Wide-area Corn Hybrid Assessment Tool

An important concept that you’ll need to understand...

in order to appreciate the potential helps that this website can provide – as well as its limitations, is that of “competitiveness”. Unlike most hybrid performance summary reports that generate “head-to-head” (one-on-one) comparisons, our tool asks each hybrid to compete against the whole class, and not just one of the other class members (specific hybrid).

The performance metric tracked in this tool is yield expressed as a “percent of location average” (% of LA). Location average then is used to represent an estimation of the “yield level” provided by each plot’s unique combination of soil type, weather/climate, and management (planting rate & date, fertility, tillage, irrigation, weed control, etc.).

The advantages of this approach are several:

However, there are also disadvantages (limitations) to this approach:

• Advantages
  1. This tool records a “hit” for each plot location that the hybrid of concern is entered in, and not just the plots where it happens to be entered with a specific other comparison hybrid. In effect, this approach generates more hits; “more hits” equates to “more data”; and the more data -- the better.
  2. Hybrids have different “adaptabilities” as you move them from year to year, soil type to soil type, east to west, north to south, and to different yield levels, moisture regimes, fertility levels, tillage practices, and planting rates. As you change locations for any two hybrids, relative to the performance of all the other hybrids you could have planted (i.e. “yield level”) -- one of these specific two could be becoming more competitive, while the other one could be becoming less. That is why you can easily find head-to-head reports where Hybrid A beats Hybrid B, Hybrid B beats Hybrid C, and Hybrid C beats Hybrid A (akin to an Escher illusion). Each head-to-head comparison is generated from a different set of plot locations; and each set of plot locations represents a different level of competitiveness (i.e. ability to compete; ability to perform above average) for each of the two hybrids in question. From loc to loc they could be changing (relative to the “rest of the pack”, relative to the changing yield level) in the same direction but to different degrees, or they may even be changing in opposite directions. Although it seems counter-intuitive, comparing to a specific other hybrid represents more of a “moving target” than does compare to location averages (i.e. its merits (any given hybrid’s) are made less clear and obvious instead of more).
  3. Perhaps the greatest benefit of taking this approach is that it allows the patterns of the natures of hybrids (i.e. how they behave) to reveal themselves as you parse the data. Additional yet-to-be-introduced tools (have been developed but still need to be incorporated (coded) into this website) will enable you to more easily discover some of those patterns: e.g. the ability to compete as you move a hybrid east or west (“eastern hybrids” often need better disease tolerance in order to compete; “western hybrids” usually need better heat and drought tolerance), or north to south (most hybrids perform best where they are full-season (on the northern side of their “area of adaptation”; but that is not always true, especially if they flower (silk) late for their maturity). Other useful patterns that can be teased from this kind of data: the effect of elevation on a hybrid; response to planting rates, response to yield levels (low-stress vs. “tough” environments), and the interactions between these two.
• Disadvantages
  1. The primary disadvantage of this approach is that it uses an unbalanced dataset (i.e. all hybrids cannot be found at every location). As most data published by seed companies comes from plots that are predominantly hybrids of their own brand, they for the most part are not much more than just competing against themselves (i.e. how “competitive” their products are vs. true competitors is not very readily made obvious). The implications for this tool, then, are:
  2. Hybrids from companies with relatively “strong” line-ups show to perform not quite as strongly as they really are; whereas hybrids from companies with relatively “weak” line-ups actually show to perform (i.e. compete) just a little bit better than what they really do…
  3. Even though this may be true, the inaccuracies of this approach are probably much smaller (insignificant?) than one would think.
  4. The other main limitation of this approach has to do with a hybrid’s maturity relative to the rest of the plot entries. As a rule, later-maturing hybrids have a higher yield potential than do earlier ones. (The last time I checked it was about 5-eigths of a percentage point (0.625 %) per 1 day in maturity on average (i.e. with all other things being equal, a hybrid that is 8 days later should out-yield the earlier one by about 5%)) Hybrids that are consistently on the early side of the average of all plot entries are disadvantaged in this type of an assessment system. This, in reality, affects very few hybrids as they usually are scattered around enough (i.e. sometimes they are “early” and sometimes they are “late” relative to the other plot entries) that this effect washes out (i.e. disappears or goes away). The hybrids that “move south” best (i.e. “competes well with later hybrids”) are the ones that are most often over-represented as early hybrids in plots and therefore are the ones most likely to be misrepresented in this way. We offer the option of excluding all locations where a hybrid is more than 5 days earlier than the plot average (for maturity) as an attempt to minimize this concern. However, we can do nothing about the very earliest of all hybrids that are by definition always then the earliest of the plot entries (i.e. a 72-day hybrid is almost always one of the earliest entries in a plot and is therefore disadvantaged). The only thing you can do is keep this in mind and only compare it to other hybrids of the exact same maturity.
• Link
• Disabled