Difficulty
Learning paths through this topic
Are some deaths caused by the game stuttering and not by the player?
A freeze at the wrong moment can kill a player who did everything right. This entry shows how to compare deaths with stutters while allowing for the fact that stutters overlap, so you know whether performance is hurting your difficulty.
How often do players die to a hit that only just touched them?
When many deaths come from the very edge of a danger, players blame the game, not themselves. This entry shows how to measure that edge, and how to compare it with the near misses to judge whether your danger zones are too generous.
How many deaths happen while the player cannot do anything?
Being juggled by an enemy until the health bar is empty is the most hated way to die. This entry explains how to measure those helpless stretches properly, including the brief moments of control that do not really count.
Are players dying again right after they respawn?
A respawn that puts people straight back into the danger is a different problem from a hard fight, and it makes players quit fast. This entry shows how to find those quick repeat deaths and the chains they form, without mixing them up with first arrivals.
Do players who stop right after dying ever come back?
Some quit on a death and return the next day, others never do. This entry shows how to compare them with players who stopped for other reasons, and which recent sessions must be left out because their story is not finished yet.
Where do players run out of lives, and are extra lives wasted?
Adding up lives won and lost per level hides when the last life actually goes. This entry shows how to replay real runs in order, so you can see which level ends them and how many bonus lives were thrown away.
Is my level two games: one for players who breeze through, one for those who drown?
An average number of deaths can hide that one group barely notices the level while another group is stuck for ages. This entry shows how to spot such a split in the data and how clearly separated the two groups are.
Does a power-up really help, or do only the good players manage to get it?
Players holding a power-up survive more often, but that may only show who they are. This entry shows how to compare each player with themselves, so you see what the item changes and not who picks it up.
Which difficulty do players really play on after they change the setting?
The setting chosen at the start tells you little once people switch it along the way. This entry shows how to count the time spent on each setting, including players who go down and later up again.
How often does this attack hit players, if a few players try it hundreds of times?
A pooled hit rate is ruled by whoever fights the most, usually the players who struggle. This entry shows how to count every player equally so the number says what a typical person experiences.
How big should an enemy be so that most shots from real players can land?
Guessing a target size gives either a game of constant misses or one where nobody can miss. This entry shows how to take the size from the aim errors of real players, including shots taken from very different distances.
How long does it really take to kill an enemy once misses and reloads are counted?
Health divided by damage times the fire rate gives a perfect-player time that nobody achieves. This entry shows how to compute the expected time for a player who misses and has to reload, so fights are tuned to what happens.