Russ, et al;
One way this difference between "accuracy" and "precision" can be
illustrated is to visualize a bulls-eye target.
Lots of holes tightly grouped over in one corner of the target represents
"precision"; Holes all over the target but centered on the bulls-eye is
"accuracy". If the measurement has precision but not accuracy, the error can
be calibrated out. The only way to deal with accuracy without precision is
to average a large number of samples; if the error is truly random, it will
average to zero.
Regards, Neil Tucson, AZ
-----Original Message-----
From: Russel Mack [mailto:rtmack@concentric.net]
Sent: Tuesday, May 06, 2003 9:15 PM
To: John Goodman; land-speed@autox.team.net
Subject: RE: The concept of Data?
John, Mayf:
your comments to Keith and Dave re the potiential for UNRELIBILITY,
non-repeatability of the data-- are exactly what I think we need to be
concerned about (NOT small but repeatable bias or imprecision). Russ, #1226B
-----Original Message-----
From: owner-land-speed@autox.team.net
[mailto:owner-land-speed@autox.team.net]On Behalf Of John Goodman
Sent: Wednesday, May 07, 2003 7:20 AM
To: land-speed@autox.team.net
Subject: Re: The concept of Data?
"The concept is to tune the car as a result the
Data... If the Egt's aren't exactly correct, it's
not really important... what is important is
what they are normally when the car has a reasonable
tune-up. Then what they are on this last pass."
Keith, keep in mind that data accuracy comes in many
forms. Russ talked about data variability. This is a
big one because if data which you record is not
stable, it is also not reliable, even as a comparison
to other data from different runs. Being off a few
degrees may or may not be a problem for you but if the
data is variable, you have garbage and it will not
correlate to other data.
/// unsubscribe/change address requests to majordomo@autox.team.net or try
/// http://www.team.net/mailman/listinfo
/// Archives at http://www.team.net/archive/land-speed
/// what is needed. It isn't that difficult, folks.
|