Weighting factor error under KCBS Rules
By JOE
O'CONNELL, cbbqa past
President
Updated April 21, 2002
The Kansas City Barbecue Society (KCBS)
currently weights the raw scores of judges at contests.
Under the KCBS Rules for weighting factors, each raw appearance score is converted to a
weighted score by multiplying it by 0.5714; each raw taste score
is converted to a weighted score by multiplying it by 2.2858; and
each tenderness score is converted to a weighted score by multiplying it
by 1.1428 -- as explained in the applicable
KCBS rule. An accompanying
story explains the mathematical basic for the complex weighting
factors.
Weighting factors
As explained in the accompanying
story, based upon the Rule of 7 and the Factor of 4, the weighting factors can be
constructed as follows.
The ratio of 1/7 times 4 equals 0.571428571428571428571428571428571. This
can be rounded to 0.5714 (using the normal rules of rounding up and rounding
down), which is set forth in KCBS Rule 19 as the appearance weighting factor.
The ratio of 4/7 times 4 equals 2.28571428571428571428571428571429. This
can be rounded to 2.2857 (using the normal rules of rounding up and rounding
down), but KCBS uses a slightly different factor of 2.2858, which is set
forth in KCBS Rule 19 as the taste weighting factor.
The ratio of 2/7 times 4 equals 1.14285714285714285714285714285714.
This can be rounded to 1.1429 (using the normal rules of rounding up and
rounding down), but KCBS uses a slightly different factor of 1.1428,
which is set forth in KCBS Rule 19 as the tenderness weighting factor.
KCBS rounding error
For some reason (perhaps an oversight), KCBS did not use the normal
rounding rules but used instead the slightly different weighting factors shown
above.
Comparing the KCBS weighting factors with the non-rounded and correct
weighting factors shows that the KCBS number are
incorrect.
Not significant
The error is not significant. A perfect score of 9-9-9 by a single
judge converts to a weighted score of 36.000. KCBS avoided any claim that
its computer program was erroneous, because the rules include the rounded
weighting factors, rather than a claim to the effect that "taste is
weighted twice as much as tenderness, tenderness is weighted twice as much as
appearance, and the weighting factors are rounded to four decimal
places". (As described here, such a claim would have been
wrong.) Therefore, the rounding error is not significant.
This is not to say that the rounding error is without consequence.
There are many cases in which the outcome of a contest would have been different
had the correct weighting factors been used. Nevertheless, because KCBS
Rule 19 refers to the actual factors and not their derivation, no one can claim
that the computer program is incorrect.
|
