What if the WAS and the LBS are BOTH different?
This is one of the challenges of reconciling or thinking-about-reconciling. Neither person is "the same," so what you're really aiming at is a new marriage that just happens to have the same physical persons in it as before.

I wonder if, in the flurry of lust and infatuation that makes up a new squidgy romance, your interests are in changing yourself to make the other person happy
Ed: What is the referent for "the other person" -- which other person are you talking about here?

many if not most romances start with an illogical assumption that the other person is perfect,
Assuming facts not in evidence. But if you want to play with the idea, maybe ALL romances HAVE to start that way. You're not going to go on Date #2 with She Who Farted Through Dinner, now, are you?

you don't want to let go of that "perfection", so you continue it, and maybe lose a little bit of yourself as you go along
Facts not in evidence. Probably as many iterations of how as there are people.

It is my contention that the wayward spouse returns in these cases because you finally shed the facade
With all due respect, you're promoting (as is often done on the boards) a curiously uni-directional model here. Wayward "returns" because LBS "returned" to LBS's "true self." Well where is Wayward in this? Is Wayward static? Was Wayward her/his "true self" when s/he got the hell out of Dodge? What of Wayward's facade?

This is a dynamic -- not static -- interaction. Which is what makes it so infinitely complex. Perhaps Mrs. SP -- flirting as she is with the idea, if not the reality (yet?) of SP -- is doing the same thing. Maybe there's a kind of DB'ing for Waywards.

So she changes; he changes; she observes his changes and changes; he observes her changes in response to his changes and changes. And they still "miss" the target, because now neither is the same. And reconciliation is a bridge too far.

Let me wax statistical (h/t @Coach, @Thinker, and our Big Midwestern City friend who has departed these boards).

What you're sort-of promoting here, @JonF, is a kind of (wikipedia link) linear regression. It's always struck me, in fact, that that model is at the heart of the DB method -- find out what was "wrong" (the explanatory variable(s)), change the value on that variable (she hated that I wore jeans, so now I wear linen trousers), and -- boom! -- you should observe a different outcome (you become the person only a fool would leave -- ta-da!).

Fair enough.

But the Walkaway isn't a "dependent variable" -- isn't a static outcome or observation.

What we have here is a strategic interaction in which both parties are engaged in (internet link) Bayesian updating. Each person is changing his/her beliefs (in lieu of probabilities) about the other in reaction to observable somethings in the other.

My WAW went from being a loving, generous, moral, caring, amazing mother and wife, to a selfish, arrogant, affair-consuming, so-so mother. Am I to believe that I lived with a facade for almost 8 years?

Surely not. But you are to believe that (a) she has changed (though perhaps not/probably not in a permanent way), (b) that that change was (at least from her POV) necessitated by some countervailing set of changes in @JonF, and (c) that whatever changes you make to "bring her back," at the end of the day Back -- Cue Peaches & Herb -- will depend upon --

(1) her next-round evaluation or (i) @JonF, (ii) the marriage itself, (iii) herself, and (iv) @JonF + marriage + herself, and

(2) @JonF's round-after-next-round evaluation of (1).

Which is probably why reconciliation is such a low-probability event.