Stomaching the Standard Deviation

[bamafamily]

Hey all,
So, I have been trying t find a happy place between Mean And STD Deviation that won't keep me up at night..(either with too high of a STD Deviation or too low of a Me). Just trying to find that right mix...
The awesome tool at Tspcalc lets us put in the amount of "volatility" we are comfortable with and find the highest return strategy....

E.g. I am currently looking at a max 9.9 STD Dev which then produces Strategy 12466 with a mean of 23.9% and actual std dev of 9.1... and only 47 IFTs...not sure if this is the one for me, but it fits my "volatility" lifestyle, if you will...

What about you? What is your STD Dev max that you are willing to throw your real money at?
Thx
Bama

[Travis1984]

#7980forlife

(Or until Beer finds a better one)

In all honesty, I think most of these strats are pretty close to the original "Seasonal," so you are adjusting your risk volatility by lowering the amount you play in the market. I don't look necessarily at a minimum amount, as many of the strata are still doing well in 2017. And yes, I'm 100% using my real TSP on 7980.

Whatever you decide...commit to it for a set designated amount of time!

[mmmmmbeer]

Yeah same as Travis, the 7980. Standard deviation of 16.67 and time in market of 58%, still fetching 35%+.

With a STD (heh) of 16.67 and 35% your looking at somewhere between 19-51% average. Uhhhh... thats a pretty sweet spread to cover man. I think Trav hit the nail on the head with it being the time in the CSI being the real thing we want to focus on. How much market volitilaty do we want to expose ourselves too?

The 7980 is only in the market 58% of the time, the rest we're sitting safe as kittens in the G & F.

Standard deviation at that high of a return rate I'm good with. I want to be safe as kittens and still making that big rate.

[bamafamily]

I agree with the numbers....and they are very good....you make a very good case....
just everyone has a limit and I suspect there are some people that are already retired that would be ecstatic with lower returns and lower volatility.....(say 12% and 4%)
It's just something I am internally fighting with....I might dip my toe in on one of the lesser strategies, realize the water is fine, and then jump in the deep end.....

As for the safety of the F fund...while it is much less volatile than CSI, it can still bleed you slowly if your not careful....

My subscription to my other TSP system is up in less than a month...thus the reason to make a decision.....

.....The 7980 is only in the market 58% of the time, the rest we're sitting safe as kittens in the G & F.

Standard deviation at that high of a return rate I'm good with. I don't want to be safe as kittens and still making that big rate.

[mjedlin66]

My subscription to my other TSP system is up in less than a month...thus the reason to make a decision.....

Some day, I am going to "solve" the TSP seasonal, and then tspcalc will have its own subscription service. There are like 250k possible seasonal strategies in each MONTH. If I run all 250k of them for each month of the year, then I can not only put together a strategy based on highest possible return, but I can also put together "Rolling Strategies".

A rolling strategy is this. Say we are looking at a 5-year rolling:

For your 2005 strategy, you use the strategy that would have yielded the highest possible return for 2000-2004. For your 2006 strategy, you use the highest possible for 2001-2005. And so on, and so on. I would compare a 3-year rolling to a 5-year rolling and a 7-year rolling.

Those strategies won't be added to the rest of the seasonal calculator. They will be available by subscription e-mail alerts only. The calculator would remain as it is, logging every user-generated strategy.

[nrialto]

I would pay for that subscription.

mjedlin66, is there a way to find which strategy does best, on average, by month?

For example; 1234 has the best average in November, 4321 has the best average in December, etc. I'm thinking forward to the holidays, specifically.

[Travis1984]

I would pay for that subscription.

mjedlin66, is there a way to find which strategy does best, on average, by month?

For example; 1234 has the best average in November, 4321 has the best average in December, etc. I'm thinking forward to the holidays, specifically.

This thought has been discussed before, but the problem is many months have roll over IFTs from the previous month (meaning you start the month in a fund) that may not be the fund for the previous months best...not sure if that makes sense.

[bamafamily]

......
but I can also put together "Rolling Strategies".
A rolling strategy is this. Say we are looking at a 5-year rolling:
For your 2005 strategy, you use the strategy that would have yielded the highest possible return for 2000-2004. For your 2006 strategy, you use the highest possible for 2001-2005. And so on, and so on. I would compare a 3-year rolling to a 5-year rolling and a 7-year rolling.......

I like that idea as well....Personally I think using the most data possible for the rolling would work better, but what do I know.... I really appreciate all the work you have put into the system...

Bama

[evilanne]

Hey all,
So, I have been trying t find a happy place between Mean And STD Deviation that won't keep me up at night..(either with too high of a STD Deviation or too low of a Me). Just trying to find that right mix...
The awesome tool at Tspcalc lets us put in the amount of "volatility" we are comfortable with and find the highest return strategy....

E.g. I am currently looking at a max 9.9 STD Dev which then produces Strategy 12466 with a mean of 23.9% and actual std dev of 9.1... and only 47 IFTs...not sure if this is the one for me, but it fits my "volatility" lifestyle, if you will...

What about you? What is your STD Dev max that you are willing to throw your real money at?
Thx
Bama

Everyone's risk tolerance is different but I think you are placing too much emphasis on Standard Deviation, which also limits your upside. Comparing your mix to mmmmmbeer and one other mix:

http://tspcalc.com/seasonal.php?ID=12466 1,848.78 SD . 9.06 CSI% 47.20 LO 14.04 HI 47.21 Mean 23.95

http://tspcalc.com/seasonal.php?ID=7980 . 6,557.90 SD 16.40 CSI% 58.00 LO 17.61 HI 78.80 Mean 35.89

http://tspcalc.com/seasonal.php?ID=12367 4,544.23 SD 12.83 CSI% 50.80 LO 14.76 HI 58.55 Mean 32.14

There isn't that much difference in the data on the low side of annual returns but the high returns are more substantial with higher standard deviation. With being in the market less than half the time, your returns will naturally be lower. Increasing time in market 3-4% can double your return.

Here are some other mixes with SD < 15%

Mix Link - Total Return - Standard Deviation - CSI% - Mean - Low Return%
http://tspcalc.com/seasonal.php?ID=12523 6,278.79 14.86 58.00 35.32 LO 18.44
http://tspcalc.com/seasonal.php?ID=665 . . 3,034.16 14.03 54.40 28.62 LO 11.47
http://tspcalc.com/seasonal.php?ID=5248 . 3,362.22 12.50 54.00 29.40 LO 15.02
http://tspcalc.com/seasonal.php?ID=4923 . 3,294.58 12.35 54.00 29.21 LO 14.46
http://tspcalc.com/seasonal.php?ID=12216 4,327.80 11.39 55.20 31.58 LO 13.89
http://tspcalc.com/seasonal.php?ID=578 . . 2,316.66 10.65 55.20 25.99 LO 10.56

[12squared]

Standard deviations are intended to describe a bell-shaped distribution of data, in which ~2/3rd of the data are within 1 SD of the mean. SDs can be skewed when you have an outlier and a small number of samples. The bigger the outlier (positive or negative), the higher the SD. It is these outliers that put these strategies at the top of the rankings.

Performance of #12767 in 8 of 14 years (2004-2016) is less than the mean. All 8 of these are within 1 SD of the mean. Only 3 of the other 5 years are within 1 SD of the mean. This is not a "normal" distribution.

Removing the highest year (2009) from #12767, reduces the mean & SD of the 2004-16 range from 38.38 & 24.29 to 33.35 & 16.89. Removing the lowest year (2005) increases the mean to 40.39, but reduces the SD to 24.22. Removing the two highest years (2008 & 09) reduces the mean & SD of the 2004-16 range to 29.90 &12.50.

It may be more helpful to look at the "low" year, rather that the standard deviation. On this basis, 12767 is better than 11006 because it has a higher "low" year (14.26% in 2005) than #11007 (13.60% in 2014).

[Octjan2]

I was looking at the 7980 since 2004 and the lowest return was 17.61%. Even if I average that for the next 10 years I will have close to 2 million. Incredible!

[bamafamily]

Thanks Anne.....
In all of my other portfolios, either the Sharpe or Sortino Ratio is calculated to get a better handle on the Risk Adjusted return of a certain portfolio....(Since I was just doing a quick test, I used the Sharpe ratio/Annual since Sortino utilizes negative returns in the calculations and most of the ones we are following do not have any negative annual returns....I could run monthly calculations but I figured annual would give me the initial picture I was looking for.)

For the 3 that were mentioned above, I calculated the following Sharpe ratios from 2004-2016(bigger is better):
07980 - 2.00
12666 - 2.31
12367 - 2.35

I may hunt other strategies down to find their Sharpe Ratios as well since that is what I usually use as a reference in my other portfolios....

Bama
P.S. These are very good numbers. You can read up about Sharpe Here.

....Everyone's risk tolerance is different but I think you are placing too much emphasis on Standard Deviation, which also limits your upside. Comparing your mix to mmmmmbeer and one other mix:

http://tspcalc.com/seasonal.php?ID=12466 1,848.78 SD . 9.06 CSI% 47.20 LO 14.04 HI 47.21 Mean 23.95

http://tspcalc.com/seasonal.php?ID=7980 . 6,557.90 SD 16.40 CSI% 58.00 LO 17.61 HI 78.80 Mean 35.89

http://tspcalc.com/seasonal.php?ID=12367 4,544.23 SD 12.83 CSI% 50.80 LO 14.76 HI 58.55 Mean 32.14

There isn't that much difference in the data on the low side of annual returns but the high returns are more substantial with higher standard deviation. With being in the market less than half the time, your returns will naturally be lower. Increasing time in market 3-4% can double your return. ........................

[bamafamily]

Went through some of the strategies.
Best Sharpe number I found was one Anne had mentioned previously.....

#12216
Sharpe: 2.59
CAGR: 31.09%
Mean: 31.58%
Std Dev: 11.39%
CSI% 55.2%

Not bad at all for the CSI%.....Might have a winner here for Bama.....

For the 3 that were mentioned above, I calculated the following Sharpe ratios from 2004-2016(bigger is better):
07980 - 2.00
12666 - 2.31
12367 - 2.35
I may hunt other strategies down to find their Sharpe Ratios as well since that is what I usually use as a reference in my other portfolios....
Bama
P.S. These are very good numbers. You can read up about Sharpe Here.

[mmmmmbeer]

Went through some of the strategies.
Best Sharpe number I found was one Anne had mentioned previously.....

#12216
Sharpe: 2.59
CAGR: 31.09%
Mean: 31.58%
Std Dev: 11.39%
CSI% 55.2%

Not bad at all for the CSI%.....Might have a winner here for Bama.....

For the 3 that were mentioned above, I calculated the following Sharpe ratios from 2004-2016(bigger is better):
07980 - 2.00
12666 - 2.31
12367 - 2.35
I may hunt other strategies down to find their Sharpe Ratios as well since that is what I usually use as a reference in my other portfolios....
Bama
P.S. These are very good numbers. You can read up about Sharpe Here.

Jump on in Bama... the waters fine!

[evilanne]

Went through some of the strategies.
Best Sharpe number I found was one Anne had mentioned previously.....

#12216
Sharpe: 2.59
CAGR: 31.09%
Mean: 31.58%
Std Dev: 11.39%
CSI% 55.2%

Not bad at all for the CSI%.....Might have a winner here for Bama.....

For the 3 that were mentioned above, I calculated the following Sharpe ratios from 2004-2016(bigger is better):
07980 - 2.00
12666 - 2.31
12367 - 2.35
I may hunt other strategies down to find their Sharpe Ratios as well since that is what I usually use as a reference in my other portfolios....
Bama
P.S. These are very good numbers. You can read up about Sharpe Here.

If Sharpe Ratio is a measure for calculating “risk-adjusted return” I don’t understand it given the 2 mixes and why 12216 would be better than 12367? Looking at the data:

G . F .. C . S . I CSI . . .Mean . StDev Sharpe
24 25 . 4 32 15 50.80 32.14 . 12.83 . 2.35 http://tspcalc.com/seasonal.php?ID=12367
16 29 . 4 37 14 55.20 31.58 . 11.39 . 2.59 http://tspcalc.com/seasonal.php?ID=12216
12367 has 4.4% less exposure to CSI stock funds, 4% less in F Fund with 216% greater return. It has 1.44% Greater Standard Deviation and 2 more IFTs but if your goal is to sleep at night, it seems like 12367 is less risky for a greater return