The hierarchy of IFComp needs

In my IFComp research, after playing 20-30 games from each of the comps, I’ve noticed some patterns in what the players and judges are looking for, and I wanted to share those here.

I’ve noticed a sort of ‘hierarchy of needs’, which I’ve abstracted for the following list. The idea is that things at the beginning are so glaring that players won’t care about things later on the list. This is of course a gross over-simplification, but it can be useful.

The hierarchy is as follows:

Tier I-Basic competency

  • Level 1: The author needs to be trying. Games like Sisyphus or Toiletworld or The Absolute Worst IF Game in History end up last because the author is just trolling.
  • Level 2: All bugs should be eliminated. Guard Duty came in 2nd to last in 1999, even though it had complex characters and a rich world, because it had a game killing bug right at the start. In general, lots of big bugs will send your game to the bottom of the list. No one can get rid of all bugs, though. Emily Short won with Floatpoint in 2006 despite a big, noticeable bug.
  • Level 3: All typos should be eliminated. As Paul o Brian said about a 2001 game that placed almost last: “It’s too bad this game didn’t give out points every time I spotted an error, because if it did, I think I’d have earned 524,000 points out of a possible 200, earning me the rank of Gibbering Grammarian.” This is just one person, but it shows how some people are very opposed to any typos.

Tier II-Amount of content

  • Level 1: The game needs a lot of content. Well done but short games can still place high, but the top games always have a significant amount of content. I’ve tried to formalize this before as a specific amount of commands (around 150 typed commands or 200+ hyperlinks in a playthrough), but it’s best just to play other top games to see what’s expected.
  • Level 2: The content needs to be visible in some way. Final Exam, Pogoman Go!, Spy Intrigue and Baker of Shireton all had a significant amount of content that was hidden. Many people rated these games lower because they didn’t know there was anything else. Birdland solved this problem by dividing the game into different Days with a ramping up of tension, and crossed-out links showing what the player was missing; Detectiveland solved it by having 4 ‘Cases’; after completing one, you had an idea how long the others were. A lot of long games suffer from ‘flail around till something happens’ syndrome, where there’s no guidance as to what the player can do next. This is similar to the hidden content problem; players have no idea when the game will end. At least one author has said that this advice I’m giving is not good, so take it with a grain of salt.
  • Level 3: The content needs to not be repetitive. The House at The End of Rosewood Street is very cool, but requires you to knock at 8 doors and deliver 8 papers for 7 days in a row. It is extremely tedious; the game is otherwise very good. Your content is only fun if it isn’t being recycled over and over. That’s why procedurally generated games are so often boring.

Tier III-Creativity

  • Level 1: The game needs an original setting. Many of the middle zone of IFComp games are games that have trite and over-used settings. Dungeons and Dragons-type quests for gems and fighting wizards are almost always down here, including in the last few years. Office and home-exploring games are often down here too.
  • Level 2: The game needs interesting mechanics. You can get pretty far with just TAKE, DROP, LOOK, EXAMINE, etc. but a consistent set of original mechanics can get you far (like Morayati’s Take or Kwak’s How to Win at Paper Rock Scissors). The mechanic doesn’t have to be a gameplay one; Untold Riches had standard gameplay, but a fun mechanic where the narrator would tell funny stories about everything.
  • Level 3: The game needs good writing. This is hard, and I can’t really comment on this too much as I’m not that qualified. Fortunately, there is a wealth of help for this online.

I debated about the ordering of Tier II and Tier III; so don’t take this all too literally. But I think it helps me focus on my priorities in writing. Please feel free to comment if you disagree or have thoughts.

Notice that creative but short games like Take, Mirror and Queen, and The Queen’s Menagerie placed below longer games with more standard gameplay. Fair and Midnight Swordfight were short but with high replay value, so they still count as ‘a lot of content’.


I wish this forum had a Like button!

At Choice of Games, we clearly see a correlation between user reviews and total code length (which can either result in longer playthroughs or higher replay value; both seem to count equally).

The length issue is a hard pill for me to swallow. IFComp’s only requirement is that a game must be shorter than two hours. A game cannot be “too short” for the comp. But people often use those exact words, and penalize games for shortness as though it were against the official criteria.

A short game might be insubstantial, of course, but it might also be substantial. A short game might be more substantial than a longer one. Doesn’t matter, because it’ll lose points anyway for being short. I don’t think voters will change how they evaluate the games anytime soon, so it’s good to know this is what’s happening. It still doesn’t sit right with me.

I thought games can be more than 2 hours but the judging has a 2-hour time frame. The judge can play longer if he/she is enjoying it, but it must be judged no later than 2 hours of game play.


You’re right, that’s more accurate. A game can exceed two hours, but games are encouraged to be shorter so that judges can complete them. Many judges penalize games they can’t finish. At least this makes some sense, because the game didn’t deliver a full experience within the allotted time.

My issue is there’s no such threshold in the other direction. By official standards, a game can’t be too short. But by the voters’ standards, a game can be too short. It’s something you just have to live with. Since I like playing and writing short games, I probably care more than other people.

These two old quotes by Zarf are connected to the above:

Zarf: "One of the original reasons for having the IFComp be a short game competition, was that we felt that otherwise, any long entries would get all the attention. Short games would never win.

I still think this is true – other things (like writing quality) being equal."

Zarf: “It’s often valuable to signal to the player how long the game is, or how far they’ve gotten, so that their expectations aren’t out of line with reality.”

There’s another Zarf quote I can’t find that says that players don’t measure a games length by how long it takes to play, but by how many times you have to stop and think about something; so, for instance, long movement sequences across a map blur into a single action.

These are very good observations. While they shouldn’t chain authors to any kind of “winning formula” and people should always write how they want to, these parameters are good to keep in mind when conceiving a potential Comp entry.

My corollary impressions:

You’re very right about “hidden content”. While one of my favorite types of game experiences is that moment where you crack the ice and find an entire labyrinth that opens a game up beyond what it initially presents, the two-hour judging limit works against this type of delicious magic. If authors do this, they need to expect that a lot of time and effort they put in won’t be appreciated during the judging. The “surface experience” needs to stand strongly on its own to be successful in Comp terms.

While an extremely well-written “brief” game can do well, the judges tend to like optional depth. When they start the game and begin rowing across the lake, the opposite bank needs to be in view, but if they want to stop the boat and fish a while, that’s ideal. This is a little easier to do perhaps in parser games where emergent gameplay can happen, especially if the mechanics involve a cool system to play with.

In the best possible scenario, you want your judge to play and go “that was cool” within one or two playthroughs, plus give them enough promise of more writing or gameplay to entice them to continue after the two hour time limit, or at least put your game in the “come back to this later when I have time” pile.

Better to serve the judges a chewy, deliciously complex truffle instead of a 13x9 sheet cake with too much frosting and some sprinkles dashed on at the last second, unless that birthday cake is good enough to take home in a box for later.

If an author plans to create an intricate puzzle box or showcase a diabolic single puzzle or puzzle chain, it’s best to present an “arcade machine” type IF (Think Captain Verdeterre’s Plunder) where it is clear the puzzle is the focus of the game instead of framed by a more elaborate narrative that runs the risk of capturing the player’s attention and then stalling if the player gets stuck in the puzzle elements. This is an example of not allowing the opposite shore of the lake to be visible from the other side.

Authors need to keep in mind that IFComp mirrors the publishing-house slushpile model: If your submission does something to put off the judge (big ol’ typo on the first page, weird formatting), there are 40 other manuscripts waiting to impress and yours risks getting “thrown against the wall”. If the publishing submission metaphor doesn’t make sense, it’s like snacking on a large bowl of peanuts in the shell. If one peanut looks odd or is difficult to open, you’re more likely to just toss it and grab another one rather than put the effort into cracking the weird one to see if the inside is good.

Ever notice how TV sitcoms are 30 minutes and TV dramas are 1 hour? I wonder if shorter games would do better if they were comedic. I think it’s harder for a serious game to have much emotional impact if it just doesn’t last long enough, whereas a comedy can deliver good laughs right from the start. Mathbrush, does your data indicate that light/comedic games rate higher on average than other short games?

I think lighter comedic games have a lesser chance of putting people off. Everyone likes to laugh. Not everyone wants to read a heavy story.

That said, I usually prefer comedy, but I found myself very moved by Ash and it was playable within 30 minutes.

My hunch is that comedy parser games have an easier time in general because players always try to mess with the parser. That “lemme poke it” attitude leads naturally to comedy. Serious games, however, have to overcome it.

Length seems like a separate issue. There are many serious traditional short stories that work very well. But people expect them to be short. It’s in the name. People expect IF to be longer, probably still due to standards from Infocom’s era.

I see the same thing in indie game festivals that have nothing to do with Infocom’s era. Very short games come off as lightweight or “appetizers”, even if they’re very well constructed.

Maybe that means it’s an expectation for games period then, and not just text games. It’s still something I find disconcerting, since short games don’t come off as lightweight appetizers to me if they’re done well. No more than short stories would. But I see I may be in a minority of one. I won’t bother anyone else about it again.

No, I agree with you, and I think zarf would too, e.g. Shade.

So, who wants to run ShortGameComp?

Oo! Then there could be a trophy called “The Emily” (like “The Oscar”).

…because her last name is… Oh. I’ll just be over here…

I’ve consciously designed my ifdb star system to be length-independent. I love a lot of the one-move Fingertips games.

Long games are regarded better, but short games get played more some times. Shade, Galatea, 9:05, and Aisle have been leads more than, say, Blue Lacuna.