if you've already decided that, come hell or high water, you are going to commence retirement when your stash hits $X
I think this assumption is valid for essentially nobody in the real world, and without it, the rest of the hypothetical falls apart.
To the extent that anyone ties their retirement date to a trigger that cannot be disarmed, it's almost always going to be a
time-based event rather than a
dollar-based event (45th birthday, when I finish this major project next July, etc.) Dollars fluctuate so rapidly from day to day, and $X in the stock market in 2009 felt completely different than $X in the stock market in 2014, that strictly targeting a dollar amount makes no sense. Picking a time-based event instead means that your gun is less-likely to fire based on a freak anomaly, and regression-to-mean makes it safer. Yes, the time-based trigger must be informed by a dollar amount, but shifting to a date is a way to let things average out a bit more.
Imagine two people in 2007. Mr. A, making reasonable projections for investment returns, sets up his trigger for 2009, and Mr. B, with a smaller portfolio, sets up his trigger for 2014. Mr. A's portfolio on trigger-day was surely much smaller than he expected it to be, while Mr. B's portfolio was probably right around his projections. But Mr. A's retirement is probably going as planned, and much better than if he spent another 5 years working in order to get to his originally projected $X amount.
If you accept that a time-based trigger makes more sense than a dollar-based trigger, then the idea of pulling any "tricks" to hit a particular dollar amount seems rather pointless, and shifting around asset allocation before and after retirement would actually prevent the time-based regression-to-mean from happening, since you'd be changing your mean around.
Furthermore, in your example, it makes no sense to ignore investment returns in your last two years prior to FIRE. I appreciate matchewed's reminder that saving is generally more important than investing for FIRE, but as the FIRE date approaches, investing
does become more important than saving. My investment returns over the last few years have been at least as important to my net worth increase as my added savings. And thus, if you're ignoring your investment returns when determining how "close" you are to FI time-wise, you're essentially deciding to work longer than necessary. Why work longer
and shift your investments more conservative for that period? Though I guess doing it that way, you essentially
are using a time-based trigger in disguise; instead of using time to smooth out market volatility, you're smoothing it out by
ignoring it!