What to measure
Learning paths through this topic
What to Measure in Your Game, and What to Skip
Measure what could change a decision about your game: where players stop, lose, retry or go quiet. Start with five events, each tied to one question.
Add Anonymous Game Tracking With AI
Have your AI write a small event log on the player's device, ask the player first, then play one round and compare the log with what you really did.
How Many Players Do You Need for Game Data?
Five people find problems, about 30 give a rough number, and 100 pin it to about ten points either way. Being twice as exact takes four times the players.
Privacy in Game Tracking: Keep It Anonymous
Record what happened, never who played: no accounts, device ids, IP addresses or clock times. Ask first, as anonymous data can still need consent.
Game event log
Your AI adds a small log to your game that writes down what happens, such as a level started or a player died, in a plain file or in the browser on the player's own device. It keeps no accounts, device ids or IP addresses, sends nothing anywhere, and records nothing until the player says yes.
Win rates look skewed because players left matches halfway
A match where someone quit in the first minute tells you nothing about whether the sides are balanced, yet it counts as a win for the other side. This entry shows which matches to set aside and why late leavers are not among them.
A rarely picked class wins every game: is it overpowered or just lucky?
A class played in only a handful of matches can show a perfect record by chance, and nerfing it on that basis would be a mistake. This entry shows how to pull small samples toward the overall picture so the ranking is not dominated by luck.
One bug sent the same event ten thousand times and ruined my numbers
Tracking data has its own bugs: retries that arrive twice, names nobody defined, broken timestamps and loops that flood the log. This entry shows a small filter that keeps the numbers believable and tells you what it threw away.
Count how many tries a level really took, including the ones that just stopped
Tries look easy to count until someone restarts by hand, closes the game mid-fight or loses a message on the way. This entry turns a messy stream of events into clean attempts, each with a result and a length.
Count how many different players you had today without knowing who they are
To know how many people play each day you must recognise a repeat visit, but you do not want a lasting tag that follows anyone around. This entry shows a small trick that counts people within a day and cannot connect one day to the next.
How many online sessions end in a crash when the game never says goodbye?
A crashed game cannot report that it crashed, so the log just goes quiet. This entry shows how to tell a crashed session from one that is still running and how to get a crash share you can trust.
Does the dungeon scale properly when two, three or four players join?
A dungeon that is perfect for one player can be a pushover or a wall for a full group. This entry shows how to compare clearing times across group sizes, and why a group size that fails a lot can make its own times look better than they are.