@mathbrush Apparently I’m even less positive than you!
I guess this does line up with my ratings for IFComp: I’ve given 7s and 8s to a lot of my favorite Comp games from the past five years, and I don’t remember ever rating one of the entries a 10.
@mathbrush Apparently I’m even less positive than you!
I guess this does line up with my ratings for IFComp: I’ve given 7s and 8s to a lot of my favorite Comp games from the past five years, and I don’t remember ever rating one of the entries a 10.
One last spreadsheet: something like IFDB top 100, but for author’s entire oeuvre.
This one is kind of scuffed because it includes voter manipulation that has later been taken off the main page, and also because it just sorts by what’s in the name field of a game, which can contain multiple names or can have different spellings of a single author’s name. So this would take a lot of work to make more usable.
(Note that the lowest author in the rankings is the author of the game called The Absolute Worst IF Game in History).
So that is “top authors based on vote averages”, with the next three numerical columns being “mean”, “number of ratings”, and “Bayesian average”?
Yeah, I’ll add that. This one is very problematic and not rigorous due to its structure, but there’s no quick fix outside of manually fixing thousands of entries
I think that can be really helpful as long as your consistent (which you are). If I get a 4 from someone and I click their page and see they never give out 5’s, that feels like a real honor!
I am in the 1% most negative and 5% most helpful. I think this is because I don’t tend to play games that are the most popular or the highest rated. From a review perspective, adding “this is a masterpiece and I loved it so much” to a game that already has 10 or so analyses on it that break it down better than I could feels a bit useless. I don’t avoid these games, there are plenty that I’ve played and liked, but it’s only once in a while.
I mainly choose games to review based on ones that have no detailed reviews already, and I enjoy reading the reviews of more obscure games myself. You never really know what to expect from these, and I’ve come across a lot that are really fun and well-implemented (Ghost, Wry, Top Floor Please, Colour Beyond Time). The amount of actual bad games is pretty low.
Out of curiosity, I actually ranked all 80 games I’ve reviewed by best to worst to investigate some trends. There were only about 8 games I outright disliked and would not play again. Similarly, my ratings between platforms are pretty consistent: both my top 10 and my bottom 10 consist of eight parser games and two choice-based ones (Bee and One Does Not Simply Fry for the former, MARIO CARDS and Textos Blast from the Past for the latter). The medians of the list are The Little Match Girl, by Hans Christian Andersen and Colour Beyond Time, which I both generally liked: they’d be around a 6.5/10, if I had to grade it that way.
And I also found out that I left a lot of one-star ratings (not reviews) for games that I would read the reviews of, and see that they were either unplayably buggy or just low-effort troll games. I don’t remember when I did this, but if those are counted in the average, it’s probably part of why it’s so low. I’ll go through and remove them, there’s not much of a point to keeping them there.
I don’t know if you’re referring to Goodreads in there? But it allows me to comment on Goodreads.
I use Goodreads to record all I’ve read. I otherwise find it quite yuck. It has a clubby feel of people high-fiving and piling on each other’s catty reviews. The primacy of the culture seems to be clicking to follow other users, then reading and commenting on the reviews of those you follow.
I noted to a Goodreads enthusiast friend my handful of isolated reviews on the site had zero reactions or feedback. This is a situation unlike with reviews I have on other sites (IFDB where I make solid efforts, IMDB where I register all films I have seen and infrequently write a review, or gamefaqs where I used to review a lot, a long time ago.) My friend said ‘You need to start following people, then you’ll start getting reactions.’
I have no interest in following people on Goodreads or soliciting reactions there. One of my reviews is for a Thomas Harris book, so you think someone would have seen it. The other is for a book of the type where culty teenagers would attack you if you didn’t like it, and I didn’t. That neither has a single comment or like suggests to me they simply are never read or seen.
What I like on IMDB is the maelstrom. People are plainly reading the reviews of people they know nothing about all the time!
-Wade
I don’t have an account on goodreads or look at it, fwiw. I am just speaking as someone who has limited time and has found a way to encounter enjoyable content pretty reliably. I don’t use negative reviews to find games I enjoy, and it’s going well.
But that’s what happens with a big, complex aggregate of content. People do whatever works for them. That’s great! If I reject anything, it’s the idea that there’s something somebody “ought” to be doing regarding reviews and scores. Or not. Anything in good faith is fine.
Goodreads is neither fish, flesh, nor fowl: A catalog of books and reviews on one hand, a social network on the other, and its Reddit-like forums tacked on almost like an afterthought. I find it a hard place to dig into.
I’m mostly on it to see when an author I like has a new book and to keep track of things I want to read later. Following a user for their reviews had not occurred to me as a thing anybody would do. Most of the reviewers spend way too long recapping the story, anyway.
At least there are no video reviews. Those are a blight on BGG.
Hey, how about that? Thanks for the graphs, mathbrush. It certainly doesn’t look like much of an upslope, if it’s there at all.
EDIT: Monthly averages vary quite a bit, but the trend at a yearly level seems to exist.
I made my own graph. (Data for 2024 is based on the first eight months of the year only.)
This is one of those graphs that cheats a bit – emphasizing the range 3 to 4 (where averages actually occur) to magnify the trend visually. Still, despite variation, the trend line for ratings of Inform and Twine games is clearly up over time. A change in the average of 0.4 (as seen for Inform games) translates to nearly half a star.
I’ve also made a list of reviews that have the most votes of one type (i.e. most upvotes or most downvotes):
Game name | Vote count | Helpful |
---|---|---|
Lost Pig | 47 | Y |
Galatea | 37 | N |
Violet | 35 | N |
Photopia | 34 | Y |
rendition | 34 | N |
Everybody Dies | 33 | N |
Galatea | 31 | Y |
Anchorhead | 31 | Y |
Violet | 30 | Y |
The Game Formerly Known as Hidden Nazi Mode | 30 | Y |
The King of Shreds and Patches | 29 | Y |
Varicella | 28 | Y |
Treasures of a Slaver’s Kingdom | 28 | N |
Aisle | 28 | Y |
Shade | 28 | N |
Cragne Manor | 28 | Y |
9:05 | 27 | Y |
Zork I | 27 | Y |
Plundered Hearts | 26 | Y |
Lost Pig | 26 | Y |
Bronze | 25 | Y |
Photopia | 25 | Y |
All Roads | 25 | N |
Kerkerkruip | 25 | Y |
The Wizard Sniffer | 25 | Y |
(so Lost Pig has a review with 47 helpful upvotes)
So I checked out that most-downvoted Galatea review, and now you need to update your chart.
I’m pretty sure that’s because that review was written by kids and people thought it was cute.
That HNM review has more votes than the game has ratings! Nearly as many votes as players! Which is perfect in a way. (Yes, I am weirdly fascinated by The Game Formerly Known as Hidden Nazi Mode)
I am in favor of positive reviews and constructive criticism. They help motivate game makers and help them make better games, and keep you off people’s enemy lists.
I also am in favor of some threads reviewing the reviewers. How else are they going to improve their reviews and make them more thoughtful and helpful?
Back in the day on the newsgroups, Mike Sousa reviewed the Comp reviewers a couple times, which was lots of fun. @mathbrush ran a competition for reviewers in 2020, too, and I wrote a review of @VictorGijsbers’ reviews as a prize for that, so I enjoyed that too - I’m definitely in favor of more of this sort of thing!
while it hardly measures up to writing reviews, this conversation did lead me to try and vote at least one review helpful per day. so far so good!
For what it’s worth, this makes sense to me! I suspect most raters in IFDB have played only games that they thought they were likely to enjoy, so were likely to rank them highly. (And personally if I don’t like a game, I will probably never finish nor rate it.) On the other hand, you have played a very large number of games of all levels of quality, including presumably a bunch of games that are just not that good.
It would be kind of interesting to calculate the “bias” of a reviewer – say, for any particular reviewer, calculate the average difference between the review score and the overall score for each game they reviewed. (So colloquially, we’d say a positive reviewer who tends to like games more than the average person who’s played them.)
I’ve been trying to do that for about an hour now, but I’m having some trouble with mariadb/SQL. If I ever figure it out I’ll let you know!