More content

The first part of this project has some important context about how this study was conducted.

I realized I had a lot of data I could find that wouldn’t really make for a good graph, and some trivia that didn’t fit anywhere. This is a more loosely structured post, so please forgive the abrupt changes of subject.

More on Metascores

The day with the most games released with a 90+ metascore was, well, a two way tie.

  • February 29th, 2000
    • Dead Or Alive 2 (Dreamcast)
    • Tony Hawk’s Pro Skater (N64)
    • Rayman 2: The Great Escape (Dreamcast)
    • Resident Evil: Code Veronica (Dreamcast)
  • November 18th, 2001
    • Star Wars Rogue Squadron II: Rogue Leader (Gamecube)
    • Madden NFL 2002 (Gamecube)
    • IL2-Sturmovik (PC)
    • Tony Hawk’s Pro Skater 3 (Gamecube)

Great days for Dreamcast and Gamecube. Wow, Dreamcast had 7 games release on Leap Day, and Chu Chu Rocket was pretty close to the 90+ club.

The greatest drought between games with a 90+ metascore was… ok, I was going to say between Realms of the Haunting and Dungeon Keeper, but in double checking I have now found 3 possible release dates for Realms and the one I have recorded is probably wrong.

Let’s instead go with Out of the Park Baseball (3/23/06) and Company of Heroes (9/13/06), a period of 174 days. Remember how 2006 saw that dip in metascores?

Two games, one name, two developers, one score.

More on Descriptors

Are you ready for more content descriptor details? The game with the most descriptors I could find was 9, and it’s not something you would expect:

I didn’t get into this before, but the ESRB combines substance descriptors if they’re both “use” or both “reference” for some reason. Not sure any game has all three. That was a bit of a pain to deal with.

Anyway, I have no idea why this singing game has so many, and oddly enough the Switch version only has 8, no Partial Nudity.

Several other games had 8 descriptors, most are current generation:

  • Constructor Plus (Switch, 3DS, PS4)
    • Blood, Crude Humor, Drug Reference, Language, Suggestive Themes, Use of Alcohol, Use of Tobacco, Violence
  • Constructor HD (Xbox One, PS4)
    • Blood, Crude Humor, Drug Reference, Language, Suggestive Themes, Use of Alcohol, Use of Tobacco, Violence
  • Night in the Woods (Switch, PS4, Xbox One, PC)
    • Fantasy Violence, Mild Blood, Sexual Themes, Drug Reference, Language, Crude Humor, Use of Alcohol, Use of Tobacco
  • The Red Strings Club (Switch)
    • Blood, Drug Reference, Nudity, Strong Language, Suggestive Themes, Use of Alcohol, Use of Tobacco, Violence
  • Thimbleweed Park (Switch, PS4, Xbox One, PC)
    • Crude Humor, Drug Reference, Language, MildBlooo, MildVio, Sexual Themes, Use of Alcohol, Use of Tobacco
  • Def Jam Rapstar (Wii, PS3, Xbox 360)
    • Drug Reference, Mild Blood, Mild Violence, Partial Nudity, Sexual Themes, Strong Lyrics, Use of Alcohol, Use of Tobacco
  • Grand Theft Auto V (PS3, PS4, Xbox 360, Xbox One, PC)
    • Blood and Gore, Intense Violence, Mature Humor, Nudity, Strong Language, Strong Sexual Content, Use of Drugs, Use of Alcohol
  • Duke Nukem Forever (PS3, Xbox 360, PC)
    • Blood and Gore, Intense Violence, Mature Humor, Nudity, Strong Language, Strong Sexual Content, Use of Drugs, Use of Alcohol
  • Skylight Freerange (Vita)
    • Violence, Blood, Sexual Themes, Nudity, Drug Reference, Strong Language, Use of Alcohol, Use of Tobacco
  • Skylight Freerange 2 (Vita)
    • Violence, Blood, Sexual Themes, Nudity, Drug Reference, Strong Language, Use of Alcohol, Use of Tobacco

The only game with both Mild Fantasy Violence and Mild Cartoon Violence was Xbox Live Arcade Unplugged Vol. 1. Okay, a compilation, but it doesn’t seem like there should be a reason those don’t ever appear together otherwise.

The only game with Nudity and Simulating Gambling was Dead or Alive Xtreme Beach Volleyball.

I had to check the box to make sure it wasn’t an error on the ESRB website. But yes, Loons has two degrees of violence. I don’t think any other game I looked at had two “tiers” of something. I have to imagine it was a mistake.

Is there any actual content differences between these versions? I believe the Vita version was released before the “rate it yourself for cheaper” program.

I’m again not familiar with these two games, but could the Cartoon Violence really be significantly more major in one version? Usually when a game is released on multiple platforms, even months apart, it gets one entry if they have the same content. Maybe for some reason different people handled these two and had different opinions?

Fantasy Violence wasn’t a descriptor when the original was released, so fair enough there, but that is quite a difference in descriptors.

 

More on the ESRB Website and ESRB Weirdness

This game was known as Speedster in Europe, and Rush Hour in North America. Why does a North American rating system website have the European title at all? This was the only two name case like this I came across.

Released on “Nintendo”, I came across a few of those.

I was pretty confused by this when I saw it, but I found an article where the creator discusses what and why he tones down the Wii U version.

Should a crossword game where the clues are a bit off-color (this actually sounds extreme for a newspaper puzzle) count as much as if there were actual characters engaging in these acts? It’s a bit of an edge case. It’s also funny that they censored “ass” on this website aimed at adult parents.

The ESRB’s website will ignore spaces when searching titles. Finding the rating of a Playstation game called “One” was quite an experience. It was on something like the 37th page (results are chronological) and you can only go forward one page at a time. You can filter by system, but only for some newer systems because why would would anyone care about something old.

If you were wondering why I didn’t just look at a picture of One’s case on eBay, I did. Games are supposed to list their descriptors on the back, below the ESRB rating, but this one didn’t.

This My Little Pony game doesn’t have any ESRB information at all on the back. I can’t find much information about game box art requirements, like what size things have to be, where they need to be placed, just “in June 2003 — the ESRB announced new labeling procedures (requiring the prominent display of back-of-the-box information)“, which is odd since the overwhelming majority of games already had that. I suppose publishers handle box art, but there is still enough structure to them that they must have a lot of requirements.

After a bit of browsing, I found this E.T. game for Game Boy Advance without any ESRB information on the back.

The Best, worst and Most Average game

I realized that with all of the data I have that I can construct what would theoretically be the best and worst game possible, based on averages. This is a for fun exercise, please don’t take it too seriously.

The worst possible game, based on average metascores:

  • System: Wii
  • Release Date: August 4th, 2007
  • Genre: Educational
  • ESRB Rating: Everyone (Early Childhood games don’t even get ratings so I won’t assume)
  • ESRB Content Descriptors: None
  • Length: As short as possible, definitely less than 4.5 hours to beat

The best possible game, based on average metascores:

  • System: Xbox One
  • Release Date: December 28, 2019
  • Genre: Role-Playing and Compilation
  • ESRB Rating: Mature
  • ESRB Content Descriptors: Sexual Content, Nudity, Use of Tobacco, Use of Alcohol, Use of Drugs, Intense Violence, Blood and Gore
  • Length: As long as possible, at least 15 hours to beat

That best possible game sounds a bit like a Mass Effect trilogy compilation, which has been rumored as of this writing (May 9th).

Most average game, based on average metascores:

  • System: Game Boy Color
  • Release Date: February 3, 2010
  • Genre: Simulation
  • ESRB Rating: E10+
  • ESRB Content Descriptors: Crude Humor
  • Length: 7.5 hours to beat

Ignoring system, but considering proximity to the actual average score of 70.25, we might label The Urbz: Sims in the City to be the most average game of the past 25 years. It doesn’t quite fit every parameter, but nothing will. I’m sure there’s better ways of finding “most average” than looking at what parameters line up in the middle of metascores, too.

That’s all for the 25 year project. Is there a specific thing that I could find quickly with the information I have that you’re curious about?

Sources

GameRankings.com for metascores and some release dates. Archives: 1, 2, 3

MobyGames.com for genres, some release dates, some ESRB ratings, and some ESRB content descriptors

HowLongToBeat.com for main story and completionist times.

ESRB.org for ESRB ratings and ESRB content descriptors

Wikipedia for its many lists of games

Content Descriptors Background

The first part of this project has some important context about how this study was conducted.

The ESRB does not just give games ratings, it also uses content descriptors to more specifically describe content that consumers may find objectionable. Very similar, appearing in the same place as content descriptors on the back of game boxes, are also “Interactive Elements” that describe online features that consumers may want to know about before buying a game, such as In-Game Purchases or Shares Location. I did not include interactive elements in this study.

I find these content descriptors so interesting for two reasons. One, because it’s a look at how an organization tries to organize and sort thousands of games based largely on what parents might not want their children to be exposed to, and two, because it’s a formal description of what kind things happen inside a game. No one else is going to make a list of games that have “Cartoon Violence, but it’s pretty mild” but a ratings agency.

By my count there have been 49 content descriptors used during the life of the ESRB. Several have been retired. A couple seem to be “retired” in that they have been replaced with other descriptors but have been used a few times since their retirement, perhaps in error. Two, as far as I can tell, have never been used.

Some Adult Assistance May Be Needed shows up on many lists of content descriptors, including Wikipedia’s, but does not appear on the ESRB’s list of descriptors, though presumably it did at some point. These lists say that it is exclusive to Early Childhood games. MobyGames claims a single game has received this descriptor, Disney’s Little Einsteins. This game is rated Everyone, not Early Childhood. The ESRB’s page for the game says it has no descriptors. GameFAQ’s scans of the box also show no descriptors.

Real Gambling, used for games where real money is exchanged while gambling, is listed on the ESRB’s list of descriptors, but I can find no evidence it has ever been used. You may be aware that Peak Entertainment Casinos is the only game to receive an Adults Only rating for having real gambling, and that is true. However, it has the content descriptor “Gambling”, a descriptor that was retired, seemingly split into Simulated Gambling and Real Gambling. The Gambling descriptor was also used for Dead or Alive Xtreme Beach Volleyball, which only features gambling for in-game currency.

There are many descriptors with “mild” prefixes, but the ESRB does not acknowledge these as separate descriptors from the non-prefixed versions (except for the substance ones, which are worded differently), on their website, but do list the “strong” prefixed versions.

The ESRB categorizes content descriptors, but I did not realize this before I had already categorized them myself. The only difference ended up being that I put blood and violence in the same category, and I included retired descriptors.

ESRB Content Descriptors – Distribution

I’m starting with this mammoth chart so I can list all of the descriptors and their approximate dates of use. You’ll notice some descriptors used once or twice and then not used for years before being used regularly, like Tobacco Reference appearing in one game, 4 years before any other substance reference descriptor appeared.

I looked it up and that game was Commandos: Beyond the Call of Duty (an expansion pack, but a standalone one, so it counts). If you go to ESRB.org (where I got most of my ESRB information) and search for that game, you’ll see this:

But if you go to Steam, where it is currently for sale, you’ll see this:

So, which is correct? The Steam listing has “Animated Violence”, one of those “retired but not” descriptors, which makes it look much older than the ESRB one. It’s possible Steam used the game’s old, original ESRB rating for some reason (the back of the box seems to confirm this was the original rating), and then the game got re-rated by the ESRB some time later. Why would the game get re-rated if if not for the Steam release, though?

I spent at least 20 minutes putting together the evidence for this one specific game and I still don’t know what should “count”. The ESRB website has a lot of odd conflicting information and I didn’t have the will to double check and investigate every single one.

Furthering my point that it’s very hard to determine the history of descriptors let’s look at MobyGame’s list of games with Animated Violence. One game in 2018, one in 2011, a handful until 2003, and then dozens every year before that. The ESRB’s content descriptor list in February 2004 doesn’t mention it, but a list from February 2001 has it under “discontinued content descriptors”. And yet that MobyGames list has dozens of uses of the descriptor in 2001 and 2002! And looking at that ESRB page from 2001 page I am seeing for the first time the descriptor “Reading Skills, Fine Motor Skills, Higher-Level Thinking Skills”. Searching for that exact phrase in Google right now returns 25 results, all pretty much verbatim copies of that outdated list, no games mentioned as using it, so I guess that’s three never used descriptors. No one has found it interesting enough to discuss until now, either. I have really been down a rabbit hole on this subject.

So, as I discuss when these descriptors began and ended I will be ignoring some outliers. If nothing is stated, then the descriptor has been in continuous use since the start of this study, in 1995.

  • Substances
    • Alcohol Reference – Started in 2003 with the other [Substance] Reference descriptors
    • Use of Alcohol
    • Drug Reference – Started in 2003 with the other [Substance] Reference descriptors
    • Use of Drugs – Started in 2002, it seems odd it took longer than Alcohol and Tobacco.
    • Tobacco Reference – Started in 2003 with the other [Substance] Reference descriptors
    • Use of Tobacco
  • Violence
    • Animated Blood – Animated in these descriptors means “cartoony”, not that it is moving.
    • Animated Blood and Gore – Probably meant to be retired in 2002, but has popped up once in a while since then.
    • Mild Animated Violence – Probably meant to be retired in 2002, but has popped up once in a while since then.
    • Animated Violence – Probably meant to be retired in 2002, but has popped up once in a while since then.
    • Mild Blood – Probably started in 2006, that was the first year I have with more than one use.
    • Blood – Started in 1997, or I just got very unlucky with my small number of games from 1995 and 1996.
    • Realistic Blood – Stopped being seen after 2000, the other “Realistic” descriptors were last seen in 2001.
    • Blood and Gore
    • Realistic Blood and Gore – This may have been intended as a stronger version of Blood and Gore because they coexisted, but last seen in 2001.
    • Mild Cartoon Violence – Started 2003.
    • Cartoon Violence – Started 2003. One of “4” (they didn’t include mild versions as separate, so it should be 6) new descriptors announced in June 2003.
    • Mild Fantasy Violence – Started 2003.
    • Fantasy Violence – One of the “4” new descriptors announced in June 2003. I have several from before then, somehow.
    • Intense Violence – One of the “4” new descriptors announced in June 2003.
    • Mild Realistic Violence – Discontinued in 1996 or 1999.
    • Realistic Violence – Gone after 2001.
    • Mild Violence
    • Violence
    • Violent References – Began in 2007.
  • Humor and Mischief
    • Comic Mischief
    • Crude Humor – Probably started use in 2003, although never got a press release like the “4” did.
    • Mature Humor – Probably started use in 2003. Likely a “Strong” version of Crude Humor as we’ll see later.
  • Educational
    • Edutainment – Last seen in 2008.
    • Informational – Last seen 2001.
  • Gambling
    • Simulated Gambling – Started in 2004.
    • Gambling – In use 2000-2004, probably split into Simulated Gambling and the never seen Real Gambling.
    • Gaming – Last seen 2000. Seems to have been replaced with the short lived Gambling. This word is occasionally used as a way to say “gambling” but feels a bit old fashioned, perhaps due to the rise of video and computer games.
  • Language
    • Mild Language
    • Language – Either started in 1998, or I got unlucky with my sample. MobyGame’s list makes it look like it may have indeed started then. It would be odd if it was introduced after its Mild and Strong versions.
    • Strong Language
    • Mild Lyrics – Introduced in 2001.
    • Lyrics – Seemingly introduced in 2004, also after its Mild and Strong versions.
    • Strong Lyrics – Introduced in 2001.
  • Sexual Content
    • Partial Nudity – First used 2001, maybe 2002.
    • Nudity – First used 2002.
    • Sexual Content – First used 2008, or maybe 2006.
    • Strong Sexual Content – Probably around for 25+ years. Oddly, I have no 2019 games with this after a 17 year run.
    • Mild Sexual Themes – Introduced 2008, significantly after its non-prefixed version.
    • Sexual Themes – I’m inclined to think this started in 2004, if not then it’s an odd case of a descriptor suddenly rocketing up in use.
    • Mature Sexual Themes – Last used 2004. It’s possible this was replaced with Sexual Themes, which seems like maybe a way to downplay it.
    • Sexual Violence – This is so rare it’s hard to say. Only seen in 2014. In fact, it was only seen in one game in this study, which was ported to 4 systems: Metal Gear Solid V: Ground Zeroes.
    • Mild Suggestive Themes – Began use in 2004.
    • Suggestive Themes

Alright, now I can get to the graphs I usually start with. Please note the Y-axis of these graphs, some categories of descriptors are much more common than others. Alcohol is the most common substance overall and drugs the least. There seems to be more and more smoking in video games over time, quite the opposite of the movie industry.

It’s odd how Animated Violence was semi-retired just as Violence took off (but it did already exist). They are distinct things so it wasn’t a replacement.

A sharp decline in Comic Mischief after a huge spike. I have to wonder how the raters are trained on this kind of thing, do they just read a list of descriptors? Has the public perception of what is Comic Mischief changed over time, or have game developers decided to employ it less than a tenth as often as they did a decade ago?

Two very rarely used descriptors that haven’t been seen in a long time. Informational is an odd one, it was for games with reference material, data, that kind of thing. The only games I have using it were Need for Speed: Porsche Unleashed, Virtual Kasparaov, Timeline, and Timescape: Journey to Pompeii.

It took a few tries, but the ESRB eventually decided what term to use. Gambling is pretty low overall, perhaps PEGI (the European equivalent of the ESRB) standards for games with gambling influenced developers worldwide.

It took a while before technology allowed the playing of music with potentially objectionable lyrics, but it’s never been very common. I wonder if the shift towards real voice clips over text influenced the rise of bad language.

Many of these descriptors sound similar. Aren’t they all “Sexual Content”, why is that a separate descriptor? This article is a good overview of how all of these differ. Mild Suggestive Themes was sure big for a while, but then drastically fell in use.

I see now that the long bar graph is missing Alcohol Reference, sorry about that. Violence is by far the most common descriptor used, with Blood at #2 with just over half as many uses . Four ports of Metal Gear Solid V: Ground Zeroes makes Sexual Violence the second rarest descriptor. But the rarest, Mild Realistic Violence, was only seen in Noir: A Shadowy Thriller and Professional Bull Rider.

Overall, all kinds of violence dominate games.

 

Here is how common every combination of descriptors is. The white cells are the overall percent of games that have that descriptor, and then the pink cells at the bottom are what percent of games have that descriptor as their sole descriptor.

It’s safe to say that games have added more and more things to warn about over time, but the ESRB has also expanded what kinds of things they have descriptors for over time. Referring to substance use wasn’t deemed worthy of note until 2003, for instance. I’m sure a number of games released before then would have qualified.

This graph fits pretty well with the general trends of more descriptors over time, and how descriptors relate to ESRB ratings, which we’ll look at later.

ESRB Content Descriptors + Metascores

The better a game is, the more content descriptors it has on average, with quite a jump in the highest tier. Do you think people are more satisfied the more realistic the violence, the more graphic the sexual content?

Games with Sexual Content are more liked than games with any other descriptor. Yet Strong Sexual Content doesn’t do quite as well. Sex in general is a pretty strong indicator of quality, remember that the overall metascore average is 70.25, every sex descriptor is above that. The three substance use descriptors are very close in scores and above average, but merely referring to those same substances scores several fewer metascore points.

Simulated Gambling, perhaps an in-game casino, really seems to turn off reviewers. Cartoon Violence may be so low due to being common in cheap licensed games or just kiddie fare in general. Referring to violence is evidently less satisfying than seeing it.

ESRB Content Descriptors + Genres

In some genres a substance will be alluded to more often than used, and in other genres it’s the other way around. Makes sense for the narrative genres to have more substance descriptors in general.

Role-Playing heavily favors Fantasy Violence due to usually being have some kind of magical or supernatural combat. Compilations have the most Cartoon Violence for some reason.

Again, please note the Y-axis on these graphs will change. Action really favors Crude Humor over Comic Mischief. Not a single Educational, Puzzle, or Simulation game had Mature Humor.

No surprise that Educational games have educational descriptors.

It’s odd that a genre as general as Action is so low in gambling compared to others. I don’t associate Compilations with gambling, I have no idea why that is so high.

The lyrics descriptors more or less work as an indication of how much English language licensed music appears in games, very rarely in Role-Playing, but fairly often in Sports. Simulation includes the likes of Rock Band, so that’s why it has such a showing for lyrical content.

I’m baffled as to why Racing / Driving has the highest rates of Nudity. Role-Playing games overall rank pretty high when it comes to sexual descriptors, while Educational, Puzzle, and Sports games are decidedly unsexy.

Role-Playing attracts a lot of potential content descriptors, Compilation’s high number makes sense, every game can add a few more.

All that Genre + Descriptor data in a chart.

ESRB Content Descriptors + Game Lengths

The longer a game is the more descriptors it has on average.

ESRB Content Descriptors + ESRb Ratings

 

 

The two (of 22) Early Childhood games with descriptors had Edutainment descriptors. There were only two Adults Only games included, but there’s also only a few dozen ever published. Steady increases in descriptors as rating become more severe.

Games rated Everyone rarely have any sexual or substance descriptors, and the violence tends to be fantasy or cartoon based. About 11 descriptors appear in more than 2% of E rated games.

Fantasy Violence suddenly becomes much more prevalent with E10+ games. More descriptors in general, 20 over 2%.

Violence now rockets up in use. Only slightly more descriptors used above 2% of the time, 21.

Blood and Gore and Strong Language now become much more prevalent. Only 18 descriptors used over 2% of the time. Teen games seem to employ the largest variety of descriptors, although it’s pretty close and 2% was an arbitrary cut off.

And here is how common every descriptor is with every rating. I wonder if the ESRB has hard rules for which descriptors can be used with which ratings. It would make sense for a game with Use of Drugs to not be allowed less than a Teen rating, but is there a rule against Mild Fantasy Violence in a Mature rated game, or has there just not been any (any in this study, at least) games that have done that?

And that’s all my graphs and charts for content descriptors. There will be one more bonus post before the next project. It will have some trivia and odd things I found in the course of making all this.

Sources

GameRankings.com for metascores and some release dates. Archives: 1, 2, 3

MobyGames.com for genres, some release dates, some ESRB ratings, and some ESRB content descriptors

HowLongToBeat.com for main story and completionist times.

ESRB.org for ESRB ratings and ESRB content descriptors

Wikipedia for its many lists of games

System Infographics

The first part of this project has some important context about how this study was conducted.

As a bonus today I have system infographics summarizing some of the information I’ve gone over so far. It’s nice to directly compare systems on a variety of statistics.

There is some important information about how these numbers were obtained in the other parts of this project, please see Part 1 if you haven’t already.

Nintendo Systems

Sony Systems

Sega Systems

Microsoft Systems

Personal Computer

Sources

GameRankings.com for metascores and some release dates. Archives: 1, 2, 3

MobyGames.com for genres, some release dates, some ESRB ratings, and some ESRB content descriptors

HowLongToBeat.com for main story and completionist times.

ESRB.org for ESRB ratings and ESRB content descriptors

Wikipedia for its many lists of games

ESRB Ratings Background

The first part of this project has some important context about how this study was conducted.

The Entertainment Software Rating Board gives content ratings to games and is recognized in the United States, Canada, and Mexico. The first games to receive ratings were released midway into the 4th generation in September 1994 and included Pitfall: The Mayan Adventure, Sonic Triple Trouble, Super Punch Out!!, Donkey Kong Country, and Doom (32X). The data we’re looking at starts at the fifth generation, so most of the ESRB’s history is included.

The following summaries of the ESRB ratings are based on the ESRB’s descriptions of them, the content descriptors post will go into more detail:

Early Childhood was aimed at young children and was mostly educational games. This rating was retired in 2018 with no announcement until the ESRB replied to a tweet asking about it in 2019. I find it interesting and odd that there being few eC games would be a reason to ditch a useful label. These games had no objectionable content and can only have a small number of content descriptors, I found one used on two games in this study.

The ESRB removed any mention of this rating from their rating description page. In fact, they removed the image of the rating from their website completely:

I took that screen capture several months ago and I was double checking a few things while writing this and it now seems to be impossible to find Early Childhood games by searching the ESRB website:

But they can still be found with a google search, although there’s just a blank space now instead of a broken image:

Everyone was known as “Kids to Adults” until 1998. These games can have only mildly objectionable content such as mild language or violence. Some games rated Everyone have content descriptors, some do not.

Everyone 10+ was introduced in March 2005 as a rating for games in between Everyone and Teen and suitable for players at least 10 years old. All E10+ games have at least one content descriptor.

Some games that got re-rated later were changed from E or T to E10+:

Teen rated games are intended for audiences at least 13 years old. Games with this rating may include simulated gambling or some blood. All Teen games have at least one content descriptor.

Mature rated games are intended for ages 17 and up. They can include intense and realistic violence, language, and sexual content and they all have at least one content descriptor.

Adults Only rated games are for those 18 and up. Games with this rating can have extreme violence, graphic sexual content, and gambling with real money (Peak Entertainment Casinos is the only game to have received this rating for gambling). The big four console manufacturers never allowed games with this rating to be released on their platform, and few retail stores will sell PC games with it, making it quite rare. Wikipedia has a nice list of these games.

ESRB Ratings Distribution

To start with here is the overall distribution of ESRB ratings. The extreme ends of the scale are quite small, but roughly 1 in 1,000 games are eC, and roughly 1 in 10,000 are Ao. Eight of the 22 eC games in this study were Sesame Street games, the two Ao games were Hatred and Seduce Me.

Games rated Everyone have only recently been overtaken by Teen games and the overall distribution of the ratings have become closer. What was it about 2001 and 2008 that kicked off a trend of games becoming less for Everyone?

This 6 year old article claims that E ratings represent 70% of all games. It seems like the source is the ESRB itself, which would probably want to paint games in a child-friendly light, but that number seems incredibly high to me. Sure, the kinds of games not included in this study might trend towards inoffensive, and there are mobile and other platforms not included here, but that still seems very high.

The N64 and Wii have the largest number of Early Childhood games, and Nintendo systems in general have more games on the child-friendly side.

The Game Boy Color has the highest percent of Everyone rated games, and at 93% this is the largest share of a rating for a system. If you don’t count the PC (and you shouldn’t, the unrated games take a chunk out of everything) the PS4 has the smallest percent.

The Wii U has the largest share of E10+ games, while the PS3 has the smallest, if you discount the systems that were out when the rating started to be used.

Meanwhile the Xbox is the most Teenage system, while the Game Boy Color is the least.

Most Mature goes to PS3, and there were zero Game Boy Color games rated Mature, and not just of games included in this study.

All two Adults Only titles included were for the PC.

Handhelds definitely get fewer games with objectionable content. For a while it certainly helped that they were less capable of rendering anything realistic, but even past that developers just don’t put those kinds of games there.

ESRB Ratings + Metascores

But what ratings do people prefer for their games? None of the Early Childhood games had metascores, and the very small sample size of two Adults Only games averaged 43.16.

Everyone, Everyone 10+, and Teen rated games are quite close, and their average, mean, and mode paint slightly different pictures. There is a small bump in metascores for Mature rated games. This may partly be because few cheaply made games are rated Mature and there is almost certainly a correlation between budget and metascore.

Mature rated games come out on top, both in overall 90+ metascores and in the percentage of games scoring that high.

ESRB Ratings + Genres

Adventure and Role-Playing, the narrative-driven genres, have the highest ratings overall. Is it harder to tell a compelling story without graphic violence, sexual situations, and bad language?

Educational, Puzzle, and Sports are all understandably low in objectionable content.

ESRB Ratings + Game Lengths

Games tend towards the Mature rated side of the scale the longer they are, although the longest main stories are a bit of an exception. Role-Playing games are such a large part of the longest buckets that their overall more adult ratings push things in that direction. Longer games also have more opportunities to present scenes that earn severe content descriptors such as Strong Language, Strong Sexual Themes, and Mature Humor.

But more on that next time, where I’ll go into more depth about content descriptors than you ever imagined possible.

Sources

GameRankings.com for metascores and some release dates. Archives: 1, 2, 3

MobyGames.com for genres, some release dates, some ESRB ratings, and some ESRB content descriptors

HowLongToBeat.com for main story and completionist times.

ESRB.org for ESRB ratings and ESRB content descriptors

Wikipedia for its many lists of games

Game Lengths Background

The first part of this project has some important context about how this study was conducted.

Today’s post is all about how long it takes to complete a game, both the main story, and to complete everything, which I will refer to as “game lengths”.

The data is from HowLongToBeat.com which takes user-submitted times which are averaged. This is the smallest data set involved in this study, only about half of the games included in the study had main story times, and about 40% had completionist times. The site also allows users to submit “Main + Extras” times in between Main Story and Completionist, which I found too vague and not very interesting. Most people aren’t submitting “I did the bare minimum” times for Main Story.

I didn’t include games that use the Solo/Co-Op/Vs. categories, which are mainly used for games that are solely PvP. There were a number of games that seem to have been given this designation in error.

There is a quirk with how HowLongToBeat.com displays information. Games under an hour long will display the number of minutes, while games an hour or longer will display X or X.5 hours. I suppose this is to make things look tidy, but it’s an unfortunate loss of detail.

Game Lengths – General

Games have gotten longer over time, but the gap between just beating the game, and doing everything the game has to offer has grown larger over time. Xbox’s achievement system debuted in 2005, while Playstation’s debuted in 2008, which may be responsible for that big completion jump in 2008. Game developers may have started designing their games with an idea in mind of long or difficult it should be to get every achievement and were afraid of making their games too short. A contributing factor may also be that people were less sure when to call a game “completed” when playing games without achievement systems.

Every game with game length data. These scatter plots give you an idea of the overall distribution of game lengths and how far the outliers are from the norm. There are so many types of games, and so many ways to play them. Less than a fourth of games make it beyond the first minor gridline of 25 and 50 hours, respectively.

I know you’re curious about those longest games. The longest main story games were 7 Football Manager games (2010-2016) at 287 to 519 hours. After that are The Secret World (281) , Destiny of Spirits (248), and Final Fantasy XIV (242).

For completion times we have Rock Band 4 on two different systems at 937 hours, followed by Dragon Quest IX (746), Animal Crossing: City Folk (690), and Gran Turismo 5 (636).

As someone that has played a lot of 3DS I never would have guessed it has the longest games on average. Sega was known for arcadey games that emphasized replaying over and over for better scores or times, so it makes sense to see Saturn and Dreamcast so short.

The structure of games can vary quite a bit, beating a fighting game’s campaign mode will almost always take under an hour, so systems with lots of fighting games may be skewed shorter, while people are likely to do many optional side quests in already long RPGs.

With these direct comparisons it we can see that handheld games don’t really have a tendency to be shorter than their console siblings, with the DS, 3DS, and PSP beating their console counterparts.

Although the trend is for games to get longer over time, the Wii is a notable example of games getting shorter compared to its predecessor, though only for main story.

Game Lengths + Metascores

I broke up the lengths of main story and completion into 4 buckets that each have approximately the same number of games.

People like long games. There may also be a correlation between game length, budget, polish and refinement.

Game Lengths + Genre

No surprise that Role-Playing comes out as the longest genre, with a big gap between main story and completion. I might have expected Strategy / Tactics to take second over the grab bag of games that make up Simulation. Despite a fair number of long visual novels Adventure comes in short due Western episodic games. It’s interesting that a Compilation of several games is still on average shorter than an individual game of several genres.

 

And here is how the genres are distributed in the 4 buckets. Adventure and Puzzle are the only genres that consistently become rarer as we increase time. Role-Playing really has the biggest difference between the shortest and longest bucket.

This post was a bit short, but we’re adding more and more combinations of metrics. Next time is ESRB ratings.

Sources

GameRankings.com for metascores and some release dates. Archives: 1, 2, 3

MobyGames.com for genres, some release dates, some ESRB ratings, and some ESRB content descriptors

HowLongToBeat.com for main story and completionist times.

ESRB.org for ESRB ratings and ESRB content descriptors

Wikipedia for its many lists of games

Genre Background

The first part of this project has some important context about how this study was conducted.

I searched several sites with large video game databases to decide how I was going to approach the genre section. I wasn’t totally satisfied with any of them, but I found MobyGames to be the best compromise. GameFAQs has an odd tiered system with a variable number of tiers that is inconsistent about describing certain aspects that I wouldn’t exactly call genres. GiantBomb was inconsistent and less complete. I made a list of 30ish tricky games to assign genre to and looked them up on 4 sites, and MobyGames seemed the best to me. I’m still about to point out a lot of issues, but no genre system is going to be perfect. MobyGames is also essentially a tightly controlled wiki, but things slip through the cracks and don’t get fixed, so I made changes and made things more consistent when it seemed obvious.

MobyGames has 10 genres: Action, Adventure, Compilation, Educational, Puzzle, Racing / Driving, Role-Playing (RPG), Simulation, Sports, Strategy/Tactics. That’s verbatim, and yes it drove me crazy how Racing / Driving has a space on either side of the slash, but Strategy/Tactics doesn’t, and we’re reminded how Role-Playing games are abbreviated. Games can be labeled with as many genres as needed.

Action is very general, covering fighting games, music and rhythm games, platformers, first-person shooters, and sometimes seemingly anything that isn’t turned based or menu driven. Some Sports games included it and some didn’t, confusingly. I made an effort to make it more consistent by removing it from any games that were more on the “sim” side, that weren’t arcadey or had an Actiony side mode.

Adventure, as MobyGames defines it, focuses on narrative over action, and emphasizes dialog and puzzle solving. Includes visual novels, point and clicks, walking simulators, and many games without combat.

Compilations have multiple games in one package. Does not include minigame collections like Mario Party. I removed compilation from many games that included DLC that was sold separately in a previous release, only full games put together counted for this project.

Educational doesn’t seem like it should be a genre at all to me, but these are mostly games for young children about the alphabet or basic math.

Puzzle seems to be used exclusively for games that are all about puzzles, not games that have some puzzles to solve, like the Legend of Zelda series.

Racing / Driving is included in some games with vehicular minigames or sections, like Grand Theft Auto.

Role-Playing games includes games like Dark Souls and Monster Hunter as well the The Elder Scrolls, Final Fantasy, and Persona.

Simulation for our purposes includes Cooking Mama, Nintendogs, most of the Imagine and Petz series, Phoenix Wright, and Guitar Hero. This is probably the most eclectic genre.

Sports includes wrestling, hunting, billiards, fitness, and some horse games.

Strategy/Tactics includes the Jackbox games, city builders, war games, board games, card games, and the Worms series.

Genre – General Distribution

Overall, nothing comes close to action. Yearly releases of many sports games keeps them the second most common. Role-Playing is higher than I would have expected, especially compared to Racing / Driving games.

I really like this chart, you can see how the industry has changed and how what kinds of games it makes has evolved over time. Although classic point-and-click adventures are rare now, the genre has managed to become the second largest. Some of this is because many adventure games are released episodically, and then bundled, resulting in a lot of separate game entries.

Sports games have actually become less common over time, perhaps it’s become too hard to compete with the big franchises?

PC is really an outlier here, where in other metrics it’s very close to the overall average due to making up about a fifth of the games included. It’s difficult to really see any trends among hardware companies. The time period a system exists in is probably much more important.

A bit easier to see some minimums and maximums here. Playstation Vita just barely coming out on top for RPGs. Switch really has a different makeup than other Nintendo consoles or handhelds have had. Lots of Racing / Driving games in the 5th generation. The N64 somehow has the largest percentage of Sports titles in its library.

Genre – Genre Combinations

Since games can have any number of genres, let’s look at what Action appears alongside. Not too hard to apply Action to any other genre.

Adventure games come packaged in Compilations fairly often. Almost every Adventure game released episodically eventually had a Compiled edition. I can’t think of any Racing / Driving or Sports games that were also Adventure games. Let’s see, Barbie Horse Adventures: Riding Camp is listed as Adventure and Sports, Yakuza Kiwami 1 and 2 are too (with Action as well). That’s one thing they have in common.

As said before, lots of Adventure Compilations, but a surprising number of Puzzle games in Compilations too.

There are so few Educational games that we can expect a very different graph.

“What is a Puzzle and Sports game?”, you may be asking. A few include Pocket Card Jockey, Vertigo, and Clubhouse Games.

Surprisingly few Racing / Driving games are purely their own genre. Simulation seems like a natural pairing. Racing / Driving games that are also Adventure games include Nancy Drew: The Mystery of the Clue Bender Society and L.A. Noire.

Strategy/Tactics games are a natural fit for RPG mechanics, while driving a car isn’t. I just checked and Final Fantasy XV didn’t count as a Racing / Driving game for some reason.

Simulating playing a Sport or Racing a car make sense, while other genres probably have a side mode or minigame with some kind of Simulation.

Most Sports games don’t intermingle other genres into their gameplay.

Strategy/Tactics games are often about solving the puzzle of how to win, but few are also Puzzle games.

 

The above graphs in chart form. Here we can see the least combination of genres is Role-Playing and Educational, which makes good sense, these genres serve very different groups and tend to have very different budgets. There were only three such games: Fossil League: Dino Tournament Challenge, Bookworm Adventures, and Bookworm Adventures: Volume 2.

Here is an overall look at what percent of games are one genre. This is different than the previous “This Genre Alone” statistics because those were only considering the subset of games with a particular genre.

Genre and Metascore

The y-axis of this graph starts at 60 to make the differences more apparent, but keep in mind the range of values is 5.13. Action is by far the most common genre, but has the second-lowest average score. Compilation’s high scores could be influenced by the perceived value of a good cost to gameplay time ratio.

And here are the the genres of the best-reviewed games. Adventure, Puzzle, Simulation, and Strategy/Tactics all pretty low considering how many games of each there are.

That’s all for genres, next time it’s completion times.

Sources

GameRankings.com for metascores and some release dates. Archives: 1, 2, 3

MobyGames.com for genres, some release dates, some ESRB ratings, and some ESRB content descriptors

HowLongToBeat.com for main story and completionist times.

ESRB.org for ESRB ratings and ESRB content descriptors

Wikipedia for its many lists of games