07Quarterly
Weight And Momentum
How movement feels, measured
Weight And Momentum
How movement feels, measured
This is the slowest character we have measured in an action game that expects the player to dodge, and the four frames of commitment on a direction change are the single most consequential number in the design. We spent twenty-six hours deciding whether they are a filter or a fault.
26 hours · Controller measured at 240fps · Default configuration, no accessibility options enabled
The first hour of this game is a negotiation and most players will lose it. You hold a direction and the character does not immediately go there; you release it and he does not immediately stop; you press dodge during a swing and nothing happens for four frames. Everything a modern action game has trained players to expect from a stick has been slowed down, and the game offers no explanation and no tutorial acknowledgement that anything unusual is happening.
What the game is doing becomes clear at about the six-hour mark, and it is not the usual claim about weight for its own sake. The commitment window is not there to make the character feel heavy. It is there to make the player's decisions visible to the enemies, who are reading the same four frames the player is spending, and who behave differently depending on what was committed to.
| Property | Frames | Cancellable |
|---|---|---|
| Standing to full walk | 11 | Yes, after frame 4 |
| Full walk to standstill | 19 | Yes, at any point |
| Direction change, ≥90° | 4 | No |
| Dodge, start to invulnerable | 6 | No |
| Dodge, total | 28 | After frame 22 |
| Landing recovery | 9 | No |
| Light attack recovery | 14 | After frame 9 |
| Heavy attack recovery | 31 | After frame 24 |
Two numbers in that table are doing most of the work. The first is the four-frame uncancellable direction change, which means that at sixty frames per second the player is committed for sixty-six milliseconds every time they turn — long enough to be felt, short enough that most players never identify it as the source of what they are feeling.
The second is the six frames before a dodge becomes invulnerable. That is a long pre-window by the standards of the genre, and it is what makes the enemy design legible: attacks in this game are telegraphed for between eighteen and thirty-four frames, so a six-frame dodge start is always affordable if the player reacts to the telegraph rather than to the hit.
The armour matters. Movement figures change with equipment, and unlike most games in this genre the changes are large enough to constitute different characters rather than different statistics. In the heaviest configuration acceleration goes to fifteen frames and the direction-change commitment to six; in the lightest it drops to seven and two. We measured all four tiers and the spread is consistent, which means a player who finds the default intolerable has a real remedy rather than a cosmetic one.
Equipment weight is displayed as a category rather than a value, so a player comparing two chestplates cannot see that one costs two frames of turn commitment and the other four. The information exists, it is decisive, and it is hidden behind the words medium and heavy.
This is the only interface complaint in the review and it is a serious one: the game's most interesting system is the one it declines to expose.
The camera is also correct, which sounds like a small thing and is not. Camera acceleration is slower than character acceleration by roughly four frames at every tier, so the world appears to lag slightly behind the character's intent. That lag is what produces the sensation of mass, more than the character's own numbers do, and removing it in a configuration test made the same movement feel substantially lighter without a single locomotion value changing.
Weight in this game is mostly in the camera. The character's numbers make it fair; the camera's numbers make it felt.
Falling. The landing recovery of nine uncancellable frames applies identically to a step off a kerb and a drop from four storeys, and the game contains a great deal of low terrain. In an hour of ordinary traversal we counted forty-one recoveries, of which perhaps six were from heights the fiction would consider meaningful. It is the one place where the commitment model is applied without thought.
Stairs are worse and for the same reason. Descending a staircase registers as a sequence of small drops, each incurring recovery, which turns a four-second walk into a stutter. We are confident this is unintended because the game's own scripted sequences avoid stairs entirely.
Reader one holds that the first hour is a filter and a legitimate one: the game is about commitment, the opening teaches commitment by punishing its absence, and a gentler opening would produce players who fight it for twenty hours instead of one. Reader two holds that the opening does not teach anything, because nothing in it explains what is happening — the player is simply worse at the game than they expect to be and is given no vocabulary for why.
The evidence is on reader two's side. There is no moment in the first hour where the game demonstrates that a direction change costs frames, and the equipment screen actively conceals the mechanism. A single scripted encounter with a slow enemy and a visible commitment indicator would resolve it, and its absence is not restraint but omission.
Both of us think the game is very good and neither of us would change the numbers, which is worth stating plainly. The complaint is about disclosure, not design.
Against the weightless platformer this game is a different species: eleven frames against two, nineteen against zero. Both are correct, because a precision platformer needs the player's intent expressed instantly and this game needs it expressed at a cost. Against the twenty-two-tonne walker it is nimble, and against the driving game it is honest — everything this game's art promises about mass, its physics delivers.