I've expressed skepticism about the usefulness of simplistic online tools, including but not limited to, Monte Carlo simulations. Getting meaningful results entails making accurate assumptions and using tools sophisticated enough to model them. This is quite difficult even for professionals, as the linked page (Wade Pfau) illustrates.
https://retirementresearcher.com/rise-not-rise-stock-allocation-retirement/Pfau and Kitces in 2014 found that rising equity glidepaths in retirement produced better results than conventional wisdom declining equity allocations. They "included three sets of capital market assumptions with a range of stock and bond returns" when simulating possible outcomes. But they hadn't built "in a mechanism to allow average bond yields to change over time."
Once their tool was enhanced to support this, and they "allow[ed] bond yields to gradually increase over time toward their historical averages", Pfau came to the opposite conclusion. Under current market conditions and assumptions, the traditional glide paths in retirement work better. He needed souped up tools, different starting assumptions, and a different model to reach this diametrically opposed conclusion.
Given how dependent results are on the model and the input assumption, how much faith can one place in the outputs of the "tinker toy" online tools? As the saying goes, they're for amusement use only.
I've chosen to highlight a minor aspect of the published work, because it serves to illustrate how unstable and unreliable conclusions that amateurs draw from relatively simple online tools may be. The core of the papers, concerning how one might modify allocations based on market valuations, is of more practical value.
Here's the later of the two Journal of Financial Planning papers discussed:
https://www.onefpa.org/journal/Documents/March2015_Contribution_Kitces.pdf
