Dave's Account Talk

[Dave M]

imported post

As the noon trading deadline approaches, the indices are up. Therefore I will let it ride.Being evenly distributed across the funds I am in position to benefit whichever way the market -- or the dollar -- bounces.

Dave

 

[Dave M]

imported post

I have made no changes and remain 40G 20C 20S 20I. For one thing I am on night shift this week and it is inconvenient for me to hang around untilthe noon trading deadline.For another I have no clear idea thatthe current weakness in the market is going to continue.

My index has dropped over 80basis points in the last ten trading days. I boughton the 2nd at 4339; it climbed to 4387 and was4304 at close Wednesday. (I have discarded the decimal point.) Every changeof 45 points equates to a gain or loss of $1000 for me so I am not happy at this time.

I shall hang in there and hope for good news to finish off the week. There is a lot of $$ out there looking for a home and eventually investorstired of being put off by high oil prices will re-enter the market which will be good for everybody. Good Luck to all Irish men and women!

Dave

 

[Dave M]

imported post

And so we end the week at 4293 down 41. We absorbed thebad oil news and are still standing but I think we can safely say, this quarter is shot growth-wise.It would be nice to stop the slide and reverse the trend beforethe end of the month, give us some momentum going into the 2nd quarter. Good luck, see you next Friday.

Dave

 

[Dave M]

imported post

The slide continues. Add the three stock funds together and discard the decimal point, the index stands at 4212 at close today, down 81 points for the week and 127 for the month,3%. This is 4% off the month's high, 4387achieved on the 7th.

On Jan 1 the index stood at 4272 so now we are eating into those wonderful gains we had in December. Some years are like this, I know, but I had such high hopes! I don't have 20 years to look forward to, so I feel I have to protect myself, and I will be staying in the G-fund for a while it looks like. Good luck everybody.

Dave

 

[Dave M]

imported post

Another down week, with the 3-fund index closing at 4214 on Friday, off by 5 points from the previous week'sclose. (I think I added wrong last week, or they amended one of the numbers while I wasn't looking.) Okay, so that is essentially unchanged which is an improvement I guess. However it was a volatile week, wasn't it. It opened Monday morning at 4219, sank to 4186 and then recovered to 4234 on the 31st before losing another 20 points on the lastday of the week.

What drove it down in the end would seem to be the grim forecast of high oil prices put out by somebody or another, and a poor showing in job-creation numbers for themonth just ended:Jitters, essentially. Yet as we have seen for a long time now, oil and stocks are negatively correlated, strongly.

So it looks to me as though we are going to require some time to get over the jitters and realize there is NOT going to be a crash and burn here in the West, and for portfolio managers -- the Big Boys with the billions -- to feel the pressure for positive results. For this reason I am not sanguine about the 2nd quarter, and look to the latter part of the year, maybe 4th quarter, for gains. (Last year I had almost all my gains in the 4th quarter, according to my TSP statements.)

Thus I remain 100G and will do so until this thing turns around. Patience pays. Good luck everybody.

Dave

 

[Dave M]

imported post

We closed on Friday the 1st at 4214. We closed today Friday the 8th at 4241 for a net gain of 27 basis points for the week. This is the first up-week in a month -- a very good sign!

But, we were down 25 points today from Thurday's close -- a bad sign. We lost in one day half of what it took all week to gain. This is sufficient deterrent for me, I will remain 100G for another week and see what happens. The market has not lost its jitters evidently.

On March 22nd we stood at 4238, very close to today's 4241. So we can say that this week clawed its way back to where we were two weeks ago, erasing those intervening losses.But where are the gains? I wanna see some actual gains before I take another plunge. Since the first of the year I have suffered losses which are just offset by my contributions so I am right where I was on Jan 1in terms of Net Asset Value, which is what I track.

I'll just say that when I was trading my IRA actively, back in the 80's, my broker and I took the stance that we were not interested in wringing every possible penny out of the market. When we had enough, we sold and left to others any further gains. That is coming back to me, now. I am content to ride thebroad movements, and I leave it to others to wring each penny out of these swings.

So if we have another good week, gain some momentum and get back another 50 points in the all-funds index, I maybuy back in. Good luck!

Dave

 

[pyriel]

imported post

Dave, I like that philisophy. I've been meaning to go long term but I am always getting sucker to analyze the market and move my funds. I'll rethink my strategy and look for a longer term with my TSP...

 

[Dave M]

imported post

It was not a good week, was it. In fact we have erased all the gains since lastyear and then some. Last Friday's close was 4241 and we closed today at 4108 for a loss of 133 basis points, 3% from where we stood.

Our1st quarter statements are available at TSP.gov. Personallyit was a disaster. Every move I made was 180 degrees out of phase except when Iretreated to the G-fund in March. My loss for the quarter wiped outhalf of mygain for2004. Fortunately my contributions made up for it and NAV is hanging in there unchanged.

Good luck in the coming week.

Dave

 

[Dave M]

imported post

Hey, what a week, eh? These 200-point days are are kind of nerve-wracking. Up, down, up, down, kind of like Cowboy riding a bucking bronco! (I avoid the word "bull.") That cartoon of the bear riding the bull is appropriate.

For the week the all-funds index closed up 39 basis points at 4147. In the last eight weeks we have had three up and five down and are now 192 points below where we were on March 2nd (4339). This is not a buy signal tome. Thissignals continued market jitters. Every day it is a new world, seemingly. One day inflation fears cause a big drop, next day inflation is "under control" and we rally. One day techs report a drop in earnings, next day techs lead the rally. It reminds me of somebodywho has hadtoo much caffeine! All I can see to do is watch and wait. To those of you fully invested, I can only admire your fortitude and wish you Good Luck!

Dave

 

[Dave M]

imported post

The bleeding continues. The close today was 4122 C+S+I, for a loss of 25 basis points for theweek. BUT...

If you got in and out on Monday and Friday, you could've gained a couple percent because these two days were up. Something told me on Thursday that the week's losses were extreme and there would be buying on Friday. So I moved 10% of my balance into the C fund and gained a percent on it. Before noon on Friday I snatchedit right back out, so I will start next week 100G again -- got a penny coming.

I had almost 16k on the line and made a couplehundred onmy gamble. It isn't much but it made up about 6% of the losses I sustained in the 1st quarter. In addition, that 1.2% gain on 10% of my balance will just about equal the $.01 coming this week in the G-fund, so I doubled the net return there. Do thisa few more times and I'll be getting somewhere. *smile*

Dave

 

[Dogdaddy]

imported post

The G-Fund paid the penney today..never can tell when it'll be the 5th or 6th day, but I'll take it! Other Funds are much too scarey for me right now.........

 

[Dave M]

imported post

Today worked out very well. Together with last week I have cut an eighth off my accumulated Q1 deficit. At noon I pulled my chips off the table in anticipation of profit-taking Thursday and Friday. I got mine; if there's more to be made, y'all can make it while I observe and ponder.

The best news is that twice in a row I have made the right move. This is a big change from previous happenings! From January to March, every time I put up they took some away. Then when I ducked to the sidelines, I missed the few opportunities there were.

This is a different style of investing for me, which by nature I am a buy-and-holder. In fact I can't really call this investing, can I? It is more like playing Wheel of Fortune -- take a spin! But the market has been so choppy lately, holding is certain curtains. Rather I have to discern when an up-day, a rare and precious up-day, is going to occur and then get on it 24 hours in advance. The real difficulty is that I cannot afford any more losses so I have to be cautious. Once I have recovered those remaining seven eighths, and have some actual profits, maybe I will get more aggressive. Good luck!

Dave

 

[Spaf]

imported post

Dave M wrote:

This is a different style of investing for me, which by nature I am a buy-and-holder. In fact I can't really call this investing, can I? It is more like playing Wheel of Fortune -- take a spin! But the market has been so choppy lately, holding is certain curtains. Rather I have to discern when an up-day, a rare and precious up-day, is going to occur and then get on it 24 hours in advance.
Dave

Every once and a while I look in the Tech tools and ponder something. Try this one:

http://www.incrediblecharts.com/technical/slow_stochastic.htm

Let me know what U thunk!

Rgds! Spaf

PS: It's on the performance charts at Yahoo, so it's easy to activate!

 

[Dave M]

imported post

P, the stochastic oscillator on the linked page compares today's -- or recent -- close to a past close. Looking at the computation, you can see that if today's close is the highest high of the period, %K will go to 100. If today's close is the lowest low of the period, %K = 0.This quotient will vary daily but always describes the market's movement in a consistant manner, a number between 0 and 100. The consistency is the advantage, I think. By constructing a moving average the variation is smoothed. Since we have a 12-24 hour turnaround in TSP, an integral multiple of thatwould be a natural time frame for averaging, for us.

Iwas thinking of something else. If I have a collection of points on a graph, I can fit a curve to them. The curve may not go through any of the individual points, but instead will minimize the total distance of all the points from the curve. Call this the line of best fit. Using the spread of the points, I can construct a bell-shaped curve that gives the probability thata point will liewithin a certain distance from the line of best fit.The probilities define our risk tolerance.

The curve of best fit does not have to be a straight line -- a 1st-order polynomial. 2nd-order or 3rd-order would be better because that way you get inflection points. 3rd-order curve-fitting is called cubic splines. Here is a link.

http://mathworld.wolfram.com/CubicSpline.html

Dave

P.S. This is what happens to a man when he becomes a meteorologist. And see my post above Dec 30th where I fit a straight line to my balance.

 

[Dave M]

imported post

This week appears to have sustained a turn-around. We closed today at 4197, up 75 from last Friday's close of 4122. (C+S+I times 100)

It marks the best week since I began keeping track in early March. Inthe last three weeks, two have been up. Of the last five weeks, three have been up. Over the last five weeks starting April 1st, we are down just 22 points, mostly due to one terrible week ending April 15th where we lost 133 points.

We have a long way to go, a lot of ground to make up. However it does look like the market is making an effort to recoup past losses and re-enter positive territory for the year. (4272 is where we stood on Jan 3rd; take away the week of April 15th and we are at 4330. The high for the year was 4387 on Mar 7th.)

We have cause for optimism. Good luck everybody.

Dave

 

[Dave M]

imported post

I put in 10% today, in effect for tomorrow, with the intention of staying in for a while, a couple weeks maybe. I think our G fund will give us a penny today, woo-hoo!If the trade-balance figures to be released mid-week don't tank us, I may increase my percentage. Ilooked, but saw no 'bad news' out there today, so what the heck. I still have losses to recoup.

Dave

 

[The_Technician]

imported post

G'morn Dave....I see you do some curve fitting.....

Have you tried anything higher than 3rd degree polynomials.....do you know about coefficient of determinations???

interesting data from such animals....it could tell you alot about the future....

The Technician

 

[Dave M]

imported post

Curve fitting is a delight to the technically-minded such as I. You start with a collection of points and end with convenient mathematical description. Polynomials are especiallyconvenient because they are so easy to differentiate and integrate. Third order polynomials are especially useful because -- as you remember from 9th grade algebra -- they have three zeroes. That means they will have two inflection points and can approximate a sine curve.

In meteorology we are confronted with cyclical datasets all the time. What is sometimes useful is what is called Fourier Analysis, named after a Frenchman. Here you fit a curve to the data that is not a polynomial but a mix of sine and cosine functions, something like this expression: Y= (a)sin(b)X + (c)cos(d)X + higher orders, sin squared of X squared, cos squared of X squared, etc. You come out with a waveform. In this way periodicities are revealed that might not be apparent at the start.

For example, say you are looking at daily rainfall totals at a station in in the mid-latitudes. There will be a fundamental period of one year as the summer rains give way to winter snows and then back again. Then there might be a 3-5 day period superposed on this which represents the passage of fronts and changes of airmasses. On top of it all might be a longer-term secular variation representing droughts, or maybe your reporting station has gotten urbanized as a city grew up around it, that sort of thing.

The advantage is that it gives you clues as to what to look for, physically. In itself it explains nothing. In fact, for the statistical methods to be valid you must start with the idea that the data points are independant of one another and that there is no inter-correlation. Then you reject this hypothesis at various levels of confidence. From there you investigate plausible physical mechanisms.

I am no expert on this stuff, I just know it's out there. When I read a paper which uses the techniques, I nod my head going, "Yep, yep, I recognize that, uh-huh..." I did my thesis on cloud seeding and used multiple linear/log-linear regression techniques to fit curves to historical datasets enabling me to predict what the rainfall would be in the target area based on nearby control areas. Then I compared the totals achieved during the project to what was expected based on the upwind controls.The deviation gauged the project's effectiveness, canyadigit? Stratification of the variables revealed "windows of opportunity" that existed. (That was in the 70's, which seem like the dark ages now.)

So what we need are true experts with knowledge of more modern computer methods and access to great gobs of data. It's fun stuff if you're into it. Thanks for reading, hope I've stimulated your thinking.

Dave

 

[The_Technician]

imported post

Well Dave, you seem to have some background in math....

Actually, if you take differential equations, and control system theory you could learn how to apply all those poles and zeros to "read" systems such as weather.....economics (which Greenspud is using), social interaction....electronic circuits, auto suspension systems.....they all are described in mathmatical models.....each pole and zero factor is a component and how it affects the system output....plus some constant =0. They are used to control output out of the system from the inputs.....

Take this for instance...

y = 5E-17x6 - 1E-11x5 + 1E-06x4 - 0.061x3 + 1742.3x2 - 3E+07x + 2E+11 is a sixth degree polynomial...it currently describes the S&P for some years back.....with approx 90% accuracy.....its trending down in the future.....you see the various power factors plus the constant = some value in the y direction....3rd degree poly's are not the only animal you can use.....generally the higher degree, the more accurate the description of the item you are describing.

I'm not really sure about the degree polynomial Greenspud is using, but I imagine it's 6 or 7 degree or maybe higher....it's called active economics (to control an active system -control system theory).....remember Reagan Financial advisor....that got canned later for his voodoo economics......I think he is responsible for getting it started......Greenspud I'm sure is using GDP, inflation, deficit and other factors in the equation to control our economy.....beats the way Jimmy Carter and those before him did it!!!

Thought you would like to know....

The Technician:dude:

 

[Dave M]

imported post

That's it, Tech. See, I knew someone was out there doing it, it is such an obvious idea. To gain access to the results we would need to be familiar with the literature, conference papers, abstracts, etc. It's a lot of work.

The trouble with high-order poly's is that they are well-behaved only within the original domain. They come in from infinity, oscillate about their zeroes, then zoom off to inifinity again. Typically the domain of interest is where they are oscillating. Outside that, they bomb.This is avoided by splines, which are piecewise continuous and fit smoothly together. But again, they only describe what we know, the historical period, and must be used with caution outside the domain, i.e., as a forecast tool.

Yes, math is (was) sort of behind all my studies. In meteorology we try to solve simultaneously a system of non-linear 2nd-order partial-differential equations derived from Newton's laws, laws of thermodynamics, ideal gas laws, etc., in three dimensions all in a rotating reference frame. It's a mess. Only by making a wholeseries of simplifying assumptions can they be solved analytically. Otherwise you have to use numerical methods which can only be approximately correct. Inevitably, the answer is a wave of some sort. And there is the problem of the data -- garbage in, garbage out.

But in the natural sciences we have the advantage of dealing with a rational system, one subject to natural law and nothing else. No amount of wishing can change the weather. In the market that is not the case. If enough people want something to happen, it will happen, in the market. Like say today every seller became a buyer; could happen. Or if some piece of bogus news causes every buyer to become a seller, suddenly. How do you model that?

Now we're getting into the realm of psychohistory, invented by Isaac Asimov in "Foundation". What we lack is a Hari Seldon! But...maybe I am talking to him right now? What are you working on, Tech?

Dave