Players treat par as a target to beat, but par is really a piece of information about the puzzle. It is the length of a shortest known route, which means the gap between par and your move count tells you how much slack the ladder gave you.
When a ladder has a lot of slack, the gap between par and a normal human solve stays small, because there are many reasonable routes and most of them are short. When a ladder has tight slack, the gap widens fast, and the difference between a good and a bad solve is one wrong fork. That is why a five-move solve on an easy ladder can feel worse than a six-move solve on a hard one: you are not comparing move counts, you are comparing how much of the difficulty you converted.
There is a second use for par that almost nobody talks about. If you record your move count relative to par for a month, the average of that difference is a stable measure of your play, and it moves much less than your raw scores. Raw scores swing with puzzle difficulty. The relative number is your own baseline, and it is the number I would track if I tracked only one thing.
Reading par well does change how you play, because you stop celebrating a hard puzzle solved one over par and start noticing an easy puzzle solved three over. Both are the same amount of lost control. If you have never looked at the answer archive this way, pick five recent routes and count the forks rather than the rungs. It is a small reframe and it makes the numbers mean something, which is the only reason to keep numbers at all. The solver page will give you a shortest route for any pair of words if you want to check a claim about par.
One warning about tracking relative move count: the first month is noise. You need something like twenty or thirty solves before the average settles, and the early numbers will swing with puzzle luck. If you want a lower-friction version, just record whether you finished at par, one over, or more. Three buckets is enough to see the trend.