A player dies five times in the same encounter. A system could interpret that as a request for help. The player might instead be practicing a difficult technique, experimenting with equipment or deliberately accepting risk. The count tells us what happened, but not what the player wants.
View full image
That ambiguity is the central UX problem in adaptive difficulty. Adjusting the experience is only useful if the adjustment addresses the actual obstacle and preserves what the player values about the challenge.
Decide which difficulty is changing
Imagine a hypothetical encounter where a player understands the strategy but struggles to execute a narrow timing window. Reducing enemy health might shorten the fight without making that window any more manageable. Slowing the relevant action or offering an input adjustment could address the obstacle more directly.
A different player may have no idea which attacks can be interrupted. They need readable feedback or an opportunity to learn. Quietly weakening the enemy might let them pass without understanding the mechanic they will need later.
Difficulty is not one variable. Execution, planning, resource availability and perception can create different barriers in the same encounter.
Offer help without deciding what failure means
One option is an explicit, dismissible offer after repeated failures. Another is allowing players to adjust specific parameters before they need help. The Last of Us Part II provides a concrete reference for the latter: Naughty Dog documents separate custom difficulty settings for aspects including enemies, stealth and resources. That is player-controlled customization, not evidence of automatic adaptation.
Even explicit offers need care. A prompt after every death could feel intrusive or suggest that persistence is a mistake. Timing, wording and the ability to decline matter as much as the available adjustment.
Completion is an incomplete success measure
If more players finish an encounter, the system may have helped. It may also have removed the part they wanted to master. Ask whether they understood what changed, felt responsible for the result and would choose the assistance again.
For competitive play, consistency between players adds another constraint. A design that supports a personal campaign may be inappropriate for a shared contest.
I would evaluate an adaptive system by how well it supports a player's chosen experience. Finishing faster is one possible outcome. Feeling that the game silently moved the goalposts is another, and completion data alone cannot tell them apart.
Reference
Naughty Dog: The Last of Us Part II accessibility and difficulty options. The adaptive scenarios in this article are design examples, not claims about a released feature or a patent implementation.