|Metrix » Tuning|
By "having a reason", we meant that you should not exclude a point merely to improve the fit (that is, the R2 value). If a point is an outlier, and you know why it doesn't lie on the fit line like the other points, then you may be justified in excluding it. If the point is due to a one time event, that most likely will not occur again (such as a gas leak, a fire, a convention, an equipment breakdown), then go ahead an exclude it. You would not billmatch this excluded point because you do not expect this anomalous event to occur in the future. Be sure to make a note somewhere why you are excluding the point, as you may at some point have to justify the exclusion to a customer, his consultant, or a jury.
If there are several points that don't fall on the fit line, and you can identify why they do not fall on the fit line, and you expect them to recur annually (such as for a school's summer break), then the best course of action is to find and tune with an additional variable, such as number of school days.
If you are unable (or unwilling) to tune with an additional variable that accounts for the several points' deviation from the fit line, then you can, as a last resort, manually exclude those points. If you do so, you should billmatch those manually excluded points, as you do expect the deviation from the fit line to occur annually. For example, in the case of a school's summer break, you could manually exclude the summer bills in the tuning screen. Then, since you expect the school to shut down every summer, you would attach a modification, and billmatch the manually excluded points.
Otherwise feel free to contact our Tech Support staff at (805) 329-6565, or via email at email@example.com.