Earth takes shape, Fire spreads, Water finds its flow

Day 3 connected Earth abilities into a timing test and made Water performance a question worth measuring. Day 4 extends that boundary in three directions: Earth needs weight and player control, Fire needs a legible gesture across a larger space, and Water needs to keep flowing while its shape changes.

The work now includes two playable Earth abilities, a broader Fire sweep with coordinated footwork, and the first Blender study of two successive whitewater waves. The common process stays deliberately modest: record feedback, change one bounded part of an ability, test the relevant behavior, produce a comparison, and preserve the revision that produced it.

G, Rolling Stone starts by raising 24 stones from the ground and assembling them around the visible earthbender over roughly 0.9 seconds. Once the shell closes, the player has five seconds of camera-relative WASD movement. Releasing the keys brakes the roll, an obstacle collision shatters the shell, and F10 cancels during assembly or movement.

The implementation initially inherited ordinary walking-speed limits from the sample locomotion system. The revised movement keeps the faster rock-form velocity while the existing character movement system still supplies actual travel and ground support. The current test measured a peak near 12 metres per second, zero travel during assembly, and five seconds from readiness to automatic return. Direction changes, braking, early cancellation, repeated input, and invalid activation were checked at 30, 60, and 120 Hz.

This is a controlled movement form, not a freely simulated ball. Its shattered pieces use physics, while the shell itself remains governed by the player and the existing character systems. There is no cooldown or impact damage in this prototype yet.

F, Crushing Walls raises two solid walls, each roughly six metres long, on either side of the space ahead and closes them together. Review moved the design away from stacked objects and toward continuous slabs with a sharper closure.

The previous breakup still read as cubes. The revision reuses R's irregular shard language: three distributed bursts produce 72 smaller fragments that scatter, tumble, and settle. A shared pool caps the physics fragments and retires them after their short lifetime.

The brown painted treatment, broad fractures, and intentional color changes support the stylized fantasy direction. This is cosmetic breakup rather than an exact fracture of the original volume, but it makes the impact easier to judge. The cleared review captures are silent; native gameplay retains impact audio.

A wide arm swing was not enough while the feet remained fixed. The current Fire motion allows a lead step, hip rotation, chest follow-through, and rear-foot toe pivot before recovery into the right-hand punch. The anatomical left hand still drives the sweeping attack.

For this study, the reference supplies the ability concept and intended area instead of imposing a frame-for-frame pose match. The latest effect is a 270-degree authored arc with greater thickness, taller tongues, and a longer-lived broad shape. At one measured source frame, its outer radius grew from about 4.3 to 6.8 metres. That is preview geometry, not an implemented damage radius.

This remains a Blender review. It does not establish Unreal performance, aiming, gameplay coverage, or a combustion simulation. The effect is painted, animated geometry and still needs angle-by-angle review.

The first twin-wave study takes canonical source frames 2800 to 2865 and builds two successive whitewater waves. The character loads into a raised-knee step, sweeps out the first wave, crosses and reloads the torso, then drives a second taller wave forward.

The waves release at approximately 0.667 and 2.083 seconds. Each begins at its own release location and continues independently, so the later character motion does not drag the first wave along. Painted cyan and white regions move across curved surfaces, with separate spray instances supporting the crests.

There is no previous twin-wave animation and no Unreal implementation or input mapping yet. The selected source ends while the second wave is active, so the study preserves that limit instead of inventing a recovery. A related Water E review isolates moving painted flow while holding geometry, body motion, water shapes, and ice-projectile timing in scope.

These are controlled artistic studies, not a fluid-solver result. A moving crest can still look like a sheet, and a correct surface can still look too uniform. No controlled Earth, Fire, and Water runtime performance comparison exists, so Blender render times are not evidence of gameplay cost.

The latest Earth, Fire, and Water work is preserved in checkpoint 13278ea, alongside the main commits for rolling Earth movement, twin whitewater waves, Fire sweep thickness, and coordinated recovery. That checkpoint records local history and evidence. It is not a remote backup or blanket approval of every experiment.

The next reviews are concrete: does Earth's formation delay feel satisfying before movement, does Fire's larger sweep retain a readable gesture, and do Water's waves feel like advancing masses rather than moving sheets? Those questions are more useful than adding detail everywhere at once.