09202017, 08:09 PM



1,040 posts, read 484,865 times
Reputation: 1435


anyone have suggestion for online Monte Carlo simulation calculator? thx

09202017, 09:05 PM



Location: Florida
4,361 posts, read 3,698,498 times
Reputation: 4095


Try online mutual funds and brokers

09212017, 02:33 AM



71,524 posts, read 71,712,424 times
Reputation: 49105


fidelity's planner is one of the best and uses monte carlo simulations .
firecalc's default is actual historical but it has a selectable monte carlo option . they are very very close in results.
why so hot on monte carlo ? i am not a lover because when it comes to investing the time frames leading in and leading out can be just as important as the time frame you are looking at .
as an example , the youngin's tell me all the time how lucky i was to be in the markets when the greatest bull in history started back in the 1980's .
the real time data though showed a different story . it was all well and good the greatest bull arrived but the time frame leading in was horrible . it was near impossible to have grown any savings as markets were dead for years , bonds were crushed and inflation was soaring .
there were few 401k's as well at the time . so it was really only the wealthy who had money to have the bull act upon that made use of it ..
so pulling things out of sequence without the lead in or after effect can be a bit misleading to how events really play out .
all failures in draw rates we have had occurred the same exact ways ,with the same lead in's being the death blow . the failures occurred because real returns fell below 2% as an average over the first 15 years of a 30 year retirement ..
Last edited by mathjak107; 09212017 at 02:47 AM..

09212017, 04:10 AM



71,524 posts, read 71,712,424 times
Reputation: 49105


here is what michael kitces found when using monte carlo simulations vs historical data .
"For instance, when comparing a Monte Carlo analysis of 10,000 scenarios based on historical 60/40 annual return parameters to historical returns, it turns out that 6.5% of Monte Carlo scenarios are actually worse than even the worst case historical scenario has ever been! Or viewed another way, a 93.5% probability of success in Monte Carlo is actually akin to a 100% success rate using actual historical scenarios! And if the advisor assumes lowerreturn assumptions instead, given today’s high market valuation and low yields, a whopping 50% to 82% of Monte Carlo scenarios were worse than any actual historicallybad sequence has ever been! As a result, despite the common criticism that Monte Carlo understates the risk of fat tails and volatility relative to using rolling historical scenarios, the reality seems to be the opposite – that Monte Carlo projections show more longterm volatility, resulting in faster and more catastrophic failures (to the downside), and more excess wealth in good scenarios (to the upside)!"
https://www.kitces.com/blog/monteca...ricalreturns/

09212017, 04:22 AM



Location: Central Massachusetts
4,800 posts, read 4,845,678 times
Reputation: 6379


Quote:
Originally Posted by mathjak107
here is what michael kitces found when using monte carlo simulations vs historical data .
"For instance, when comparing a Monte Carlo analysis of 10,000 scenarios based on historical 60/40 annual return parameters to historical returns, it turns out that 6.5% of Monte Carlo scenarios are actually worse than even the worst case historical scenario has ever been! Or viewed another way, a 93.5% probability of success in Monte Carlo is actually akin to a 100% success rate using actual historical scenarios! And if the advisor assumes lowerreturn assumptions instead, given today’s high market valuation and low yields, a whopping 50% to 82% of Monte Carlo scenarios were worse than any actual historicallybad sequence has ever been! As a result, despite the common criticism that Monte Carlo understates the risk of fat tails and volatility relative to using rolling historical scenarios, the reality seems to be the opposite – that Monte Carlo projections show more longterm volatility, resulting in faster and more catastrophic failures (to the downside), and more excess wealth in good scenarios (to the upside)!"
https://www.kitces.com/blog/monteca...ricalreturns/

Extrapolating this line further. A person looking at higher failure rates and increased volatility will be more apt to over save in preparation for retirement (not a bad thing) or underspend in retirement. On the surface those do not seem like bad things but as a retiree I and my wife saved up so that we can enjoy retirement not fret and worry on how to make ends meet.

09212017, 05:01 AM



71,524 posts, read 71,712,424 times
Reputation: 49105


the point kitce's is making is it can be overly pessimistic , not good if you are trying to save up to a level you think is good and delaying because you are not quite there , or setting a lower spending budget initially than you need to . i it can be overly optimistic when things are better causing to high of a spending expectation .
it can be a double edge sword depending on view and assets .. i like fidelity's as it is stricter than firecalc . but if i was not in the position i was in it would have me spending less or waiting to retire building more savings.
fidelity uses not only monte carlo , but unlike the default in most planners it takes healthcare and long term care costs and accurately inflates them by far more than general inflation .
it currently figures 5.50% on both healthcare and long term care . it was 7.50% on healthcare and 5% on long term care but surveys show it came down .
the real story is using historical date like firecalc has actually left you with more money than you started 30 years later 90% of the time with a 60/40 mix and 2x what you started with 67% of the time as well as 3x what you started with 50% of the time .
so in reality using historical at 4% swr has been to conservative almost all the time so to go with something even more pessimistic makes little sense for many of us . odds are we will be taking raises along the way anyway .
Last edited by mathjak107; 09212017 at 05:16 AM..

09212017, 06:19 AM



1,040 posts, read 484,865 times
Reputation: 1435


Thanks MJ and others.....was just curious

09212017, 09:37 AM



Location: Haiku
4,060 posts, read 2,571,078 times
Reputation: 5983


The trouble with historical data for modeling is that it is limited  we only have 120 years or so of data. So people try to get better statistics by using Monte Carlo, which replaces historical sequences of data with analytical data distributions. It gives a false sense of providing a more robust result with better statistics. But it is unwarranted comfort  Monte Carlo cannot magically tell us what the future holds any better than using historical data. So I get a little suspicious of MC simulations because they can be abused and misinterpreted. It is important to realize that MC is just a guess, same as using historical data.

09212017, 09:49 AM



71,524 posts, read 71,712,424 times
Reputation: 49105


but all we need is the math from the worst case scenarios that exist .
the math says we need a 2% average real return the first 15 years to hold 4% .
what else do you need to test for ? all the back testing is used for now that we have powerful computers is to determine the math that caused the failures . we know all that now that kitces crunched the numbers .
we have the results of the failure , the math that caused it and how to monitor ourselves . it is basically done .
even if we had 1,000 more sample cycles that were worse , the math is the math and we already know what that math is that leads to failure . any condition that falls below 2% real returns during the first 15 years will have to have a pay cut or risk failure if drawing 4% inflation adjusted ..
it was different before the numbers were crunched and converted to a common denominator that caused all failures to happen .when we worked off a percentage of pass vs failure based on sample size more samples would be better. but today once we identified the conditions that caused the failure basically we are done.
think of it as analyzing all the diseases we ever had that had fevers that led to death .
once we learned that if the human body goes over xamount of temperature it fails , we can stop analyzing diseases to find the link between disease and temperature ..
Last edited by mathjak107; 09212017 at 10:15 AM..

Please register to post and access all features of our very popular forum. It is free and quick. Over $68,000 in prizes has already been given out to active posters on our forum. Additional giveaways are planned.
Detailed information about all U.S. cities, counties, and zip codes on our site: Citydata.com.

