Where players quit
Learning paths through this topic
A story is played over many evenings: count one quit per playthrough, not per evening
Story games are played in pieces, and every time the game closes looks like a quit. Counted that way almost everyone has quit many times, and the players who come back make the numbers nonsense. This entry shows how to join the evenings of one playthrough and decide when it is truly over.
Did my rewritten scene keep more players? Comparing before and after a patch
After rewriting a scene it is tempting to compare the quit numbers of last week and this week and cheer. But players who were already halfway through, and plain luck with small numbers, can both fake a win. This entry shows who to count and how big a change must be to believe it.
Do players leave after a shock in the plot, like a death? Measured in scenes, not seconds
Story games have moments that might push people away, such as a loved character dying. Whether players really leave after them cannot be read from a raw count; it needs a comparison with the usual chance of leaving over the same number of scenes. This entry shows how, and the runs you must leave out.
Find the one story line where many players stop dead, a sign of a stuck game
Normal quits spread out over a scene. When most of them sit on one single line, something at that line is broken: a button that does not appear, a sound that never ends, a choice that cannot be clicked. This entry shows how to find such a line from the log without reading every scene.
After players reach their first ending, how many start another run, and see a new one?
Games with several endings live from replays, but a player who abandoned a run before ever reaching an ending is not a replayer, and one who replays the same ending is not exploring. This entry shows how to count the groups that matter and what each number tells you about the story.
Players who come back but start a new game despite having a save: worry or not?
Someone who returns and starts over might be a happy replayer, or someone whose save was lost, or someone who gave up on a story that confused them. These look alike in a simple count. This entry shows how to separate them and where in the story the unhappy restarts cluster.
A laptop going to sleep made one scene look like it took all night
When time in each scene is computed from time stamps, one closed lid or one changed clock can add hours to a single scene and ruin every average. This entry shows how to compute time per scene from the log so sleep and clock changes cannot spoil it, and what to keep track of on the side.
Showing quit numbers by group without exposing a handful of players
Splitting players into small groups, for example by one rare ending or one unusual choice, can leave a group of two or three people who could be recognised or guessed. This entry shows how to hide such small groups in a report in a way that cannot be undone by subtracting from the total.
Your own test runs and bots make the quit chart look strange: how to filter them out
A story game that you or a helper click through to check a fix, or a program that walks it, produces runs that read impossibly fast or with identical timings. They pull every rate toward a finish that real people never reach. This entry shows how to recognise such runs from timing alone.
Why does almost nobody get my true ending? Find the requirement that stops them
A secret ending needs several things to be done first, and when few reach it you do not know which of the conditions is the wall. This entry shows how to follow players through the conditions one by one and, more useful, how to find the one condition that would unlock the most players if you changed it.
At what deficit does a comeback become so rare that offering to end the match is kind?
A surrender button that appears too early robs players of comebacks. One that appears too late leaves them in a lost game. Your own match history can say where the line is. This entry shows how to read it from data instead of guessing.
How long until new players first have fun, counting the ones who left before?
The average time to a first win or first big moment looks great if you only look at the people who got there. The ones who left first are exactly the story. This entry shows how to measure it without leaving them out.