Where players quit
Learning paths through this topic
Do players quit more when their sanity or health is low? Count time, not quits
If most quits happen at full health, it does not mean full health is the problem, because players spend most of their time there. This entry shows how to compare how often people quit in each state of a meter by taking into account how long they spend in it.
How much progress do players lose when they close my horror game?
A player who quits right after a checkpoint loses nothing; one who quits just before the next one loses a lot and rarely comes back. This entry shows which numbers say whether your save points are too far apart, and why mixing in players whose game saved on exit hides the problem.
Do players stop my horror game because of the story or because it is bedtime?
Quits pile up at certain hours of the evening, and it is tempting to blame the scariest scene. But the real cause may be the clock. This entry shows how to find hours where people leave far more often than usual, even when sessions run past midnight.
Does a particular scare make players close my horror game? Compare to normal
Some players leave right after a scare, but players also leave at any other moment, so a raw count says nothing. This entry shows how to tell whether a scare really pushes people out, by comparing against how often they leave at random, and when too few people saw it to say.
How long can a horror player go without a safe place before they quit?
Fear needs relief, and a stretch with no safe room can wear people out. But looking at how many leave in each stretch is misleading, because fewer and fewer players are still there in the long ones. This entry shows a fair way to read the curve and where to put the next safe place.
Find the minute the city stopped growing, long before the player stopped playing
Players rarely quit in the middle of a fast climb. They usually leave after a stretch where nothing improved. This entry shows how to find where that flat stretch begins in a growth curve, so you can look at what the game offered at that point.
Tell apart players who gave up a lost battle from those who just walked away
Not every early exit from a match is frustration. Some players concede because they are losing, some leave while ahead, and some lose their connection. Lumping them together points you at the wrong fix. This entry shows how to sort them and in what order.
Do players leave because the game loads too slowly?
Some people close a game before pressing a single key. Whether that comes from a slow start or from a boring first screen, the numbers can separate the two, but only if you also count the ones who left while it was still loading.
Rank maps by how often players give up on them when some maps have very few games
A map with two games, both abandoned, looks like the worst map in the game. A map with a hundred games and one in five abandoned is the real problem. This entry shows how to rank maps so that thin data does not shout louder than thick data.
Find when a match was really lost and how long the loser stayed in a lost game
A losing player often keeps going for minutes after the match is decided, and that stretch is where frustration builds. To measure it you must pick the right moment in the score curve, and the obvious one is wrong. This entry shows which moment to use.
How much of a session does the player have nothing they can afford to build?
Waiting for money is part of a builder, but waiting with nothing to spend on feels like a stall. A simple count of moments hides it because your log is not evenly spaced. This entry shows how to measure the empty stretches properly.
After which building action do players pause for long, and after which do they leave?
A long pause in a builder can be a break to think, or the first sign of being lost. What matters is the last thing the player did before it. This entry shows how to attribute pauses and exits to actions, keeping the two apart.