Difficulty
Learning paths through this topic
How much of a session does an idle-game player have no goal within reach?
The game may be full of things to buy, yet the player stares at a screen where the next one is far away. This entry shows how to measure that stretch as a share of time, and why counting samples instead of time gives a different answer.
Will players reach this event's goal in time, judged halfway through?
A straight line from the first days overestimates how far players will get, because events are front-loaded. This entry shows how to project the final result from last time's shape so you can adjust the goal while there is still time.
Find the typical player's progress when idle-game numbers span many orders of magnitude
In a game where numbers grow without end, the average is nothing but the one furthest player. This entry shows how to store and compare progress so that the typical player is visible, and how to measure how much one player dominates.
Choose how many hours of offline earnings to give, from how long players are really away
A cap that is too short punishes the people who come back the next morning, and one that is too long makes checking in pointless. This entry shows which share of returns to cover and which short absences to leave out of the sample.
Check that each building tier in an idle game still pays back in sensible time
If a higher tier pays back faster than the one below, nobody wants the lower one; if it pays back far slower, players hit a wall. This entry shows the one number per tier to compute and the two warning signs to look for in the list.
Spot idle-game players who have stalled, judged by their own rhythm and not a fixed time
A pause that is normal for one player is a warning for another, so one fixed limit marks the wrong people. This entry shows how to compare each player with their own pace and count at which step the stuck ones are waiting.
Find the upgrade where the waiting time suddenly jumps in an idle game
In a game that earns while you wait, a wall is not an expensive upgrade but a sudden jump in waiting time compared with the step before. This entry shows how to compute the waits from your price list alone and where to look for the jump.
Is this player still learning the fight or really stuck?
Ten failed tries can mean steady progress or going in circles, and the game should react very differently to each. This entry shows how to tell them apart from the tries so far without being fooled by one lucky or unlucky run.
Why do my later levels look easier on paper than they feel?
Average deaths per level often drop toward the end, not because the game got easier but because the players who struggled are gone. This entry shows how to compare levels using the same people so the curve tells the truth.
Which level is the real wall, and which one is only a bit harder than the last?
Completion rates fall naturally as a game gets harder, so comparing every level to a fixed target flags the wrong ones. This entry compares each level with the levels around it, to find the single level that sticks out.
Measure how long a level really took without the time the phone sat on the table
On a phone, people put the game down mid-level and come back minutes later. If you time levels from start to finish, those pauses make an easy level look slow. This entry shows how to count only the time someone was actually playing.
Do not count an attempt as lost just because a phone call interrupted it
Phones pause games at unpredictable moments, and a timer that keeps running can end the attempt for the player. If those get counted as failures, a level looks harder than it is. This entry shows how to recognise and set them apart.