A small note: the rotations are detected by accelerometers+gyroscopes+magnetic sensors apparently. I suppose the latest (earth magnetic field sensor) gives a bearing reading and keeps down the need for calibrations of the gyros. Anyway the accelerometers in the Wii are 100Hz, I don't think the rotation data for the Sony wand will be slower than that.
More importantly: if I were designing a "pointer" based game I wouldn't be using exclusively the rotation data. I'd be mostly using the translation of the glowing sphere as read by the camera.
I think that's precisely the problem. Likely the games that are using a pointer are using the software that analyses images from the PSEye and gives us the position of the wand relative to the camera (fixed to the TV). ALL the games demonstrated have shown lag (similar to Natal's lag), and that lag is fair enough when you're doing full gestures, I think people can appreciate the amount of work the console is doing to replicate what you're doing onto the screen.
It's just that when you go and use the same method for something like a pointer, which you definitely want to be fast and responsive (otherwise it's not much of a replacement for dual analogue), you have a lot of overhead for something that should be simple and easy. The fact that the IR camera on the Wii is very simple means it's no burden to the Wii.
Maybe they'll solve this if they can figure out a way for the software to analyse the images from the Eye more efficiently, especially for non "augmented reality" games, but I have a feeling this is going to be a permanent problem because they wanted to go with the Eye camera rather than one designed specially for the task.
A game I'm developing with some friends:
It is largely a technical exercise but feedback is appreciated.