Plot Wars

The annual battle of the plots between companies will soon be into high gear.  This year we are seeing variability across fields that leaves experienced farmers shaking their head.  Yield monitors are telling the less experienced the same thing.  Yields from one end of the field to the other can vary from 80 to 200 bu. This is an unprecedented spread, at least in this part of Ontario. 
The biggest contributor to this is the lack of available moisture deep in the soil profile.  One wise person made the comment that yield maps this year will be revealing more about the variability of our sub-soils than anything else.       
Against this backdrop, seed company reps including yours truly will be flogging a variety of plot results.  How much validity is in the data presented?  An experience I had recently illustrates how misleading one plot can be.
I was with a customer who split his 12 row planter between Pioneer P0474HR and Dekalb DKC 50-45. 
The first side-by-side produced the following result.

P0474HR   -  191 bu @ 21.7
DKC 50-45 -  172 bu @ 21.5

These numbers would leave the folks wearing Pioneer underwear breathless with ecstasy.  Those with Dekalb underwear would claim foul because it was weighed by a Pioneer guy and totally disregard the whole experience.

We moved over 8 rows and replicated the weigh off.

P0474HR   -   182 bu @ 21.9
DKC 50-45 -   182 bu @ 21.6

Significantly different numbers 20 ft apart in a field that appears to be perfectly uniform.  We stopped there because it was time to go get something to eat.

The Pioneer underwear types would say we won by an average of 9 bu.  Happy days are here again.  The Dekalb underwear types would say the only reason Pioneer looked a little better was because P0474HR is a later hybrid.  You can see that by the difference in moisture.  I would tell both types to take a seat in the "time out" chair.

Does either result have any significance?  Not really.  Statistic analysis of side by side data tells the truth.  A 12 bu/acre difference has only a 60% chance of predicting the best hybrid for next year if we rely on only one plot.  If we had 10 plots, a 12 bu advantage increases the probability to 90%.  The problem today is with so many hybrids to choose from, getting a large number of comparisons in a local area is very difficult. 

As far as individual plots in our neighbourhood having a lot of significance this year, I tend to think not.  
My take home message is stay wary of the "painted underwear" types.