Early Retirement & Financial Independence Community

Early Retirement & Financial Independence Community (http://www.early-retirement.org/forums/)
-   FIRECalc support (http://www.early-retirement.org/forums/f36/)
-   -   FIRECalc glitch? (http://www.early-retirement.org/forums/f36/firecalc-glitch-85186.html)

mistermike40 01-27-2017 11:54 AM

FIRECalc glitch?
 
1 Attachment(s)
I ran a "what-if" scenario using 95%... I'm attaching a picture of the results. It should have chose $100,605 (a little greater than 97%) instead of $83,837 (which is called in in the results).

Has anyone seen this before (and know what might cause it)? Other "what-if"s seem to work perfectly.

Thanks in advance!

CardsFan 01-27-2017 12:24 PM

It looks like you actually searched for 100% success, from the wording above the graph. But I am not sure why it did not say $92,221

mistermike40 01-27-2017 12:39 PM

Thanks for your reply CardsFan! I did choose 95% though... it states it correctly in the first line under "FIRECalc results".

And you're right... even if I had chose 100% I would think it would indicate $92,221

Spudd 01-27-2017 12:47 PM

I wonder which is wrong, the graph or the words!

SecondCor521 01-27-2017 01:00 PM

@OP, the investigate tab does contain the slight disclaimer "(usually within 1%)". It is possible that the goal-seek function, given your particular inputs, can't get within 1% of 95% but can get within 5% of 95% (i.e., 100%). I'm not sure why that would be exactly...my guess is that it is an iterative solver that sometimes can't zoom in quickly enough sometimes, so it gives a result that is "somewhat close" quickly rather than a "really close" result slowly.

Usually when I have done searches of this type it will get within 1%. This is the first time I've seen it not do so. My guess as to why is that you're looking at a relatively high spending rate, and it is possible that the site is tuned towards more median spending levels. As an experiment, you could divide all of your inputs by 2 (divide income by 2, spending by 2, assets by 2, etc.) and see if it does better with the investigate function. (Just don't forget to double the resulting spending level again ;-) )

@mistermike, the result of the investigate window will always show the result ($83,837) in the middle and a few data points above and below. The fact that the next higher data point also hit 100% was probably an artifact of it not hitting very close to 95% in the first place.

mistermike40 01-27-2017 06:12 PM

Quote:

Originally Posted by Spudd (Post 1831470)
I wonder which is wrong, the graph or the words!

I think the graph is correct. When I do the reverse (input a spending amount instead of a success %), it gives 100% success at 92K and 97.3% at 100k.

mistermike40 01-27-2017 06:16 PM

Quote:

Originally Posted by SecondCor521 (Post 1831483)
@OP, the investigate tab does contain the slight disclaimer "(usually within 1%)". It is possible that the goal-seek function, given your particular inputs, can't get within 1% of 95% but can get within 5% of 95% (i.e., 100%). I'm not sure why that would be exactly...my guess is that it is an iterative solver that sometimes can't zoom in quickly enough sometimes, so it gives a result that is "somewhat close" quickly rather than a "really close" result slowly.

Usually when I have done searches of this type it will get within 1%. This is the first time I've seen it not do so. My guess as to why is that you're looking at a relatively high spending rate, and it is possible that the site is tuned towards more median spending levels. As an experiment, you could divide all of your inputs by 2 (divide income by 2, spending by 2, assets by 2, etc.) and see if it does better with the investigate function. (Just don't forget to double the resulting spending level again ;-) )

@mistermike, the result of the investigate window will always show the result ($83,837) in the middle and a few data points above and below. The fact that the next higher data point also hit 100% was probably an artifact of it not hitting very close to 95% in the first place.

I'll give that a try but dividing everything by two will keep the same spending rate. I tried the reverse and input the spending amount instead of the success % and it gave me the results on the graph.

I've used FIRECalc often in the past few years and never run across this situation.


All times are GMT -6. The time now is 06:09 PM.

Powered by vBulletin® Version 3.8.8 Beta 1
Copyright ©2000 - 2019, vBulletin Solutions, Inc.