Question Valorant — USceneComponent ComponentToWorld Offset and Bone Conversion

[byte_corvus]

Anyone currently digging into the bone conversion logic for Valorant? If you are trying to map bone transforms to world space for an external ESP and getting zeroed-out results, you are likely hitting a wall with the ComponentToWorld offset.

When your debug logs show

c2w_s=(0.00, 0.00, 0.00)

it is a dead giveaway that your driver is reading garbage or the offset has shifted in the latest build. Here is a breakdown of the current implementation hurdle.

Technical Implementation
The logic revolves around reading the FTransform from the mesh (USkeletalMeshComponent), which inherits from USceneComponent. If the scale and translation are coming back null, your world-to-screen flow is dead on arrival.

1. Reading the Mesh Transform:
   

   meshCompToWorld = driver.Read<FTransform>(mesh + ComponentToWorld);

2. Bone Array Handling:
   Typically hitting

   BoneArrayCache (0x0AB8)

   with a fallback to

   BoneArray (0x0AA8)

3. Transformation Logic:
   Applying the rotation matrix and translation from the component space to get the final world position.

Spoiler: Current Debug Trace

[bone-trace] bone_world ok mesh=0xCDD9AAB0 bone=3 local=(-10.1,-3.9,116.8) c2w_t=(0.0,0.0,0.0) c2w_s=(0.00,0.00,0.00) c2w_q=(0.000,0.000,0.000,1.000) world=(0.0,0.0,0.0)

The Reversing Hurdle
Offsets like

0x250

and

0x2D0

are floating around, but neither seems to yield valid data for the current USceneComponent structure. For those reversing the binary, the most reliable path is tracing CalcNewComponentToWorld in IDA. This Unreal function handles the update of the transform member—look for where the SIMD instructions (like

vmovdqa

are dumping the calculated transform back into the component memory.

If the scale is zero, you might be reading padding or a deprecated member. Has anyone verified the exact shift for the latest patch?

Drop your IDA findings or updated offsets below.