View Single Post
Old 31 January 2021, 02:13 AM   #409
saxo3
"TRF" Member
 
saxo3's Avatar
 
Join Date: Dec 2020
Location: .
Posts: 3,105
Quote:
Originally Posted by HiBoost View Post
… The +/- 2 spd spec relates to (average) precision. Thus if you had a watch which lost between 59 and 61 seconds a day, every single day for a year, it would actually have near perfect precision (+/- 1 spd) and be completely within the Rolex target standard … The expectation is that because the precision is so good, a simple regulation can now achieve the accuracy (actual deviation from an atomic clock over time). But if precision is lacking, the accuracy will be much more of a crap shoot and likely needs to employ compensators (setting the watch in position X overnight to undo the deviation during the day, etc).
I am afraid this description is incorrect because of a wrong understanding of precision.
Accuracy refers to how close measurements are to the "true" value.
Precision refers to how close measurements are to each other.


Quote:
Originally Posted by EEpro View Post
Many members seem to expect all 6 positions to be inside the 2 seconds box.
2, 4, -6, 4, 5, 3 is actually in spec to give an example
As is: -5, -4, -3, -2, -1, +3
Yes, you are right, not each position needs to be in the range -2/+2 sec/day.
That is already a big misunderstanding.

Your listed number (or rate) sequence:
I understand your calculations, but both examples are incorrect.
The 12U position rate must not be taken into account; it is also not used by Rolex.
Instead, one has to add all rates of positions DU, 6U, 9U, 3U, DD, and then divide by 5.

The topic Accuracy and Precision merits another thread.
Please, let's not get lost in this data thread!
saxo3 is offline   Reply With Quote