The Ultra Popular 7980 Daily Strategy - Why?

[Sad Al]

Why do so many follow #7980?
(See TSPCALC.COM if you are unfamiliar with the daily strategies).

Sure, the return is good at Mean of 36.6%, but with a Standard Deviation of 16.05, the Sigma is just above 2.2. That means you can be relatively sure (99.9%) that you won't have more than one loss every ten years, or about three or four times in your investing years. Conversely, you can be certain every year that there is a 0.14% chance of a negative return. There are several strategies with a higher return and a lower SD, for example #26845 with a mean of 37.51 and SD 15.72 sigma = 2.4

I'm willing to give up a few percentage points of returns for a higher sigma ratio. (Note: Mean divided by Standard Deviation equals Sigma Ratio). I've found that a Standard Deviation of about 5.5 or lower can give you a 6-Sigma ratio, which virtually eliminates the possibility of a down year.

For example, limiting SD to 5.5, I find that strategy #28862 yields 33.05% Mean, with a Standard Deviation of 5.33, which computes to a 6.2-Sigma ratio. Strategy #7980 gives you 3.3% more yield, but the mathematical safety of the former strategy gives you a likelihood of a negative year somewhere around 1 in a million. Is the extra 3% yield worth the margin of safety? If you are young, perhaps it is.

Note: Following #24823 - CAGR 30.12% - SD 1.78 - Sigma = 16.92. This should be ultra-safe. I'm retired, and willing to give up 6-7% yield for an easy night of sleep.

[Hoss2112]

It's very simple: Greed.

[crondanet5]

You both have a negative return for the year. Based on the performance of whatever seasonal you are following do you still believe you will have a positive gain of significance for the year 2018? If so, what leads you to think this and how will you achieve that positive return?

[TSPNerd]

It's very simple: Greed.

Greed? As in, I want to be able to retire comfortably and not have to worry whether the government gets rid of FERS. Then yes. I'm greedy.

Or, more accurately, when the earlier strategies were developed people were still trying to maximize their average return as the daily seasonal strategies were brand new. After the maximum average return was achieved, people started focusing on lowering sigma and strategies that only used specific funds.

[ewok55]

You both have a negative return for the year. Based on the performance of whatever seasonal you are following do you still believe you will have a positive gain of significance for the year 2018? If so, what leads you to think this and how will you achieve that positive return?

QFT ... This is the danger of "Torturing Data," and how statistics can mislead people to thinking that there's a "99.9% chance" of these strategies performing so greatly in the future. Standard deviation doesn't take into account family-wise test error, and creating a few, a posteriori seasonal mix patterns while denying any selection bias is foolhardy at best.

[SnareMV17]

Ewok55, using your logic, you just made the case that any statistic ever is essentially invalid and meaningless. I think you’ll find that you are mistaken.

Crondanet5, you pretty much miss the entire appeal of the daily strategies for the lamen. You seem a bit cynical about it too which is too bad. My question is what do you consider a significant return? And, it would be foolish even given past results to assume that a down January and February eliminates any possibilty of a “significant” return for the remaining 10 months of 2018.

[Hoss2112]

It's very simple: Greed.

Greed? As in, I want to be able to retire comfortably and not have to worry whether the government gets rid of FERS. Then yes. I'm greedy.

Or, more accurately, when the earlier strategies were developed people were still trying to maximize their average return as the daily seasonal strategies were brand new. After the maximum average return was achieved, people started focusing on lowering sigma and strategies that only used specific funds.

The balance of greed and risk usually determines the strategy a person chooses.
The higher the risk, the higher the potential reward. 30559 with a CAGR of 36.89 and a standard deviation of 31.69 would probably appear too risky to someone utilizing 31312 which has a CAGR of 28.41 and a standard deviation of 0.78.

[evilanne]

mmmmmbeer did a great sales job last year for 7980, but I think he moved to something else now

[jlozano042]

mmmmmbeer did a great sales job last year for 7980, but I think he moved to something else now

Yeah he did .... now I'm trying to return mine and he wont take it back!

[Sad Al]

Great question, Crodanet! I am looking for input, pos and neg. I believe it will still sustain 30% this year (I'll be happy with half that) because of past performance, and because of statistics/mathematics. In past years the strategy I am using has had negative months, but came back with strong returns later that still resulted in a positive 30% plus return. 2009 is the closest performance, where it was down 9.5% in January and February, suffered another 4% loss in a later month, but still yielded 30.4% that year. Nearly every year has had some negative months but still averaged 30% for the year. I believe the statistical average of past returns indicates a very strong likelihood of future returns, based on mathematical predictions.

[Sad Al]

QFT ... This is the danger of "Torturing Data," and how statistics can mislead people to thinking that there's a "99.9% chance" of these strategies performing so greatly in the future. Standard deviation doesn't take into account family-wise test error, and creating a few, a posteriori seasonal mix patterns while denying any selection bias is foolhardy at best.

Ewok55, can you further explain family-wise test error? Do you feel such selection bias negates the validity of the computations and patterns created in TSPCalc? (Still searching for "what's the catch?" while I follow something that has obviously done a lot better than my nebulous guessing). Thanks.

[cashworth]

I agree with the general sentiment of Ewok. That's why i run about 25 strategies side by side to determine which moves are solid and which are negotiable. For example, every strategy I watch moved to G fund today, to me that's a clear seasonal signal. Another example, last week some went to G, some went to F, and some stayed in S. For me, that's not a complete picture so I factored in trend analysis and stayed in.
I also dont love the S fund right now, so I may split some off to C a/o I.
Just my 2 cents.

[jlozano042]

I agree with the general sentiment of Ewok. That's why i run about 25 strategies side by side to determine which moves are solid and which are negotiable. For example, every strategy I watch moved to G fund today, to me that's a clear seasonal signal. Another example, last week some went to G, some went to F, and some stayed in S. For me, that's not a complete picture so I factored in trend analysis and stayed in.
I also dont love the S fund right now, so I may split some off to C a/o I.
Just my 2 cents.

That is usually how I do business, just not that many strategies yet.

I tend to follow this line of advice ...

[mindofmush]

It's very simple: Greed.

It doesn't matter if you think greed is good or not, the word "greed" is a negative term and greedy people are considered baaaad. Only the "rich" are "greedy" because they only got rich by depriving some poor person.

Seeking a good return is smart and proper stewardship of your family's future. Our CNN & CNBC talkings heads will tell you that your choice of 28% return is greedy and 25% more than you deserve and could only come from stealing the food from the mouths of babes.

If you make a good return and are conservative, you're smart; if you're liberal, you're greedy.

Your choice.

[fordest]

mmmmmbeer did a great sales job last year for 7980, but I think he moved to something else now

Here is the real answer to the original question however. Anne got it right. When TSPCalc began gaining in popularity, the strategies were limited to around the 11000s . Most people were following the leader. #10617 I believe. It had a mean of 38ish% and a StD of 26ish and nobody was talking about that aspect. A few people started noticing the 100+% 2008 and pulled it out as an outlier. That made those leaders no longer look very appealing to a few. mmmmmmbeer was very vocal about his new 7890 which had a lot less variation and appreaed more stable. Enter 12squared and his explanation of sigma and StD in general. Huge sigma strategies started being created. But 7980 and 10617 remained very popular. There are probably still a lot following the 10617 or the current leader as well as the 7980.