| Our default values are 25' plateau and 100' search radii. So that is 50' and 200' diameters respectively. 100' may be a little aggressive (or would that conservative?). I mean by using 100', we may be allowing a yield point from too far away to impact our estimate at a specific site, introducing more smoothing than is appropriate? But, those settings are user preferences. Yes, a 1,000 BPA yield within the 50 circle around a site being estimated could still produce a bump, but if it is that high, it is because the area is small. But in real life, we don't see those bumps happen much. I suspect it is for two reasons. First, in our yield data import algorithms, we filter for "small area". If the combine is moving too slow, or if it is cutting too narrow of a swath, we just flag that site as non-mappble (with the reason), so it never figures into any of our spatial calculations (still figures into total quantity). The second is that our combine dynamics algorithm doesn't adjust 100% of the grain flow by "12 seconds". We use a "normal" distribution over 5 seconds, centered at 12 seconds. I think we use 50% at 12 seconds, 20% at 11 and 13, and 5% at 10 and 14 seconds. I don't remember. I wrote it too long ago :( But, anyway... I believe we have those two things going on that reduce the need for some of the data cleaning that I see people talk about. |