OK well I ponied up the dollar to try it out (it is refundable after all) out of boredom or curiosity, I don't know...but I wouldn't call the results wildly inaccurate?
For example, 4 mode BLF A6 with 5A (~4000K) tint:
Modes 1-3 read between 3900-4100K
Mode 4 read 4600K
SC52w on most modes except the very highest read either exactly 4400K or slightly above.
L11C with Nichia 219b (4500K) read exactly 4500K except on the highest mode
So it has issues being accurate on higher modes (unless there's actually a tint shift going on here), and on some lights, if you move the spill into the center of the viewfinder, you can clearly see how much cooler they get (over 1000K cooler on some lights according to this). It also helps to be in a completely dark room. I haven't tried a true flooder, but I'd think they could be more accurate as the percentage of hotspot to spill in the viewfinder can change the reading. Every time, you hit the button to take a reading, an inset picture in the corner shows the intensity of your light across the viewfinder; therefore, with a flashlight, you see the hotspot clearly and its degrees of intensity from center to spill. This helps you to keep the hotspot centered or move the spill to center if you prefer.
In short, I wouldn't use this to replace a well-manufactured and calibrated piece of equipment costing several hundreds of dollars, but for a dollar, it seems to be in the ballpark enough to have fun playing around with.