I am a new user and having difficulty understanding the Earliest Safe Retirement Date results. We have completed various scenarios, each with 90 to 100% success rates. I am 64 and using a desired retirement age of 65 (end of this year). My wife is not working any longer. When running the optimization I'll get a date where I retire at 72. Run it again and it says 66. Run it again and it is 70. Each time I get these wild variations, which makes no sense to me when my MC analysis has >90% success rate. Is there an explanation for this. How do I know which outcome is valid? Any help would be greatly appreciated.
Joseph, I agree that this amount of variation makes the value questionable. I think the reason is two-fold: the Monte Carlo algorithm contains a certain amount of randomness by its nature and the retirement date optimization algorithm stops looking for a better solution when the success rate of a given retirement date is 90% + or - 5%. With that said, though, I'm going to spend some time investigating the behavior of this algorithm and see if I can find a way to tighten it up a bit. If you'd be willing to email me an export file, that would be useful to me to test this with your actual case.
Stuart
@smatthews51 Hi Stuart - you may recall I had a similar problem last year with my Monte Carlo analysis results. CSS spend was varying by quite a bit (up to $16k annual difference) from scenario to scenario even when all three scenarios were configured the same. I saw happen this again last week.
@jjmcleer Was this ever resolved? I am having the same problem, two years later.
@afranklin Apparently not. May I ask how much variation you're seeing from one run to the next?
Thanks,
Stuart
@smatthews51 Using Monte Carlo analysis, the success rate is 100% in 2056, but drops to 99% in 2057. Using historical data, the success rate is 100% in 2059, and 95% in 2060. These are both based on retirement at the end of 2025, and we are planning to survive through 2060. The calculated earliest safe retirement date (in the optimization area) is 11/26/2033. I'm wondering what the cause of the discrepancy is.
@afranklin Hmmm, interesting. Would you mind sending me (mail@pralanaconsulting.comn) an export file so I can investigate this?
Thanks,
Stuart
