Everything here is either a survey across several games or a short piece about one specific mechanism. Full reviews live on the reviews page.
Survey
Median and worst-case checkpoint gaps for every game we have played this year, manual save availability, and the real cost of an unexpected quit measured by deliberately forcing one at the worst moment.
Note
When losing progress is the entire design, a game has to be honest about what carries over. We compared nine roguelikes against twelve narrative games and the difference in clarity was not close.
Dispatch
Seven survival games tested by quitting without warning at the worst possible moment. Three handled it perfectly, two lost the current task, and two lost an entire day cycle of accumulated work.
Essay
Games increasingly assume a replay and rarely do the work of making the second run genuinely different. We think a replay must be earned, and our scale reflects that rather than treating length as value.
Audit
We have now verified nineteen games that advertise multiple endings by playing every one and diffing the text. Eleven were accurate, five overstated by a factor of two or more, and three were not really branching at all.
Note
A loop that carries knowledge forward is a save file that respects you. A loop that makes you re-execute what you have already proved is a checkpoint placed badly and dressed up as a theme.
