Last time I did this, I wrote that I didn't think that I should do it again.
I have a practice of documenting the reasons for decisions, largely in order to close the loop: if it turns out, afterwards, to have been a bad plan, then I can use the logs to find the bug and fix it†.
That goes double when reversing an earlier decision. So here's the deal:
Your identity is what you have proved to yourself. But that proof expires, and my last marathon was in April 2014 (!). If I want to keep thinking of myself as an endurance athlete, I have to get recertified.
I've also written before about the hedonic treadmill, and about how capability-improvement is the only way off. There will come a time when I can't do that any more - when, each year, I'm a little slower, a little weaker, a little less smart, than the year before.
But today is not that day. I am better now than I was in 2014. By the time I've finished getting back into marathon shape, I'll be tougher, too.
Which is why my normal race triple-goal of "finish it, don't hate it, don't get hurt" is getting a fourth goal: "set a marathon PB". If I run this in more than 4:14:57, I would have been better off not running it at all. (I think I can probably do 3:45, but that's an ideal, not a requirement).
And, more importantly, because I need to prove it.
† Well, at least in theory. In practice, the counterfactual is often sufficiently ambiguous that it's not possible to know whether it was actually a bad plan. But having a detailed log of decisions helps to prevent me from revisiting them unnecessarily; it's effectively a regret-minimisation system.