Because the period (4 years) is even, the moving averages first fall between two years and must then be centred with a further 2-yearly average.

Step 1 — 4-yearly moving totals and averages (each average falls between the 2nd and 3rd year of its group of four):

Step 2 — centring (2-yearly moving average of the above averages), to place each trend value onto an actual year:

Independent check. Recomputing each 4-year total directly from the original series — e.g. $50+54+49+61=214$ and $54+49+61+60=224$ — reproduces the same totals, confirming the averages before centring; the centring step itself is then a simple, verifiable pairwise average.

As with the 3-yearly case, no trend value is available for 2015, 2016, 2021, or 2022 — the years lost at both ends grow with the length of the moving-average period.

Reporting the uncentred 4-yearly averages (53.50, 56.00, ...) as the trend values against actual years — an even-period moving average is meaningless until it is centred, since it does not correspond to any single year on its own.