This is a public-interest archive. Personal data is pseudonymized and retained under GDPR Article 89.

Re: Est weight


D1109@aol.com wrote:
> 
> I have a pumpkin that has an ovt total inches of 291. If you use the formula
> in Don Langevins book, "Dill's method", it shows an estimate of 552 pounds.
> By using the "Stelfplug method" of circumference only, it has an estimated
> weight of 500 lbs. Can anyone shed any light on this discrepancy.

I think your own words are the key: these are "estimates".  I dont know
much about pumpkins (my first attempts are just developing in the
backyard), but know more about engineering models.  Pumpkins are quite
complex;  these predictors are VERY simple.  To an extent, a predictor
with more parameters will tend to give better results.  It might be
interesting
to consider what tuning factors could be added to various models:
eg, genetic history, growing latitude, degree days since planting,
stem circumference, some measure of oblongness, some measure of
symmetry, ... .

There are statistical methods that would do this automatically.  One
feeds in say 2000 data points containing readings on many different
factors (say the 30 most plausible you can think of); and factor
analysis
will generate essentially an ordered list of the most relevant ones.  I
think
this little project crash lands at: noone cares THAT much.  The rough
rate
of growth per day is probably as important to many as the absolute
figure;
and its only the real weigh-off that counts (in terms of weight) anyhow.

Bill Sakoda

_________________________________________________________
I agree that the RATE is the most helpful statistic as this lets me compare
one to the other for deciding which to keep and allows me to compare growth
from one year to the next.

I use "average rate of growth". ie. I divide the circumference by the
number of days since pollination.  This easily shows if one is speeding up
or slowing down.  I do the same for the OVT total measurements.  This helps
to compare pumpkins with different shapes and provides a way to verify the
straight circumference measurement.  Graphing these curves gives an easy to
interpret visual picture of all this data.

Kurt Frederick
Edmonton, Alberta
Other Mailing lists | Author Index | Date Index | Subject Index | Thread Index