If it is unavoidable, I have already made a feature in the Software to ignore the point if it is detected outside the calibrated area and always on.
This will be great, because my sensor bar emits 3 lights at the left and 3 at the right, totaling 6.
I find that with single-light stationary ir is good in small range such as <2 meters, because multiple-light will provide bias, the cursor can move in another loop, meaning when we move the cursor to the leftmost of the screen, suddenly it will appear at the center, then after we move the second time to the leftmost, the cursor will stop.
But when the distance is getting longer, single-light make cursor stop randomly, but only few times, it could be my "handycam-nightshot-mode-ir" light that is not strong enough, but could be something else, so if any information can confirm what is the longest distance between the two vishay ir leds, will solve the issue.
Oh... I miss that, it seems that I have to read every links there...
Using the accelerometer of the Wiimote for motion is not suitable like you have mentioned. It will be difficult to control the cursor.
Maybe I was overemphasizing the difficulty too much, the navigation is tolerable, here I attach my modified glovepie script, use it without sensor bar/stationary ir, if the cursor is running away while the wiimote stand still, use Wiimote Calibration.
Actually this is my first experiment using wiimote as mouse, because at that time, I didn't have sensor bar.
Going up and down is the same, but going to the left and right is different, the wiimote must be rotated clockwise for the right, counter-clockwise for the left, there is no need to swing to the right or left.
