I like creating daytrading systems
 

Hi,

I like creating daytrading systems, and the temptation to "optimize" them is incredible. I mean I create a system and then apply it to some historical data, and it occurs to me that if I changed the rules in a certain way, the system would work even better, like gangbusters in fact! (The problem, of course, is that systems optimized for a particular dataset don't work very well on any other dataset.)

Personally I think the stock market is getting ready to take a big dump. In fact, earlier today I was saying "Later they're gonna call today Bloody Friday!" (My prediction didnt' turn out correct.)

Best,

- Boyd

Harry Dent's model....
 

Hi Michael,

What you say is right about various businesses peaking at different times. Harry Dent already covers that in his books, and has different peaks for different businesses. (Of course, I can't summarize his whole book in a short post.)

Does that mean that those investments could represent good possibilities? Maybe. But keep in mind there are large lengths of time when Buffett has stopped buying stocks completey. If you read the recent articles, you'll see Buffett is at that stage again - he says he's not selling his stocks, but he's not buying any new stocks either.

As for the 46.5 years vs. 49 years, yes, he has changed it - but I don't see it as a problem. This is partly due to culture, since if people start having children later, then the point of "peak" spending will change. There's nothing wrong with making a small adjustment like this.

If you say that you can't adjust for this, then you're saying that the age of peak spending should never change. That's unrealistic. The POINT is that the age of peak spending roughly correlates with the stock market, which seems to be true most of the time.

Does the model hold recently? The model is not an exact fit, as you can see for yourself, and may be even less of one in recent times. Does that mean the model is no longer valid? I don't think so - but it is a theoretical model.

This is different from day-trading - this model is ridiculously simple. Take the birth rates - shift them by 46-49 years. None of the day trading models I'm aware of are this simple. Plus, there is rationality behind this model - whereas day-trading models often don't seem to have a "reason" as to why things should behave the way they say.

However, there's an inaccuracy in the model....

He looks at BIRTH RATES. What he really wants to look at is the POPULATION SIZE at the age of peak spending (around 46-49 years). While there will be a large correlation of the population size at those ages with the shifted birth rates, it won't be exact - because of immigration.

If you adjusted for this, I wonder if it would make a difference.

Anyhow Michael, like I said - I can't summarize everything in his books here. He's taken into account many of your points already. I've only talked about a very small part of what he's written about.

- Dien

P.S. There is another effect - the more people know about this model, the more it will affect their behavior. So, the more this model is well-known, the less accurate it will be - because the knowledge of the model itself will start to affect people's stock-buying and selling behavior. The effect should be that everything the model predicts should start happening sooner. So, for example, if everyone expects a depression starting around 2008 - they might start selling their stocks earlier than that to be safe. This could mean that the depression could occur even earlier than the model predicts.

Re: Harry Dent's model....
 

> As for the 46.5 years vs. 49 years, yes, he
> has changed it - but I don't see it as a
> problem. This is partly due to culture,
> since if people start having children later,
> then the point of "peak" spending
> will change. There's nothing wrong with
> making a small adjustment like this.

> If you say that you can't adjust for this,
> then you're saying that the age of peak
> spending should never change. That's
> unrealistic. The POINT is that the age of
> peak spending roughly correlates with the
> stock market, which seems to be true most of
> the time.

I am not saying the offset cannot change. But, from my understanding, he has changed the offset for the entire time period. He has not adjusted the offset as time has gone along.

So the first chart was offset by 49 years. Now his new chart didn't fit he had to change the entire chart's offset by 46.5 years. THAT is changing his chart to fit what he want to prove.

Doing that, I bet I could make a radom ECG stip match the stocks.

> However, there's an inaccuracy in the
> model....

> He looks at BIRTH RATES. What he really
> wants to look at is the POPULATION SIZE at
> the age of peak spending (around 46-49
> years). While there will be a large
> correlation of the population size at those
> ages with the shifted birth rates, it won't
> be exact - because of immigration.

I also question his peak spending idea...

At 49 (or 46.5) most parents don't have kids living at home. Certainly not the Boomers. Tha would mean, if they had a child at 25, the child is 21-24 and at home.

That's a little unrealistic in my opinion.

I think he is finding arguements to fit his model. And I am sure we could find equally strong prof that peak spending happens at other times instead.

> P.S. There is another effect - the more
> people know about this model, the more it
> will affect their behavior. So, the more
> this model is well-known, the less accurate
> it will be - because the knowledge of the
> model itself will start to affect people's
> stock-buying and selling behavior.

Yes. That's what I said... just knowing about it changes it.

The
> effect should be that everything the model
> predicts should start happening sooner. So,
> for example, if everyone expects a
> depression starting around 2008 - they might
> start selling their stocks earlier than that
> to be safe. This could mean that the
> depression could occur even earlier than the
> model predicts.

If that is the case, then the peak spending reason why goes out the door. How would he adjust for that? Change the offset again, perhaps :o))

I actually find it interesting. A piece of info to keep up my sleeve and bring out and consider.

One thing I wonder... seeing as the stockmarket didn't go as high as he said... and it goes as low... that wold keep it about it's current level for the next whole bunch of years, wouldn't it?

I wonder why the mainstream hasn't picked up on his theory?

On another note... Beating The Streat (Peter Lynch). Have you read it? Would you recommend it as a good book for a novice know-nothing-about-the-stockmarket guy like me?

Michael Ross

Re: Harry Dent's model....
 

Hi Michael,

> I am not saying the offset cannot change.
> But, from my understanding, he has changed
> the offset for the entire time period. He
> has not adjusted the offset as time has gone
> along.

> So the first chart was offset by 49 years.
> Now his new chart didn't fit he had to
> change the entire chart's offset by 46.5
> years. THAT is changing his chart to fit
> what he want to prove.

> Doing that, I bet I could make a radom ECG
> stip match the stocks.

Doesn't changing the age from 46.5 to 49 only shift the graph left or right?

Here's a random ECG strip....



How will you make this fit the Dow Jones Index?

> I also question his peak spending idea...

> At 49 (or 46.5) most parents don't have kids
> living at home. Certainly not the Boomers.
> Tha would mean, if they had a child at 25,
> the child is 21-24 and at home.

> That's a little unrealistic in my opinion.

Should it be the average age of having the FIRST child, or the average age of having the "average" child? (Eg. if you had 3 kids, you would have to take the age at having each of the 3 kids, then taking the average.)

For example, when he had children, my Dad was aged 29, 31, and 40. If we were using him as an example, should we take age 29 as the key figure (the age he had his first child), or the average age of these three ages - that is, (29+31+40)/3 = 33.3 ?

Further questions.... Does a child have to live at home to be supported by their parents? What about children who are university students?

In the USA, the culture is such that usually children leave home to go to university (often in a different city or state). However, these kids are still often supported by their parents. Could that affect their parents' spending, even if the kids are not living at home?

> I think he is finding arguements to fit his
> model. And I am sure we could find equally
> strong prof that peak spending happens at
> other times instead.

> Yes. That's what I said... just knowing
> about it changes it.

> The

> If that is the case, then the peak spending
> reason why goes out the door. How would he
> adjust for that? Change the offset again,
> perhaps :o))

> I actually find it interesting. A piece of
> info to keep up my sleeve and bring out and
> consider.

> One thing I wonder... seeing as the
> stockmarket didn't go as high as he said...
> and it goes as low... that wold keep it
> about it's current level for the next whole
> bunch of years, wouldn't it?

> I wonder why the mainstream hasn't picked up
> on his theory?

I was told (by a friend who's read it) that Kiyosaki's latest book seems to draw on Dent's theories. (I haven't read it myself though.)

> On another note... Beating The Streat (Peter
> Lynch). Have you read it? Would you
> recommend it as a good book for a novice
> know-nothing-about-the-stockmarket guy like
> me?

Yeah, it's a good book. :)

- Dien

Re: Harry Dent's model....
 

> Hi Michael,

> Doesn't changing the age from 46.5 to 49
> only shift the graph left or right?

Yes. However, the peak spending is either 49 or it's 46.5. If it changes over time due to lifestyle then that means a constant adjustment.

Just up and changing the peak spending claim and applying it to the whole graph - as he appears to have done - is something you yourself have said is wrong due to lifestyle changes.

Now you're arguing that it only makes a slight difference.

Come on Dien. Stick to your arguements, eh? :o)

> Here's a random ECG strip....

> How will you make this fit the Dow Jones
> Index?

Well... actually it's quite easy...

See where the graph is labeled (1). What follows that is a drop, then a slight rise/level off, followed by another drop and then a rise.

This fits Dent's graph from now and further into the future... a drop, level off slight rise, drop, large rise.

See it?

To make it fit better just "adjust it" for graph recording speed or some other reason and voila!

> Should it be the average age of having the
> FIRST child, or the average age of having
> the "average" child? (Eg. if you
> had 3 kids, you would have to take the age
> at having each of the 3 kids, then taking
> the average.)

> For example, when he had children, my Dad
> was aged 29, 31, and 40. If we were using
> him as an example, should we take age 29 as
> the key figure (the age he had his first
> child), or the average age of these three
> ages - that is, (29+31+40)/3 = 33.3 ?

> Further questions.... Does a child have to
> live at home to be supported by their
> parents? What about children who are
> university students?

> In the USA, the culture is such that usually
> children leave home to go to university
> (often in a different city or state).
> However, these kids are still often
> supported by their parents. Could that
> affect their parents' spending, even if the
> kids are not living at home?

Now you're just being ridiculous. Use your same arguements for the entire graph. From the beginning of the century.

Either peak spending occurs at 49 or 46.5 (why not 46 or 47?). Either people have kids late in their 20s and into their 30s or they have them earlier. And earlier this century, they certainly had them earlier.

So a 22 year old has a child. That means peak spening occurs when that child is 25 to 27. Come on... they are having their own kids at that age.

Dent's theory is interesting. But I am not justing going to accept it without question. Especially when he changes his graphing formula to make it fit. Small technicalities? Maybe. Maybe not. But either it works or it doesn't.

As I said... it's something to keep up my sleeve.

Michael Ross