Firecalc returning vastly different numbers with one minor change...

[stevemac]

Problem: Changing "Spending" amount on the first page by $1,000 and then back (over and over) resulted in a different Result every time, and by often-times huge numbers.

E.g., at $134k per year, for 221 cycles, the first there would be success for all but 1, and that one would be around -$174,000. Then changed to $135k/year, 3 cycle failures with lowest at say $400k. Change back to $134k/year, and the same number of cycles but now several failures, with a bottom of around $215k. Same thing happens going back to $135k (with now a low of around $700k and 5 failures).

Setup: Fairly standard, with SS starting in 2029, a major "Add" of $1M in 2026 (anticipating a house sale). Spending Model was Constant Power. Various selections used under "Your Portfolio". Each time, there were big disparities in the Results from one (e.g.) $134k to another.

Perhaps I've missed something about how this works, but I expected the same result for the same yearly Spending Amount each time I visited that one.

Thoughts? Thanks.

 

[REWahoo]

Be sure you are using only current or future dates. FIRECalc doesn't play well with any date inputs prior to the current year.

 

[USGrant1962]

That sounds like you are using "random performance" on "Your Portfolio". That is FIRECalc's version of Monte Carlo, but it is not very good - it does too few runs leading to wide variances in results.

 

[stevemac]

Thanks REWahoo and USGrant, but neither of these items are "in play": I'm only using current and future dates, and the issue appears for various options under "Your Portfolio" (although IIRC under "random performance" there might have been only one Result, but for the ~220 Results under the other options the Results kept changing each time I re-checked the same Spending Amount. Thanks again.

 

[SecondCor521]

Thanks REWahoo and USGrant, but neither of these items are "in play": I'm only using current and future dates, and the issue appears for various options under "Your Portfolio" (although IIRC under "random performance" there might have been only one Result, but for the ~220 Results under the other options the Results kept changing each time I re-checked the same Spending Amount. Thanks again.

For debugging these types of issues, it's helpful if you post the input data (click the link in the upper right corner of your results window and copy/paste).

 

[stevemac]

For debugging these types of issues, it's helpful if you post the input data (click the link in the upper right corner of your results window and copy/paste).

Thanks, I wasn't aware of that option, and will try that when I re-setup firecalc to try repeating the problem.

 

[cathy63]

Thanks REWahoo and USGrant, but neither of these items are "in play": I'm only using current and future dates, and the issue appears for various options under "Your Portfolio" (although IIRC under "random performance" there might have been only one Result, but for the ~220 Results under the other options the Results kept changing each time I re-checked the same Spending Amount. Thanks again.

Something is wrong with the number of results you're getting. The data only goes back to 1870, so the max number of cycles for a total market portfolio is 2021-1870-[yrs of retirement]. If you leave the default retirement length at 30 yrs, that's 121 cycles. Even if you reduce the retirement years to 1, you max out at 150 cycles.

For a 30 yr retirement, a mixed portfolio gives 49 cycles and a consistent growth portfolio gives 91 cycles.

The only way I can get as many as 221 cycles is to use the random option, so I suspect that it is somehow getting selected. Maybe something got scrambled in your browser and it's sending the wrong info to the server. If you close all your browser windows and start fresh in a new one, that will probably solve it.

 

[stevemac]

Something is wrong with the number of results you're getting.

That's a really interesting observation, one I hadn't thought about. I know for certain that Random wasn't chosen in various Results I checked, but you may be right about the browser doing odd things. It appears I should do a more scientific test and see if I get the same problem. I really appreciate your help on this.