I have found some scoville unit reports for certain peppers to be confusing or misleading because it is commonly expressed as a simple number. However, the Scoville unit rating could be a mean, median or mode average of the peppers tested, it could be the measure of the hottest pepper in the pile, or (worst of all) it could be a reading from one pepper that may not be representative of the group. For example, what if five pods each of two peppers were tested and came back with scoville ratings as follows:
Pepper 1: 950,000; 1,100,000, 950,000, 950,000, 1,300,000
Pepper 2: 1,100,000; 1,100,000; 950,000; 1,000,000; 900,000
In this example Pepper 1 would have a median and mode average of 950,000, and the mean average is 1,050,000. The hottest pepper in the pile is 1,300,000. Which of these three numbers should be reported?
In this example Pepper 2 would have a median of 1,000,000, a mode of 1,100,000, and a mean average of 1.010,000. The hottest pepper in the pile is 1,100,000. Which of these four numbers should be reported as the scoville figure?
This gets more confusing when you try to figure our which pepper is hotter. Pepper 1 produced the hottest pepper in the bunch, but pepper 2 had more pods over a million scovilles and had a better mode average. If only a single pepper was tested, you may end up with pepper 1 reading as hotter (e.g. pod 5 of pepper 1 versus pod 1 of pepper 2), pepper 2 reading as hotter (e.g. pod 1 of pepper 1 versus pod 1 of pepper 2), or about the same (e.g. pod 2 of pepper 1 versus pod 2 of pepper 2) depending purely on which peppers were selected.
Another point that was already touched on but needs to be highlighted again is that the number of peppers tested is extremely important, because that plays a role in calculating the standard deviation (SD) and the margin of error (MOE) in the sample. Without knowing the SD and MOE the number you get is pretty abstract and only meaningful in the most general "hot" - "really hot" - "sounds nasty" sort of way. The rub is that in order to get an accurate reading you need to test a large number of peppers: even in the largest gardens about a hundred randomly selected peppers would get you a MOE of just under 10%, while 300 would get you closer to 5%.
Of course no one would bother with testing a large number of peppers because it would really only represent that particular crop. Peppers are notorius for having variable heat depending on factors such as watering, soil conditions, etc etc etc. My hottest bhut may have less capsicum than your coolest one.