Various measurements of the "size" of a game

Some recent threads, as well as my own experiences writing my first game, have me thinking of various ways to have an impression of an IF’s “size.”

This can be subjective, and I want to embrace that! There are several different axes by which it could be measured. From a player POV, it may involve depth of implementation, time to beat, or minimum number of moves to beat.

From a development POV, it might involve amount of code, amount of printed text (hard to measure in I7, I think), number of objects and so forth. This probably varies by system and may imply something like “effort.”

What would you consider a “big” game? In what senses is a game “large” for a player? For an author? Note that “large” does not equal “good.” By many measurements, Zork Zero is Infocom’s “largest” game, and I really, really, really dislike it.

Points of departure:


In the early days, the number of “locations” was the standard measurement of the size of a game. Which led to a run of games which boasted hundreds or thousands of rooms… most of which were padding locations, described the same.

Measurements could be “word count”, number of locations, or number of puzzles. Length of time to complete is a possibility but could be too easily skewed by difficulty.


I know “big” games that consist of many rooms (Emily Short’s Bronze). A few of them, (Aaron Reed’s Blue Lacuna) use word count to measure the game’s size.

I think I’d consider them both big games, but I think the real measure of it is the time it takes to complete a game and discover all its secrets.


Yes! That’s another interesting consideration: “size” as a promotional characteristic. I seem to recall some games boasting about their total number of points, too.

I think this is what is most often meant by players, though as Strident says above, difficulty muddies things. Infocom’s (sorry to keep mentioning Infocom, that’s where my expertise lies) Deadline took me years to beat, I think. Nearly all of that time consisted of waiting for inspiration.


I don’t think there is one measurement that could help determine whether a game is big or not. Depending on the type of game (choice-based/parser/etc), some measurements will apply to a game better than others. I’d say word count would probably measure better choice-based IF than parsers (or other type of IF), while locations/puzzles would fit better parsers.

I know for choice-based games, the comparison of word counts between a single run and the total amount of word in the game is quite popular (esp in CoG). A game might not feel large, but because of the variation (and the replay value) players might consider them big.
Could a replay value or variation option could work as a measurement for parsers though?

Also agreeing on the time spent. At least regarding choice-based games, time spent on a game will depend on the reading speed of the player.


Here a couple I can think of, including stuff you guys have said:

Word Count:
This is a very easy measure for games, and is the main selling point of Choice of Games games (longer games get charged more for and are more popular).

For choice-based games, this usually directly correlates to longer games, but not necessarily for parser games. Pogoman GO! has over 100K in words (which is really unusual), but a lot of that is based on special responses to standard verbs and randomized text, and the whole game can be completed in a few hours.

So Pogoman Go is an example of a medium-sized game with tons of text, while an example in the other direction might be Plotkin’s Praser 5. I don’t know the exact wordcount on it, but I suspect its fairly low, as it just consists of incredibly difficult puzzles, each of which can be solved in a compact fashion.

Another small-code but giant game is Starcross. According to this post, Starcross only has 15K printable words in the code (which if it’s similar to Zarf’s estimates, would be around 45K word of codes), and Starcross is one of the Infocom games I consider really big.

Time to complete

You mentioned this one, and I agree. Time to complete can make a game really feel big, but not always.

SNOSAE is a game that takes a very long time to complete. It’s in DOSBox, and it’s just one series of very hard puzzles after another with little hinting or cluing. If someone had a perfect walkthrough, it would be fairly short.

On the other hand, a game can feel smaller than its time to complete suggests. The House at the End of Rosewood Street takes a lot of actions to complete, but most of them are completely repetitive (delivering mail to 8 different houses on a street for 7 days in a row). So it’s a 2-3 hour game that feels smaller.

A more positive example is Choice of Robots, one of my favorite choicescript games. While it is in fact very large (300K words) it feels even larger because its immensely replayable and has a lot to find in replays. Like, one playthrough is about saving your parents with cancer-curing doctor robots and another is curbstomping Alaska with giant death mechs.

Number of distinct rooms/scenes/interactions

I think this is probably the major indicator of size for me. But there are for sure counterexamples here, like that game Snowball with 7000 identical locations.

Even among games with different locations, it can feel like not that much content. One example is The Northnorth Passage, which has very lush and descriptive scenes in a variety of rooms but is completely trivial to solve. Another similar examples is Toby’s Nose, which has very dense room descriptions and a variety of locations but is relatively brief due to the constrained actions.

Why not all three?

The games that have all three (tons of text, takes a while to complete, distinct locations) are the ones I consider truly large. The problem is many of them just become a blur of memory and it’s very hard to know their actual size in comparison to each other. Here are some:

  1. Finding Martin. I suspect this is the largest parser game I have ever played. It includes stuff like a zoo with penguins, time travel repeated multiple times, a phone booth that teleports you around, flavor-based magic, etc. Very difficult, high wordcount, etc.
  2. Cragne Manor, which with 80 authors its size shouldn’t really be a surprise.
  3. Inside Woman. Andy Phillips is known for giant games filled with catsuit-wearing female assassins and this is probably his biggest game, where you have to fight your way through an immense skyscraper with different levels (like the Shinra tower in FFVII) and uncover its dark history while fighting some symbolic number of female assassins (12? 7?).
  4. The Mulldoon Legacy is one of Jon Ingold’s early games, and has a lot of portal-based time travel, lots of unusual mazes, many codes, etc.
  5. Blue Lacuna is bigger in code size and locations than many of the above, but is purposely designed to have easier puzzles, so will be shorter than others when not using a walkthrough and vaguely similar when using a walkthrough.
  6. Worlds Apart is more story-focused and I swear it’s somewhat incomplete (like maybe the author teased a sequel or more content? Haven’t played it in years despite how much I liked it), but this is very expansive. Very Anne Mcaffrey-esque, which a protagonist singing and communicating with dolphins.
  7. Lydia’s Heart is Jim Aikin’s biggest game (I think? I haven’t finished Prom Dress). A very large horror game set in the swamps of the American south. Not Just An Ordinary Ballerina was one of the very first games I played when I discovered IF in 2010, and that was huge; this game is even bigger.

I didn’t include Hadean Lands and Counterfeit Monkey because in a weird way, despite both being monstrously large, they go out of their way to feel smaller by providing many conveniences. Hadean Lands would be far larger without Plotkin’s handy shortcuts, and Counterfeit Monkey is designed to be completed. So I think they provide the same kind of content as the bigger games, but in a more player-friendly and time-friendly way. It’s like all the good stuff, without any of the filler.

Curses is big, and it’s my number one favorite game, but I don’t think it cracks the above list.

You linked a post of mine above; I’m trying to create a truly large game. I have a list of Games I’ve reviewed that are over 10 hours long, and there aren’t a ton of games on the list, so I want to add my own. I’m shooting for high wordcount, lots of variation and puzzles, and long play time, which is why I set the arbitrary goals of 10 small dimensions, each with about 2 hours of gameplay, so that the final result will be 20 hours (that way even if someone’s way faster than me, it’ll still be around 10 hours to finish).

I’ve beta tested a game by an author that easily rivals the games above, so something like that may be coming out in the future, but that’s up to the author if they want to talk about it.


Just flagging up that this recent thread covers another potential “size” measure…


How about word count in a complete play through? Complex games may well have text embedded in code in a way that is difficult to extract, but you could do a word count of the transcript. This would give an indication of the number of locations, but also the amount of description per location.

I can see issues, and you would possibly want to ignore repeated room descriptions, but seems better than just the location count.


It could. Getting transcripts from a sprint to the finish-type game vs a transcript of trying to do everything might be informative. It would at least give an average time-to-read that could be the starting point for an estimate.

In parser, depth of implementation is an x factor, too. It might reflect a ton of content, but incurious players may never see any or most of it. This isn’t necessarily the same as optional content with goals or problems to solve.

Deep implementation can lead to high word counts, definitely. I doubt the average player will see a quarter of my game, but they may all see different slices of the pie.

Yes, this is what I’m suggesting. At least there are loose standards for read time per word. It’s something to work with.


This is entirely subjective, but there’s something to be said about a game that feels big because it’s world feels expansive.

This is one of the many mystiques behind Enchanter, I think, that the player feels like they are but a small piece on a much broader playing field. The last IF Comp’s You May Not Escape! makes the player feel they’re just one of thousands (more?) who have attempted to escape. And, while not a text-based game, Grim Fandango portrays the life of a lowly salesman in the afterlife discovering he’s actually an unwitting cog in a conspiracy.


As a software engineering instructor, I’m more attuned to “size” in terms of the Implementer’s experience. In which case the size of an I7 project, for instance, might be a linear combination (or otherwise) of rooms/objects/rulebooks/phrases/extensions used, etc.


This is a great point. A player can have an experience of scope or scale separate from any measured quantity. This is a trick of authorial craft that can add a lot to play experiences.

Sure, that’s reasonable. All of those things take time and add to project scope. Custom verbs in my case were one source of story size and effort. I had no idea what that approach would require when starting out.


One more criterion, though not easily compared… total character count of non-library/extension source files. I7, I presume, is far more verbose than TADS3 to execute the same type of code statement. Then there are people like me who create about 500 macros to condense the T3 code even further…


I certainly hope it will be coming out! Trying to write some ending scenes while waiting for beta transcripts…


I saw your earlier post linked above about how many moves other games have. I think Mulldoon legacy is about 2100 moves, from what I was able to find.


How do you all get your printable words stats? Is that built into Inform IDE somehow?


When you compile the game, it shows this screen:

and this screen if you change the tab:


2100… wow, that’s big. For a minimal playthrough, eh?


Gotcha… so when you say 100k words, do you mean words in the source code or words that the player might see on screen?


It means words in source code. To find words that appear on screen, I think there are some threads where Zarf and others use python to figure that out, but I haven’t checked.

As for Mulldoon legacy, I got the info from this walkthrough:

It says:

Takes about 2100 moves.
Comments and points are in ()

I just checked Finding Martin, which I felt was bigger than Mulldoon Legacy. It has 1966 moves in the walkthrough (found by counting lines in the walkthrough, subtracting comments, headers, and blank lines)

Edit: My goal for my game is for each of the ten segments to require 200 moves, so it will be roughly the same length as mulldoon legacy. However, my puzzles are much easier than all of these, and I’ve already made some segments shorter, like the murder mystery.