 to the early 80s and even today still versions are available to buy, so I would hate to really sit down and work out a rough hour count for that alone…. Ditto games like Pac Man, Frogger, Breakout, Space Invaders and Tetris probably have more hours played on them then probably the entire Playstation 3 library ever has.

    Count mind share (perhaps media citations) – That would skew games towards anything post internet due to the sheer number of references to anything new, although early games like Pong, Pac Man, etc would also appear quite hightly,I can’t really see it producing anything really useful.

    Weigh how influential a game was on other games – That is quite subjective and given the sheer number of games out there how would you even start?

    @Gabriela

    All computers and consoles had their shovelware. From the hundreds of arcade ‘clones’ on early computers, to the glut of platformers in the 90s to the number of puzzle games that are available now on everything that seems to have a display.

    One way that could be use could be make a snapshot of every year since about 1972 (release of Magnavox Odyssey and the start of arcade games) or 1977 (Atari VCS) of a proportional percentage of important games against games released that year. So the early years would be a handful of games to say a few hundred for now. This could generate a useful weighted graph of genres / machines, although it would still be several months (years?) work. It would have to be country focused (ie America) or have someone who had a decent worldwide knowledge of games to choose from North America, Japan, Asia and Europe…

  12. @Barry How would you go about making a complete list of releases per platform?
    My guess is that it would be easier for consoles and other closed platforms, given that each release has to be greenlighted in some way by a central authority whose records one could conceivably get access to. It would be much harder to do for personal computers and other open platforms.

    My point is that no method of weighing releases is *the right one* – they each show different things and come with their own challenges. # of releases though probably is the one with the least methodological difficulties all things considered, and the one that we can most clearly explain what means…

Leave a Reply to Nick Cancel reply

Your email address will not be published. Required fields are marked *