Fun with Numbers

[FundSurfer]

imported post

2004 Average Daily Gain by Day of the Week
Fund Monday Tuesday Wednesday Thursday Friday

G 0.03% 0.01% 0.01% 0.01% 0.02%

F 0.00% 0.03% 0.01% 0.01% 0.03%

C 0.06% 0.12% 0.09% -0.04% -0.01%

S -0.02% 0.23% 0.10% -0.08% 0.11%

I 0.05% 0.09% 0.02% 0.07% 0.14%



2004 Maximum Possible Gain

(If you had picked the right fund to be 100% invested each day.)

149.34%

2004 Maximum Possible Loss

(If you had picked the worst fund to be 100% invested each day.)

-123.99%



What I didn't post was the standard deviation for those weekday numbers. As you might imagine, they were very high, which means highly variable.

However, I think we can say a couple things. G gives the penny most often on Monday. Interest accrues daily whereas the market only goes up/down on trading days. Therefore if the penny was due Sat/Sun/Mon, then Monday get it.

If we were to use this to allocate, we would have allocated:
Monday C, Tuesday S, Wednesday S, Thursday I, Friday I

Your return for 2004 would have been 30% if you'd used this formula.

My bet is that this method would not back test well. Probably very badly.

 

[jgpalmerdds]

imported post

Fund Surfer,

Your the man (or women) How long did it take you to figure all that out? Impressive.

 

[Rod]

imported post

This seems very similar to my "system" I was trying out back in Oct/Nov:

Monday: 100 G

Tuesday: 100 S

Wednesday: 30 G 50 S 20 I

Thursday: 100 G

Friday: 50 S 50 I


Of course, my missing some of the election rally threw a wrench into it.

 

[FundSurfer]

imported post

Spreadsheets are a wonderful thing. It allows you to play all sorts of fun games with numbers. I just took the table off tsp.gov; calculated daily percentage move; then sorted by day of the week; and finally summed the result. Actually didn't take that long. At some point in time I need to do this with a longer term data set.

 

[Rod]

imported post

FundSurfer wrote:

Spreadsheets are a wonderful thing. It allows you to play all sorts of fun games with numbers. I just took the table off tsp.gov; calculated daily percentage move; then sorted by day of the week; and finally summed the result. Actually didn't take that long. At some point in time I need to do this with a longer term data set.

I did the same thing using pen & paper. But it looks like we came up with the same results!

Believe me though, this "system" wasn't fairing well after the election. You can read more aboutthat on page 2 of my account talk. I believeI mention a thing or 2 about it.

I wonder when would be a good time to calculate this out again???

 

[FundSurfer]

imported post

Well my calcs used all of the 2004 data. I've been meaning to do the 2003 data, but haven't gotten to it yet. Rod, when was your calcs from?

 

[Rod]

imported post

They were from May-Oct of 2004.