What to measure
Learning paths through this topic
How long do players sit around dead while their friends keep playing?
Waiting to come back is the most boring part of a multiplayer match, and the average hides who suffers most. This entry shows how to measure dead time per player so that doubled log entries do not inflate it, and how to spot the long single waits.
How much of a co-op match is played with fewer players than it started with?
When teammates leave early, the people who stay play a harder game than you designed. This entry shows how to measure how much of the match was played at full strength and at less, even when the logs contain doubled or overlapping entries.
How many attempts does a group need to beat the boss, counting those who never did?
If you only look at groups that won, the boss looks much easier than it is, because the groups that gave up are exactly the ones you left out. This entry shows how to estimate the effort with every group included.
What ran out first before the whole team went down: ammo, healing or something else?
Looking at what was empty at the moment of defeat blames whatever happens to be empty at the end. This entry shows how to find the shortage that really started the slide, and why an early shortage that got refilled does not count.
Which loading times hurt a group game, the average one or the slowest one?
In a group match everybody waits for the last person, so a good average loading time can still mean long waits. This entry shows the figures that describe what a group actually experiences and how to spot matches held up by one player.