The Traffic Accident Reconstruction Origin -ARnews-
|
The Traffic Accident Reconstruction Origin -ARnews-
|
|
I agree with you wholeheartedly on the need to examine the sensitivity of your results with respect to the range of possibilities of your input data. However, if you know the acceleration is between 0.88 and 0.89 ft per sec^2, then when you convert to an acceleration factor in multiples of g, you do not want to round (or truncate) to the nearest hundredth, but rather the nearest thousandth--that is, 0.88 fps^2 produces an acceleration factor of 0.0273, while 0.89 produces 0.0276--in either case the number is between 0.027 and 0.028, not between 0.02 and 0.03. Using 0.027 and 0.028 produces a much smaller variation in the calculated time than what you reported.
Actually, to minimize round-off/truncation errors, it is advisable to do chains of calculations with techniques that do not round or truncate until the final result. Then report it with either an appropriate number of significant digits (which are not necessarily the number of digits to the right of the decimal point), or with a range of values depending upon the results of the sensititivy analysis. By converting 0.88 to 0.02 through a multiple truncation process has actually reduced the precision of your number from two to one, which does produce a big problem--as you have pointed out.
Bruno Schmidt
BFSchmidt@worldnet.att.net
For example, to continue this discussion look for a thread titled
Accuracy of rounding caution
If this thread does not exist in the current archive, you can begin another one by using that title.