Some interesting (?) stats on IFDB in past decade

tl;dr I realized while I wrote this that it was getting more and more boring and I’m probably the only interested in this, so I’ll summarise: people are making more games than ever and reviewers are more active recently than they were in the past. Both choice and parser formats are healthy in growth.

IFDB is what drew me into IF. In 2010, I bought an iPad and vaguely remembered seeing Zork as a kid, so I searched for it and found Frotz, and with Frotz, I found IFDB.

I thought at the time that IF was a ‘dead’ field. I stopped playing (Vespers and Varicella freaked me out too much), and didn’t check IF again until 2015 when I redownloaded it and tried again.

I noticed around 2015 that IFDB seemed to ‘peak’ around games from 2000-2008, which had hundreds of ratings. I felt like it was just in slow decline.

Now, in 2020, things are very different!

For instance, compare these three links:

From 2005-2009:
https://ifdb.tads.org/search?sortby=rcu&newSortBy.x=0&newSortBy.y=0&searchfor=published%3A2005-2009
From 2010-2014:
https://ifdb.tads.org/search?sortby=rcu&newSortBy.x=0&newSortBy.y=0&searchfor=published%3A2010-2014
From 2015-2019:
https://ifdb.tads.org/search?searchfor=published%3A2015-2019&searchgo=Search+Games

The expectation with ratings (with people slowly playing older games less) is that, if things are dying off, old games will have way more reviews overall than new games; if things are steady, that there should be a big step over a few years, and that if things are increasing then new stuff should have about as much or more activity than in the past.

From 2005-2009, there are a few absolute standout games (Lost Pig, Violet, and Bronze) with over 200 ratings, and several other games that are very popular (like De Baron, Blue Lacuna and Gun Mute).

Overall, only 870 games are listed, and the bottom of the page (of 20 items) has games with ratings in the 60’s and 70’s.

I think these few games with high ratings happened because IFDB started around Oct 1, 2007, and people felt obligated to help fill up ratings for games, since there were none at first. So games that people really remembered (like that year’s IFComp winner and the next) got unusually high numbers of ratings.

From 2010-2014, there were 1857 games, a far higher number. This includes the ‘first wave’ of Twine games. Of the top 20 most-rated games, exactly half are choice-based. The most-rated game is Counterfeit Monkey, with 136 ratings, and the 20th most-rated game has 53 ratings, not too far off from the 2005-2009 range, despite the fact that less time had passed.

From 2015-2019, there were 2241 games (including some pretty big comps like Twiny Jam). The highest rated game, Birdland, has 115 ratings, and the 20th most-rated has 40, again not too far off from the set 5 years earlier, showing that there’s a lot of ratings going on recently. Out of the top 20, 10 are parser, 8 are choice-based, 1 is a parser in choice-based format and the other is choice-based in parser format.

11 Likes

If I get what your numbers are telling me, there’s definitely new life brought to the medium thanks to choice based and hypertext games. But if we’re are to assume that those games make up roughly half of the total, then those numbers say that the amount of new parser games created each year is also still increasing.

Is that right? That sounds really promising if so.

1 Like

There is a representation bias. I’d interpret this as IFDB getting more popular as a website, so people are adding more games on off-comp seasons.

For example, the “first wave” of Twine games is barely present there.

3 Likes

This topic is interesting.

An “event” which had some impact on the numbers (but not on the conclusions) is that there came a huge number of Quest games on IFDB from 2010-2014. During this period, 25% of all new games were Quest games. I think I read somewhere, that it was because that at some point all people’s Quest games were automatically added to IFDB.
More specifically:
2010: 11 Quest games
2011: 28 Quest games
2012: 124 Quest games
2013: 280 Quest games
2014: 32 Quest games

3 Likes

That’s very interesting!

If we restrict to games that people ‘care about’, we can get rid of a lot of automatically generated games. One measure of ‘caring about’ is having at least 5 ratings. Then the numbers are:

2005-2009:
281
2010-2014:
644
2015-2019:
690

Having at least 1 review might be a different measure. But I’ve written about 1/4 of all reviews, so to remove my personal interests from the equation, I’ll do 2 reviews:

2005-2009:
225
2010-2014:
495
2015-2019:
557

And again, we’d expect all these numbers to go down a bit over time if growth was stagnant, since older games should be accumulating more ratings and reviews.

4 Likes

This is true and was due to a misunderstanding. I’m sure there are legacy threads here about it where everyone was complaining that Quest games were wiping all the new releases off the IFDB front page which usually displays the most recent five or ten. Which would have been no problem if they were complete, legitimate games.

Alex Warren, with all the best intentions, had set the textadventures.co.uk site to automatically create an IFDB listing for new Quest games without realizing it was going to grab every test game and random minor project that was created from people experimenting with Quest instead of only finished games people intended to make public.

4 Likes

I’m about to make a point about a very different era but it still may help illustrate why there are only so many firm conclusions you can make from aggregated IFDB data.

I know I was recently reading a discussion in which someone made a remark along the lines of “According to IFDB there were only 180 commercial parser games in the original 70s/80s era” (I forget the exact number but I believe that was the ballpark) and I knew that couldn’t be correct.

And indeed, it’s not. I’ve been adding plenty of games from that period which weren’t at all obscure, they were from as-legit-as-it-got commercial publishers of the early period. All I’m doing is going through major hint compendiums and adding games as I find them missing. And there are plenty.

6 Likes

I guess that the IFDB data from the commercial era (70s/80s) is one of the few areas that are far from complete. I was about to say that IFDB is probably pretty up to date with most of the modern IF community since 2007 but then I came to think of www.textadventures.co.uk

The highest rated game is “Victorian Detective”. On texadventures.co.uk is has 132 ratings but only 11 on IFDB. And none of the four following games in the “Victorian Detective”-series have been added to IFDB yet. This shows that there are lots of IF people who do not use IFDB actively.

I suppose there are lots of significant sub-groups of the IF community, which do not use IFDB or intfiction.org though some of us(?) regard these as the center of the IF community. Still, Mathbrush’s conclusions seem to hold as the number of “cared about”-games seems to go up.

3 Likes

I’ll counter this with “It’s easy to visit the IFDB main page, see the release of a new game, realize it will take three minutes to play and two minutes to review, and then spend the next five minutes doing that.”

(IFDB discovery also rewards those “cared-about” games which have activity as well, yes, and some older games get new entries, but…)

2 Likes

I suppose there are lots of significant sub-groups of the IF community, which do not use IFDB or intfiction.org though some of us(?) regard these as the center of the IF community. Still, Mathbrush’s conclusions seem to hold as the number of “cared about”-games seems to go up.

I got into IF through IFDB, and the one thing that has appealed most to me over the year is that I can generally keep on top of most IFDB/intfiction stuff and not get overwhelmed. I can’t keep up with buying Choice of Games, I’ve only played Choices/Episode for research for a contract I had, Philome.la was too much for me when it was around (and textadventures.co.uk is too much now), and retro gaming generally has games that take a long time to play.

So I stuck to IFComp/IFDB/XYZZY/Spring Thing. It’s not ‘the world’, but it is ‘my world’. Just about nothing means more to me than finding someone added a rating to one of my games. I check just about every day lol.

I’ll counter this with “It’s easy to visit the IFDB main page, see the release of a new game, realize it will take three minutes to play and two minutes to review, and then spend the next five minutes doing that.”

I agree completely. But every old game was once new, and would have received the same treatment then, right? Most games follow a pattern of a big burst of likes followed by a slow trail, and that trail is what I’m arguing makes older games usually get more votes. I think the difference is that when I say ‘older game’ I’m thinking like 2009 while most of the games you work with are from 70’s/80’s, long before IFDB was created, so I think in that case your arguments are completely correct.

1 Like

And I don’t have the data on this (legit question: you might, do you?) but it sure seems to me like a “median gameplay time by year released” graph plotted with “year” on the X axis and “minutes” on the Y has a significant downward slope even over 2005-2020. (A short game can still prompt deep thoughts that take months or years to process, blah blah, but it’s also just a lot easier to find three-minute games now than it once was.)

3 Likes

I have the data but can’t organize it without downloading IFDB and using SQL. I started classifying every game I play by gameplay length (only the most recent 1000 reviews), with the categories:

less than 15 minutes
15-30 minutes
about 1 hour
about 2 hours
2-10 hours
more than 10 hours

You can see them here:
https://ifdb.tads.org/allreviews?sortby=new&newSortBy.x=0&newSortBy.y=0&id=nufzrftl37o9rw5t&ratings=&tag=

But it doesn’t display the game’s publication date.

Looking through the < 15 minutes tag, the majority are from the last few years. Looking through the > 10 hours tag, it looks like most of them are quite a bit older, usually 80’s-90’s. The only parser games in the last few years that are > 10 hours for ‘good reasons’ that I’ve played are Cragne Manor, Hadean Lands, and Worldsmith. So I think that correlates with what you’re saying.

4 Likes

This just shows that Quest has its own insular community that doesn’t much venture outside textadventures.co.uk - which is really cool because that’s what they like. This tends to occur in IF communities centered around a tool with their own specific forums and distribution/rating systems. I think that was Alex’s intention in making Quest games auto-list on IFDB since Quest is a legitimate venue for IF.

There are many other separate communities. Choice of Games has a zillionfold of our activity since there are many who just play and write ChoiceScript. We’ve brushed against the tiny but loyal Choose Your Story community when someone there discovered the comps and entered. They have their own unique online-only tool and that’s all most of them are interested in. Many of us think of Visual Novels as IF even though nobody discusses them here nor actively lists them on IFDB - there’s a whole archipelago of Ren’py/VNMaker communities out there somewhere with their own forums and game services who don’t need nor would be successful trawling IFDB. It’s not the center of everyone’s IF universe. Some of us stay local and won’t cross the ocean in any interpreter that won’t play Zcode, some of us hop on the HTML cruise lines and venture to other places!

I like to think the IF community overall is more of a European Union as opposed to a Borg - needing to assimilate the entire universe and conform all to our philosophy and technology. Not everyone natively thinks of themselves as “an IF person”, but they have their own customs, game rituals, and reading habits.

7 Likes

Any analysis of IFDB is interesting, but ultimately flawed.

There are whole active communities of parser “interactive fiction” creators that don’t really engage or use IFDB; particularly those making text/conversational adventures for retro platforms, and the huge non-English communities (such as the Spanish). The historical database on IFDB for those sorts of text adventures also has plenty of gaps.

2 Likes