When talking about deviated points (values) from a line formed by points (values) changing in a consistent way in a graph, which word better describe them: outliers or offsets?
What is the difference between these two?
The term is (statistical) outliers.
Some explanation (hopefully not too technical, we are a language site ^_^)
When measuring data, the measurements will - due to various reasons - deviate from "one correct or average value" Example: A child of age X will have an average height of Y. Obviously, not all children are exactly Y tall, a certain range is considered normal.
The term outliers describes values that are outside this "normal" range of deviaton.
The term offset describes shifting an entire system or set of data by a certain value.
Referring to my example from above: If you draw a graph of N children of age X, the height Y will vary within a certain range. Unfortunately, the height differences wil show as a small zig-zag line only. But there will be an average height Y'. So, if you use Y' instead of zero, your graph will show for each child, how it deviates from the average - above or below.