# Comparing Current year to five year average, should current year be included in average?

I often calculate a five year average for my data, and compare that to the current year. I realized that for 2023, I would find the average of 2019-2023, and compare it to the value for 2023. Is this correct? Or should the average be for the past 5 years (2018-2022), so that 2023 does not affect that value?

Just to have an example, if the past 6 years of data are

(2018 ,1), (2019, 2), (2020, 3), (2021, 4), (2022, 5),(2023, 6)

My current method would be the five year average is 4, and there is a 50% difference in 2023. The average of 2018-2022 would be 3, and there would be an 100% difference in 2023.

Which of these methods is more correct? Are there other factors this decision would depend on?

• If you want to compare the past with the now you should take the middle of the past 5 years not including the present year,
– trula
Commented Aug 1 at 13:33
• It depends on why you are looking at the five-year average. If you want to see how well it predicts this year's figure, it might be sensible to use the previous five years' average as you will not know this year's figure in advance. If instead you are trying to distinguish noise and trend without predicting, it might be better to take the five years centred on the year in question (which you cannot do for the last two years). Commented Aug 1 at 14:13
• Specifically when calculating moving averages ( i.e. technical analysis ), I'm pretty sure that the convention is to include it. But, in all the other cases, as henry and truly explained, it seems subjective. Commented Aug 2 at 3:52