'Nother Retirement Calc

[Vincenzo Corleone]

Here's another retirement calc I found.

https://calculator.ficalc.app/

 

[Dtail]

Playing around with it a bit, it appears to have similar results as Firecalc with similar inputs.

 

[Morgan22]

I like it for the most part.

What I do not like is that I cannot enter in a negative number for "additional withdrawals" starting in a certain year. (FireCalc allows this.) It would come in handy if a negative # could be entered for when you finish paying off a mortgage, etc.

 

[FANOFJESUS]

Not much of a change from 50% stocks to 100% stocks

 

[chassis]

meh Interface is too jumpy. Can't change start of retirement date/age.

 

[karen1972]

I like it for the most part.

What I do not like is that I cannot enter in a negative number for "additional withdrawals" starting in a certain year. (FireCalc allows this.) It would come in handy if a negative # could be entered for when you finish paying off a mortgage, etc.

For this tool I added an additional withdrawal starting year 0 for the remainder of my mortgage. I think this is more accurate as mortgage doesn't go up with inflation, so even though your solution gets you close it would have inflated the mortgage payment over time and the mortgage thing is one of the few things I disliked about firecalc.

 

[Out-to-Lunch]

I like it pretty well!

 

[Dtail]

For this tool I added an additional withdrawal starting year 0 for the remainder of my mortgage. I think this is more accurate as mortgage doesn't go up with inflation, so even though your solution gets you close it would have inflated the mortgage payment over time and the mortgage thing is one of the few things I disliked about firecalc.

One can accomplish the effect of a non inflation adjusted mortgage in Firecalc by using the Pension section as an expense on one line with the start date and income on another line with the end date while NOT checking the inflation adjusted box.

 

[jamesplease]

Hi all! Creator of FI Calc here. Let me know if there are any questions you have or feature requests. I'm open to any feedback/questions you may have.

meh Interface is too jumpy. Can't change start of retirement date/age.

If you have time, I'd be curious to hear more about what you mean by the interface being too jumpy. If you dislike some of the motion, you can tone it by going into the User Preferences and enabling "Reduce motion". Let me know if that helps, or if you have other concerns about jumpiness I'd be happy to take a look and try to resolve them.

----

Also, sorry I am new to this forum so I don't know how to multi-quote, but regarding the mortgage / negative additional withdrawal conversation, the solution that karen1972 provided above is the most accurate way to account for mortgages as they aren't inflation adjusted. For that reason you'd likely want to add it as an Additional Expense rather than adding it into your base withdrawal to get the most accurate calculation.

 

[Out-to-Lunch]

Welcome, James!

I spent a fair amount of time with it this morning, after my note above. I must say I am very impressed with your effort. It is very thorough and very professional. It allows investigating a lot of parameter space very quickly.

One question: I assume that the graphs represent a (smoothed) histogram of, say, ending portfolio values. Is that correct?

You label the peak of that distribution "the median." If I am correct about what the graph means, then I don't think that is the median. That would be "the mode," right?

 

[Retired Expat]

Like most Monte Carlo calculators. No provision for investment real estate other than putting in anticipated net income and the property asset but RE growth rarely comes close to equities growth. Although rental return plus growth can... not too impressed.

 

[latexman]

Hi all! Creator of FI Calc here. Let me know if there are any questions you have or feature requests. I'm open to any feedback/questions you may have.

Hi jamesplease,

Can you give us a biography, please? Maybe put it here:

https://www.early-retirement.org/forums/f26/

Thanks!

 

[chassis]

Hi all! Creator of FI Calc here. Let me know if there are any questions you have or feature requests. I'm open to any feedback/questions you may have.



If you have time, I'd be curious to hear more about what you mean by the interface being too jumpy. If you dislike some of the motion, you can tone it by going into the User Preferences and enabling "Reduce motion". Let me know if that helps, or if you have other concerns about jumpiness I'd be happy to take a look and try to resolve them.

----

Also, sorry I am new to this forum so I don't know how to multi-quote, but regarding the mortgage / negative additional withdrawal conversation, the solution that karen1972 provided above is the most accurate way to account for mortgages as they aren't inflation adjusted. For that reason you'd likely want to add it as an Additional Expense rather than adding it into your base withdrawal to get the most accurate calculation.

Thanks, that is better. The default motion is annoying.

 

[jamesplease]

Welcome, James!

I spent a fair amount of time with it this morning, after my note above. I must say I am very impressed with your effort. It is very thorough and very professional. It allows investigating a lot of parameter space very quickly.

One question: I assume that the graphs represent a (smoothed) histogram of, say, ending portfolio values. Is that correct?

You label the peak of that distribution "the median." If I am correct about what the graph means, then I don't think that is the median. That would be "the mode," right?

Thanks a ton for the kind words.

Under-the-hood, I am using p-quantiles for the data displayed in the charts. The p values used are 0, 0.05, 0.25, 0.5, 0.75, 0.95, and 1. What this means is that the value at the "peak" of the graph is actually the median: half of the end portfolio values appear above that value, and the other half appear below it.

I understand how it may seem to represent the mode, though, because it is not always the case that the median would be at the peak of a probability distribution graph. This is an artifact of how I am rendering the charts right now, which is still just a "first pass" at visualizing the data. I have it hard-coded to always render the peak at the median, and every chart gets the same curve (adjusted for where the other p-quartile values land). It's been on my to-do list to do a second pass to make this more geometrically meaningful.

Because of your comment I'm going to move that on up in my to-do list, so keep an eye out for v2 of the charts! Thanks so much for asking about this!

Like most Monte Carlo calculators. No provision for investment real estate other than putting in anticipated net income and the property asset but RE growth rarely comes close to equities growth. Although rental return plus growth can... not too impressed.

This isn't a Monte Carlo calculator, although it is pretty similar. It's more accurate to say that it's a backtesting calculator that uses historical data for its calculations (as an aside, an update to introduce a Monte Carlo mode is in the works!)

One of the ways that you can support real estate right now is through Additional Income. There's a guide explaining how that works here: https://ficalc.app/configuration/additional-income/

This may not be enough, though...if that's the case, I'm always open to hearing suggestions on how this could be made more robust.

One of the challenges with real estate is that there is no historical data for all real estate in every geographic location in the U.S. dating as far back as the stocks and bonds data goes (1871). And it makes things a bit muddy when you start merging real data with "fake" data. It's similar to the difference between running a calculation assuming 10% yearly stocks growth vs. using actual historical data. Even though it may average out to be the same, it's the details of how it reaches that average that are often where interesting insights come from.

Anyway, feel free to drop suggestions if you have any. I'd love to add a super robust real estate solution into the calculator!

Hi jamesplease,

Can you give us a biography, please? Maybe put it here:

https://www.early-retirement.org/forums/f26/

Thanks!

I'd love to, but before I do, I'd like to know it you still think I should if:

- I don't make my personal FI numbers public
- I'd like to do something more substantial than a self-promo post that just says, "Hi all, I swung by because my app was being discussed here!"

Do you still think it's worth posting a bio there? I'd be happy to if you do!

Thanks, that is better. The default motion is annoying.

I'm glad the setting worked for you! Just let me know if anything else is annoying or if you have any other thoughts and suggestions. I love hearing what people think of the calculator!

 

[FANOFJESUS]

Both 50/50 and 90% stock 10% cash have a success rate of 95% Could that be right?

 

[jamesplease]

Both 50/50 and 90% stock 10% cash have a success rate of 95% Could that be right?

Perhaps...let's see what other trusted resources have to say about a 50/50 allocation:

FIRECalc
95.00% success rate
Link

Trinity Study (1926 - 1995, inflation-adjusted)
95% success rate at 4%

Bengen (1926-1976)
Figure 2. 50% stocks lasts a "minimum of 30 years", which one might interpret to mean a 100% success rate (this is admittedly a bit more opaque than the others).

A similar investigation can be done to see the similarities in the 90/10 allocation.

But why are there differences at all between these resources? Three main reasons:

1. Different data sources
2. Different time horizons (what is the date range used in the calculation? 1871-2020? 1926-1995? etc.)
3. Different algorithms for calculating what occurs in a year

I use Shiller's data source, which is available for free here. I believe FIRECalc use the same. My interpretation of the Trinity Study's source is that it uses an expensive data set from the SBBI Yearbook, and it may even be modified in some way. Bengen also used the SBBI Yearbook, I believe.

On the algorithm side, I prefer FI Calc's way of computing a year, which is documented here (https://ficalc.app/how-it-works/one-simulation-year/).

 

[jamesplease]

Both 50/50 and 90% stock 10% cash have a success rate of 95% Could that be right?

I posted a fairly long reply, and it showed up here, but then it disappeared for some reason...I don't have time to post it in full again, but surprising as they may be these results are consistent with FIRECalc, the Trinity Study, and Bengen's analysis at 30 years and a 4% success rate.

 

[FANOFJESUS]

I posted a fairly long reply, and it showed up here, but then it disappeared for some reason...I don't have time to post it in full again, but surprising as they may be these results are consistent with FIRECalc, the Trinity Study, and Bengen's analysis at 30 years and a 4% success rate.

Very true the only thing is the ending balance might be very different.

 

[pb4uski]

Both 50/50 and 90% stock 10% cash have a success rate of 95% Could that be right?

I posted a fairly long reply, and it showed up here, but then it disappeared for some reason...I don't have time to post it in full again, but surprising as they may be these results are consistent with FIRECalc, the Trinity Study, and Bengen's analysis at 30 years and a 4% success rate.

Yes, it is right. If you run the FIRECalc with 50/50 or 90/10 with the rest default assumptions in each case the success rate is 95%. Also, see graph of success rates at various AA's from FIRECalc below.

 

[jimbee]

^^^ Thanks pb4uski for running the numbers. Yep, the actual asset allocation doesn't have a big impact on success (within limits). It has a larger impact on how much is left over.

 

[pb4uski]

^^^ Exactly.

 

[latexman]

Do you still think it's worth posting a bio there? I'd be happy to if you do!

Sure.

 

[ERD50]

Both 50/50 and 90% stock 10% cash have a success rate of 95% Could that be right?

I am curious as to why you think this would be wrong? -ERD50