Fault Line Ascent
The cleanest jump we have measured in three years. A 41 millisecond response, a six-frame coyote window and an arc you can predict from the first screen. Every one of our 1,400 deaths was explicable.
Each entry shows the median response figure across four controllers, the total hours we played before writing, and the score. Games we abandoned without finishing are not listed, because we do not score what we have not completed.
The cleanest jump we have measured in three years. A 41 millisecond response, a six-frame coyote window and an arc you can predict from the first screen. Every one of our 1,400 deaths was explicable.
A fighting game that publishes its own frame data in the training mode and displays network delay on screen during matches. Twelve characters, no filler, and the most honest teaching tools in the genre.
An optional hitbox display turns an opaque bullet pattern game into a legible one without lowering the difficulty by a single frame. The best readability decision we have seen this year.
Not the most accurate tyre model available, and the most communicative. Steering weight tells you the surface has changed before the visuals do, which is the whole job of a rally game.
Deliberate weight, honestly implemented for five of the eight weapons. The other three take control away during recovery in ways the animation does not advertise, and that costs the game a point.
An excellent movement model interrupted 214 times in fourteen hours by a camera that decides it knows better. We counted every instance and timed how long control was withheld.