TSP Calc Daily Strategy challenge

[dougellen1]

Is it possible someone can extrapolate the tspcalc daily data strategies back to 1989? It would give us a better MACRO sample set on this data. I know this is a huge task.....I'm just asking. I would do it but can't figure it out. I will donate funds to TSPCalc to assist.

[mjedlin66]

Is it possible someone can extrapolate the tspcalc daily data strategies back to 1989? It would give us a better MACRO sample set on this data. I know this is a huge task.....I'm just asking. I would do it but can't figure it out. I will donate funds to TSPCalc to assist.

You could do it for S and C fairly easily by using the underlying index. This would not be perfect but it would be extremely close.

You could not do it for the I fund, since that fund does not appear to track an index.

The best you could do for G fund is to assume a linear rate. We could take the monthly gain for G and divide it up by however many days are in the month. This would be pretty close.

F fund approximation based on AGG should be fairly accurate.

[mindofmush]

I wouldn't mind a strategy without the I fund since I have difficulty researching the ups and downs of the international market, I could do without it.

[evilanne]

Someone could back into the funds prior to the conversion to $10 share prices in 2003. You could probably add a few more years with S & I (introduced in ~2000) using the monthly data & benchmarks to approximated the daily prices. Another option would be just to back into the 3 primary funds and use the C Fund prices for S & I fund since it was the only stock fund available prior to the introduction of S & I or the share prices & S&P 500 daily data is readily available. This would be a little less time intensive but I'm not sure the effort would provide any more meaningful results than what mjedlin created with TSPcalc.com

[hightide]

I wouldn't mind a strategy without the I fund since I have difficulty researching the ups and downs of the international market, I could do without it.

The I fund is definitely confusing and hard to track. But with over 22% gains this year, I wouldn't ignore it completely.

[Midway]

Mjedlin66, if we could identify four ticker symbols that have data available back to 1989, could they easily be substituted in your calculator for the index funds?

[mjedlin66]

Mjedlin66, if we could identify four ticker symbols that have data available back to 1989, could they easily be substituted in your calculator for the index funds?

Yes. I could copy and paste everything I have now to a new calculator with a new database, and import index funds. This assumes you want the same TSP trading rules to apply. I could do this in maybe 10 hours.

Then you would have a TSP Daily Calculator, and an Index Fund Daily Calculator. I would not mix them.

It would not have the automatic update capability, at least not right away.

[dougellen1]

I have no affiliation with mjedlin66. I really appreciate the work that has been done setting up TSP Calc....and it is potentially benefitting many folks on this site. With that said, I hope people can contribute some money to TSP Calc to compensate this person's efforts.

[MakeMe$$$$]

Perhaps I'm not smart enough but...

How much relevance is there in "ancient" history? There are so many dynamic changes in politics, market players and such over 20 years I just don't see a real value. I would think a 10-15 year rolling analysis would be more appropriate for seeing a pattern in relatively current context.

I would suggest that using data from 30 years ago can skew once perception of expectations.

Just my 1/50th of a dollar.

[hightide]

I agree.

[mindofmush]

I wouldn't mind a strategy without the I fund since I have difficulty researching the ups and downs of the international market, I could do without it.

The I fund is definitely confusing and hard to track. But with over 22% gains this year, I wouldn't ignore it completely.

I agree that 22% this year is great especially compared to long term growth of about 8% but that reinforces the idea of which is more important:

Confidence in a statistical model which requires 30 years of data (small sample minimum = 30)
or
Current market performance and its effects on a strategy for next year (ie I fund performance high because Europe's high growth emerging from this last recession and/or F fund performance expected to be lower due to rising interest rates)

Since the Daily Strategies lend themselves to a more "active" management, I can understand that creating a new strategy every year based on the last 3 years of data would be the most relevant plan and probably the highest performing.

I'm still interested in comparing a strategy of 30 years that survived 2 recessions to a more relevant high performance strategy. (And I'm fearful of missing those economic indicators that predict a major recession)

no money, no life

[dearsmiths]

I have no affiliation with mjedlin66. I really appreciate the work that has been done setting up TSP Calc....and it is potentially benefitting many folks on this site. With that said, I hope people can contribute some money to TSP Calc to compensate this person's efforts.

I'm with my TSPCenter friend, EllenDoug, I will gladly provide more money to mjedlin66 if he can do as suggested. Count me in.

Dearsmiths

[bamafamily]

Ok..so let's say we go ahead and ask for this tspcalc improvement...
What funds do we use?? Mutual or ETF? We need to come to a mutual agreement so he only has to do this once...Here is my uneducated very rough cut based on Morningstar website..
C Fund - FUSVX or VOO
S Fund - FSEVX or VXF
I Fund - FSIVX or IEFA
F Fund - VBTLX or AGG
G Fund - no clue....

[maui21ice]

Then you would have a TSP Daily Calculator, and an Index Fund Daily Calculator. I would not mix them.

Ah yes, THANK YOU. I was going to say, if this was a combined project into the TSP calculator and threw off all the numbers you had now with index funds and the likes, I would definitely NOT be a fan of it. But if it is 2 separate calculators, that would be cool. I would definitely hope the TSP one is the main focus since that's why we're all here!

Also, I'm definitely down for helping to donate cause! You've done and awesome job so far, and I know many many many people appreciate it!

[12squared]

I agree that 22% this year is great especially compared to long term growth of about 8% but that reinforces the idea of which is more important:

Confidence in a statistical model which requires 30 years of data (small sample minimum = 30)
or
Current market performance and its effects on a strategy for next year (ie I fund performance high because Europe's high growth emerging from this last recession and/or F fund performance expected to be lower due to rising interest rates)

Although statistical confidence increases (i.e. margin of error in the mean decreases) as the number of samples increases, there is nothing magical about having 30 data points. margin of Error in the Mean decreases as the number of samples increases. However, the MoE also increases as the Std Dev increases.



However, if you divide the MoE by the Std Dev, you will find that the ratio depends only on the number of samples.

Since the Daily Strategies lend themselves to a more "active" management, I can understand that creating a new strategy every year based on the last 3 years of data would be the most relevant plan and probably the highest performing.

I am curious as to how recent data would be used, if not statistically. I am puzzled as to why using only 3 years of recent data would be better than 5, or 9, or 13.

Any strategy other than Buy & Hold is a form of active management. Daily, Monthly and any other "seasonal" strategies are all driven by predetermined timetables, not current market conditions. They are all based on historical data. Daily strategies differ from the Monthly variety in only two ways: use of more than a single IFT per month, and are not limited to making the IFT on the 1st or last trading day of the month.

I'm still interested in comparing a strategy of 30 years that survived 2 recessions to a more relevant high performance strategy. (And I'm fearful of missing those economic indicators that predict a major recession)

Have you compared any of the Daily strategies to the classic Monthly strategies using only the 2004 -16 data?
http://tspcenter.com/seasonal.php