Benefit: Still fuzzy so far. Both of those threads were exploratory, neither of them (yet) describe a concrete example where gyroscope is critical.
Cost: Adding a gyroscope (or replacing existing 3-DOF sensor with 6- or 9-DOF sensor) would require change to the sensor board BOM, PCB layout, STM32 firmware, and Pixelblaze to expose a new API.
Unknown (to me): Is there a way to patch gyroscope data into a current revision Pixelblaze for exploration and prototyping of patterns?
Rotation acceleration would be very handy for quickly adjusting to changes in rotation, where the gravity vector might not be sufficient.
There’s also some algorithms that can integrate 6 or 9 axis information to give really good orientation and improved dead-reckoning positional information.
I’m prototyping a version of the Pico with a 6 axis (accel +gyro) for use in poi / staves / hoops.
A future version of the sensor board could use this too.