Including tax-deferred accounts in smoothing - a possible third way

I read with great interest the posts from last spring on the effect of the constant withdrawal rates for tax-deferred accounts on consumption smoothing.

If I read correctly, you have most of the code in place in the CE to employ the tax-deferred accounts in consumption smoothing, but running such an analysis would be unsatisfactory because of computation time. (I am a programmer and I understand what it means for a problem to be computationally expensive and how that impacts the user experience.)

So here's my suggestion, or maybe it's a question. Could you treat this the same way you treat another computationally expensive report, the Monte Carlo simulation? If you did, a user could choose to run a regular report, a Monte Carlo report, or a "tax-deferred-smoothing" report. (Not a catchy name, I know.)

For many people, running the regular report will produce a satisfactorily smooth outcome and running the slower tax-deferred report would not be needed. But for others, like us, ES will provide even more help when all assets are employed in the smoothing process.

(BACKGROUND: Here's why I say ES would help us more if it could do this. Like a lot of federal employees, most of our savings is tied up in the government Thrift Savings Plan "TSP". We liquidated most of our regular assets to build a vacation home in Canada. Now, as I approach retirement at age 60 this year, practially all of the funds available to smooth consumption are in TSP. The ES report is helpful in many ways, but it doesn't do much smoothing over the near term in my case. In particular, it is difficult to run comparisons of taking SS early or late, because there is nothing available except the TSP to fill the money hole if I take SS late.)