Some patterns in which games do well in IFComp and IFDB

I’ve been playing through many IFComp games from past years, and games from the IFDB top 100. I’ve noticed some general patterns that surprised me. I posted this on Euphoria chat, but I wanted to post it here too so I could access it later.

So here is my final conclusion on what makes a game do well in IFComp and IFDB ratings (such as IFDB top 100). There are three major factors. Some of this will be pretty obvious and vague.

What makes games get high scores?

The game should not be buggy

Buggy games plummet to the bottom of the list. For large complex games, there seems to be forgiveness for a few small bugs. Unimplemented items that should be implemented and typos count as well.

The game needs to seem and be substantial

Substantial games can be quantified as needing around 100 or more moves to fully complete. This has to be communicated to the player. So games where the size of the game is not obvious suffer. The Play and Midnight Swordfight clearly indicate that there are multiple endings. Six Stories (coming in 3rd in IFComp one year) was an anomaly, but contains six large stories, making it substantial in amount of text.

The game needs not to frustrate the player

Games with tedious elements such as slow typing, very repetitive actions, etc. tend to place lower (foraging for food or trying to raise money doing the same thing over and over again.). Also, top games always indicate a path forward for the player, something to try next. Long web-based games without save functions get docked.

If anyone would like examples and/or statistics, let me know. I’d be interested in seeing counterexamples or contrasting opinions (some of this has already been changed a bit based on feedback from Chris Klimas).

I will add that every game that satisfies these three requirement (in my subjective view) had an average score of 7 or higher in the ifcomp, and the top 3 games in each ifcomp (except perhaps the first) satisfied these requirements.

Thanks for sharing. I imagine you probably played the relevant games closer together than most people have.

You said some patterns surprised you, but the 3 main factors you list don’t seem to be surprising ones. Are there any patterns you didn’t mention, maybe because you’d put them a bit lower (eg in positions 4,5)? Or did the 3 main factors surprise you?


Patterns I’ve noticed:

  1. Games entered in the IFComp tend to do much better in the overall voting than games that weren’t entered.

  2. Games with a high score tend to do better than games with a low score.

  3. If a game throws an error message which causes your computer to blow up and maim you, this game is far less likely to do well than a game which plays smoothly.

I think the point is not that it’s surprising these things are important, but that they appear to come before things like good writing and plotting (however you want to measure those - I think craiglocke looked at Xyzzy nominations for that sort of data.)

@David Whyld haha, your 3rd item is the reason A New Day took 10th place, I guess.

@Wade The real pattern that I noticed is that long games that were bug free got high scores regardless of content. You could write essentially any crappy game in the world that was long and didn’t crash and get high up votes. But there were a few games that were long and not buggy that scored pretty low (maybe 6-9 games in all of IFComp). All of those games turned out to have very repetitive experiences (like farming money or experience, navigating a maze, etc.), or were difficult to install.

So the real surprise is that people seem to be voting mainly on length. Having time caves or secret content or ‘hey, after you play for a bit, the game really changes and you realize it’s bigger than you thought!’ was uniformly a negative in terms of scores.

Also, I don’t believe that these things make games ‘better’. Shade, the Gostak, Rameses, Delightful Wallpaper, etc. all violate one of these rules, all placed relatively low, and are all great.

p.s. I forgot, the 4th item that I didn’t mention was grossing people out/putting people off with content. This involves mostly Porpentine, some melodramatic Christian games, and gruesome horror games. It has an effect, bt not as big as it seems; for instance, Vespers won.

Oh I see.

This doesn’t surprise me so long as bugs continue to be a substantial issue for a sufficient number of IFComp games. A game’s content can only be delivered if it’s not blocked by bugs, or by the game itself in a way the player doesn’t find stimulating (this covers things like gruel, too much frustration, repetition). I suspect this figures for the majority of the history of the IFComp to date, so will be relevant in your (craiglocke’s)survey of top 3 games.

If a game can deliver its content over duration, that tends to be more engaging to me than a shorter or buggier game, and I know I’ll tend to rate it higher than the shorter/buggier. Then I’m rating the games that engaged me similarly against each other. So length is definitely an organising factor for me. I’d say IFComp has never handled this well (eg at the Oscars, they don’t put features and shorts in the same category), but I’d also say it remains that super short games aren’t thick on the ground in the IFComp. There are also lots of alternate venues available for especially short games now.


Without any further data to support this conclusion, I’m afraid it might be a good old “all dogs have a tail, therefore if something has a tail, it is a dog” fallacy. Did you actually find any such games?

To me it seems much more likely that because making a long and bug free game requires quite a lot of time and effort, it’s probable that a lot of time and effort was put into other aspects of the game, too.

@Juhana Here’s an example: The Erudition chamber. It took 4th place, with an average score of 7.32.

Reviews include “The Erudition Chamber falls short on these counts. In fact, the game is seriously hobbled by the prose that comprises it. The author repeatedly misuses commas, misspells words, and gets verb forms wrong. (Adam Cadre)” (bad prose)

“Troubled prose… always weakens a game for me, and it’s a pity, because this game is pretty strong in lots of other areas. I found no bugs, which always pleases me, especially in a comp game.” (not buggy, but bad writing)

“It was compulsive in a self-assessing way because you just had to find out which sect fitted you best. I also loved the way doors disappeared behind you as new areas were entered. I was left in no doubt that I should go forward with the equipment in my current inventory. No fifty-move treks across the map to retrieve an essential object discarded earlier. (Virginia Gretton)” (so, not frustrating)

"The writing is not as interesting. The story, such as it is, is all about being tested. It feels pretty artificial, both in the idea of setting up this test in the first place and in the lore that goes with the various puzzle-solving styles. " (bad writing)

Here’s another example: An Act of Murder. It was a very well-received game (and I like it too), and the writing was not bad, but what do the reviews mention?

"The characters have some interesting dialog, but they are a bit static and lifeless for a game (and a genre) that is heavily driven by NPC interactions.

The author has also gone to some effort to make the game easy to play, for which I’m always grateful. Simple touches (like listing where the exits of each room go) free up brainpower for use on solving the mystery rather than fighting with the game. A nice many-layered hint system ensures that you won’t be stuck for long.

AAOM is not an epic or elaborate game by any means, but as a quick, engrossing whodunnit with no obvious bugs or holes, it does a fine job. "

"'ll first start with the one thing I especially disliked. I felt that the characters had no feeling at all that the murder happened. They sat around playing pool like usual or reading the newest mystery novel, and not once did anybody show the slightest remorse that there had been a murder. I mean sure, some of them should have been uncaring, but a story should also have some who are Teary eyed. This in my opinion really took away from the characters.

Of course I thought most everything else was well done. I’ve played mystery adventures before which are brutally tough because of time constraints and obscure evidence, but both in this case were good. An adequate amount of time was given, and evidence took some thought, but was not too difficult. It is a game about your next move, meaning that you have to think ‘what next’ after everything you do.

I also loved the notebook. It made my life so much easier. "

“The writing is spare, but universally excellent and there are almost no typographical errors. The descriptions are not so filled with red herrings so that you go off chasing the wrong ideas, but do include just a few irrelevancies to turn you about here and there. The simple deduction puzzle that is at the heart of this game won’t overly frustrate one, but still keeps one engaged. Technically speaking, the use of an inspector’s notebook is helpful and clever and dispenses with the need for paper notes, but the notebook can inadvertently reveal the names of some objects that one has not yet unveiled through natural game play.”

So some people thought the writing was okay, some didn’t, but people universally praised the game for not being frustrating, and for not being buggy. (Others have mentioned that the writing had to be spare to allow for different suspects to be the killer).

The same author, a very talented writer, wrote the beautiful “Sunday Afternoon” in 2012, which was shorter and more frustrating (some people couldn’t get out of the chair). It came in 5th, compared to 2nd for An Act of Murder.

In other words, the voters stubbornly continue to evaluate IF games primarily as games instead of stories.

I don’t think that finding good writing important in a writing-based game means you’re evaluating it as a story rather than a game (putting aside the dodginess of that dichotomy.) If someone remarks on the graphics in a graphical game, are they evaluating it as a TV programme?

Not “finding good writing important”, but finding it more important than whether the game succeeds as a game. I should mention that I’ve played four of your games and all of them, while they have strong writing, are good games. If they were broken, or more linear, or had other game-related faults, the writing wouldn’t redeem them as games.

That parallel falls down in a couple of ways. (When I talk about TV programs, I talk about both the writing and the visual design. It doesn’t represent a one-side-of-the-coin judgement.)

The comparison here is games that are cited for good writing but poor mechanics, and games that are cited for poor writing but good mechanics. That’s a meaningful comparison. If you’re entering IFComp, you want to know how people react to those two cases!

(Obviously you’d love to create a game with superlative writing and perfect code. But, in the real world, you often wind up asking “I have a week left before the deadline. How important is fixing this bug right now?”)

It’s been really interesting hearing everybody’s thoughts. (And I agree that Robin’s games have been really great).

One thing I wanted to point out is puzzle design by itself doesn’t really seem to factor in independently from substance or frustration. So a game that had a lot of interesting, puzzle free content could do well, and you see games like that with Map, Creatures Such as We, and Exhibition.

I think that any sufficiently interactive game has an implicit puzzle of “what can I make this do?” that sustains engagement.

That’s true; even Exhibition had the puzzle of “What was the artist really like? What really happened?”

Yeah, although that particular question could apply to static fiction as well. I think the difference is that something is required of the player to discover the answer other than merely turning the (virtual) page.

Edit: I think that a novel with an unreliable narrator could qualify as a puzzle by this “more than mere page-turning” definition, but not as an interactive one.

Aw, thanks. I should put that on my business cards.

In response to some of the curated comments on An Act of Murder, I guess some of the emotional response kind of fell by the wayside. Generally, I think people confronted by a murder would probably be too shocked by the fact of “violent crime” to register “loved one is dead”, but I guess I left out all hints that this might be the case with the characters. I think I did have an idea that one character, seen reading a book for duration of the story, should have been staring at the same page for all that time without actually registering a word; but I guess I dropped that in favour of a “pot-kettle-black” observation about pseudonymous writers naming their heroes after themselves or vice-versa.

There’s something that came out in a comp review for another game of mine: apparently the size of a paragraph in parser IF matters, at least subconsciously. About three to six lines seems to be the ideal, I think.

This is all part of the principle of making the player’s experience a smooth one. Clunky writing is like a mess of rocks strewn across your path: it can be annoying, and if it’s really bad you might stub your toe, but it won’t halt your progress. Comparatively speaking, it isn’t actually all that frustrating. Buggy code is more like a rockslide: it can actively stop you from seeing anything more than the few steps you’ve already taken. That’s massively frustrating. In the end, it really all just comes down to two things: frustration and boredom. Everything, from writing to coding to plotting to game design, is directed towards managing those two beasts. And if you manage to neither frustrate nor bore your player, you’ve got it made.

Ooh, thanks for your advice about paragraph size. I’ve really been struggling with that in my current projects! I’ve only had a sentence or two in many rooms and it seemed sparse.