I tested out three different versions of IR receiver placement on your head. I'm sticking with placing these on top of the head for now instead of on the device, since you can get consistent 360 degree coverage on someone's head. When playing with friends, people tried to hide their device to not get tagged, thinking that would save them. So, instead of trusting that humans will change (and not try to cheat), I'm designing the system around them.
From bottom left to top right in the picture below, there is the first version, with eight IR receivers in an octagon shape (to give 360 degree coverage), which sits on top of your head. When testing with this one, it was hard to consistently tag people, as they would tilt their head up (changing the angle, so you couldn't hit the IR receivers, as their forehead would block the beam), or their hair would get in the way.
The second version, the one in the middle, has a tighter pattern of IR receivers (still covering the 360 degree angles, just more compact), and it is sitting on top of a one inch tall stand to make it sit an inch up from the top of your head. This one was easier to hit, despite people's hair or head tilt.
In the third version, top right, I soldered the infrared receivers onto copper tape along the RGB LED strip. That way, there was an IR receiver on the person's forehead (making it very easy to see and hit).
A closer image of the IR receivers soldered to copper tape I placed on the back of the RGB LED strip is shown below.
After comparing these three setups for IR receivers, I found that the second one was more reliable. For some reason, the IR receivers attached to the RGB LED strip performed the worst. I would have thought it would do better, but when tagging it, the device consistently missed consecutive tags (fired within 1/3rd of a second. Slower than that, and it got all of them). I don't know if there is some capacitance in the copper tape or something that the device is picking up, but from the experiment, I should do the next series of tests with the second version.