The statistics have a major flaw. You cannot blindly mix genres like that.

Seriously look at the short games. Pinball, at 9 minutes. Battle Colosseum at 10 minutes. Browse the <30m crowd and you'll see a bunch of short-session games that are designed to be short.

Dr. Mario, Pinball, and PvP Melee are all short games. But they are games you replay thousands of times.

The Final Fantasy and Elder Scrolls and Dragon Quests are long games. You might replay them once or twice, but that's it.

The site lumps the 5 minute Dr. Mario style games with the 60 hour Dragon Quest style games since they both came on the same console, then reports an average play time is 2 hours. That is a far less valuable statistic.

It tells more about the style of games -- many short games or many long games -- rather than the relative length within a genre.

I think both are important to different people.

I also don't think games are getting shorter in general. I think the genre of the average game has changed over the years. PS1 and PS2 tended to be the era of the 40 hour RPG. The industry since then has been trending more towards shooters which last around 10 hours. Shooters in the past were much shorter though. I could finish Goldeneye on the N64 in a couple hours if I took my time. The statistics in that link are much too simple to analyze the trends of game length.

I don't think the length of a game should determine its price point either. That makes no sense. I could add hours to an RPG by tuning the experience points or the random encounter rate, but making a shooter longer without affecting the difficulty would require adding levels and story (audio + animation). It's not a fair comparison.

One thing I do agree with from the article is that there should be a varied pricing structure for games that aren't quite AAA... But don't we have that already? Nothing is stopping a studio or publisher from releasing the game on the PSN or Xbox Live marketplace (or steam/origin/etc.) for a lower price point. However, I don't see this being feasible for a game being sold through retailers requiring manufacturing and distribution networks.

I think the article is misleading and stuck in the past. There is no way a game like Bulletstorm gets made to be a packaged good in today's world. It came out too early for publishers to get on the downloadable title bandwagon and therefore resulted in a shorter game at a higher price point.

What about games that are essentially multiplayer? (like Chess, Spacewar!, most 4X, etc).

I believe this stat does a poor job of reflecting the entirety of games. By choosing a sample, the author is making a consciously subjective choice.

Does anyone know of a better site the author could have used?

Does anyone know of a better site the author could have used?

Youtube search for "speedrun" since he is looking for minimal playthough time.

First of all; it's quite clear from the title, being an open question, that it's not conclusive or comprehensive. The author is a student journalist.

Secondly, scraping youtube is not an easy task and would not lead to any structured data.

The author clearly states his source. Adrian Chmielarz seemed quite impressed as he tweeted it.

Does anyone know of a better site the author could have used?

Youtube search for "speedrun" since he is looking for minimal playthough time.

http://speeddemosarchive.com/

You'll get everything you need for retro games there.

However, note that many of the 'best time' are contingent on bugs found within the last 5 years, which drastically diminishes the run for a number of these games (such as Chrono Trigger and Zelda: Ocarina of Time, as seen at AGDQ 2013).