Do stock prices really trend?

Chart Forum » Does Technical Analysis Really Work? » Do stock prices really trend?

«« Previous Next »»

Thread Start New Thread
Last Poster Posts Pages Last Post

Author

Message

david_louisson

Member

Post Number: 221
Registered: 02-2004

Rating: N/A
Votes: 0


Tuesday, March 07, 2006 - 05:47 am:

Do stock prices really trend?

Yes, I believe they do. If price movements were completely random, it would be impossible to succeed at trading, in the long run, over any time frame.

However, try the following test.

Below are six charts.

Three represent 1000 consecutive EOD closing prices for three different UK100 stocks (chosen at random).

The other three were created by generating 1000 points using Microsoft Excel's random number generator.

Of course, the question is: Which of the three are genuine prices, and which are the random sequences?

OK, punters - let's have your expert opinions!

Correct answers will be published in due course.

Good luck!
David

Let me repeat: If price movements were completely random, it would be impossible to succeed at trading, in the long run, over any time frame.

david_louisson

Member

Post Number: 222
Registered: 02-2004

Rating: N/A
Votes: 0


Tuesday, March 07, 2006 - 05:54 am:

Footnote: the vertical scale in the charts is linear, as opposed to logarithmic.

dutchdanish

Member

Post Number: 12
Registered: 10-2005

Rating: N/A
Votes: 0


Tuesday, March 07, 2006 - 04:24 pm:

Real charts: F, C, B?

maybe A?

(Message edited by dutchdanish on March 07, 2006)

"...yet with how many things are we upon the brink of becoming acquainted, if cowardice or carelessness did not restrain our inquiries."
--M. Shelley

snorter

Member

Post Number: 87
Registered: 01-2004

Rating: N/A
Votes: 0


Tuesday, March 07, 2006 - 04:40 pm:

David

Good post!
Excel charts -
Each point is a change of one point from the previous point, the direction being random? (+1, -1, 0)
Were these random charts, randomly chosen?

"Failure is not an option"

stevo

Member

Post Number: 339
Registered: 01-2003

Rating: N/A
Votes: 0


Tuesday, March 07, 2006 - 05:55 pm:

David
They are all probably random sequences!

Any of them could have generated using a random heads and tails spreadsheet - you will get 8 plus heads or tails in a row if you throw the coin 1000 times. Also they all look to have been generated using Excel - just keep hitting the F9 button.

It doesn't matter if stock prices are random or genuinely trending for some reason as long as we make substantially more than we lose.

stevo

david_louisson

Member

Post Number: 224
Registered: 02-2004

Rating: N/A
Votes: 0


Tuesday, March 07, 2006 - 07:41 pm:

Thanks, guys, for your replies.

Dutch: you're the only person bold enough (thus far) to attempt an answer. But if I tell you how right or wrong you are now, it might spoil it for others who also want to try. So I will postpone posting the answer for now.

Snorter: the three Excel graphs were plotted by generating a random number in the interval (-5,+5) for each "day", simulating a possible rise or fall of up to 5 points. This value is then added to the previous day. Re selecting the three stocks, I simply chose three popularly traded UK100 companies that didn't have "giveaway" characteristics like splits or "earthquakes". I did NOT choose ones that I thought exhibited the most "random-like" behavior. That would defeat my purpose, because there is a genuine point that I'm trying to make here. These are actual EOD closing prices being plotted.

Stevo: only three of the charts are random. You are correct when you say that numbers randomly generated will create the APPEARANCE of "trends" (e.g. likelihood of 8 successive heads in 1000 coin tosses). But it would be impossible to trade these profitably, because no matter where one "enters" or "exits" there is always EXACTLY a 50% chance of the "trend" reversing at that point (which means that if there were transaction costs, one must eventually lose, as it would be a negative sum game). That applies irrespective of whether one is trading long term or short term. The assumption TA makes is that prices (under suitable market conditions) actually create a GREATER THAN 50% chance that the next bar(s) will continue in the direction of the trend; that they also occasionally break out with increased volume, ping back from extremes, respect trendlines, support and resistance, reverse at Fibonacci retracement areas, etc. It is these types of behavior that defy randomness, and hence potentially allow a trader to profit.

The underlying point that this post attempts to make is that there is little immediate visual difference between randomly generated numbers and "trending" prices, but somehow it is apparently significant enough to allow expert traders to profit across the course of a lifetime.

I was wondering whether experienced chartists might be able to discern patterns (e.g. levels of support and resistance) in the prices which would enable them to make a correct "guess". My less educated eyes struggle to see them.

David

(Message edited by david_louisson on March 07, 2006)

ingot54

Member

Post Number: 1178
Registered: 05-2004

Rating: N/A
Votes: 0


Tuesday, March 07, 2006 - 08:41 pm:

If price movements were completely random, it would be impossible to succeed at trading, in the long run, over any time frame.

But they ARE random, David, imho.

All trading should be done from the hard right edge - the next day's price should always be seen as a lottery.

You only have your edge to help you, and even your edge will fail you on many of your trades .

I am missing the point here, I am sure.

Even if all 6 charts are randomly generated, or all 6 charts actual sequences of historic activity, does it make a difference to how the trade is managed?

I'm not sure I can agree with the contention. It's really splitting hairs, isn't it?

Nothing trends forever! And time is the essence of any trade, be it intraday, or 10 years.

Keep Smiling

Trading style :CFD's predominantly. Looking for ways to enter CFD trading over long term.

ingot54

Member

Post Number: 1179
Registered: 05-2004

Rating: N/A
Votes: 0


Tuesday, March 07, 2006 - 09:09 pm:

Travis Morien has an excellent essay on trend and randomness here

Morien seems to support your contention, and the rest of his essays are worth the time on a rainy day.

Keep Smiling

Trading style :CFD's predominantly. Looking for ways to enter CFD trading over long term.

azure

Member

Post Number: 11
Registered: 02-2005

Rating: N/A
Votes: 0


Tuesday, March 07, 2006 - 10:00 pm:

I don't believe prices can be completely random, otherwise over the long term the market would not move. Share prices must in some way reflect the value of the companies, even if adjusted by some market sentiment, and the real value of companies do change due to numerous factors such as good and bad management, commodity prices, etc.

Now, I have to admit that even if pure randomness can be disproved, it doesn't imply that trending does occur. Perhaps there is some other behaviour that still leads to a price increase in the long run. However I have seen enough charts to say that I do believe that trends occur.

Perhaps some level of proof would be to calculate the probability of a share chart such as, say RIN occurring by chance. If the odds are small enough perhaps we can feel there is some statistical significance to the behaviour we see.

By the way I guess B, C and D as the random charts as they show the smallest deviation from the mean.

justice

Member

Post Number: 489
Registered: 01-2003

Rating:
Votes: 1


Tuesday, March 07, 2006 - 10:35 pm:

From documentaries on the subject I know that during the last tech bubble bull run, there was at least one NASDAQ listing that went up consistently for months despite the only required regular report produced had but one line of text which stated quite categorically that the company had precisely NO business activity during that (or any other) reporting period. The company did nothing and owned nothing. Lots of "smart" traders made good money on this 'value' company before the $hit hit the fan after the bubble burst.
I don't know if this contributes interest to this discussion, but it sure does wonders for my sense of perspective.

(Message edited by justice on March 07, 2006)

... XXXX2 ... :-)

msparks

Member

Post Number: 352
Registered: 10-2004

Rating: N/A
Votes: 0


Tuesday, March 07, 2006 - 11:20 pm:

David
The time frame of 1000 days is roughly 3 years so i could look at the index and compare to see if there were any common events, but i didn't.

Also if the time frame was marked on the charts and the reporting dates were roughly known, maybe they are the same as asx stocks ,don't know, but then that wouldn't be only TA would it ?

I would say A,C and E are the stock price charts.

Trend lines, breakouts ,support /resistance can be drawn on all the charts.

BDF are random generated i reckon and i will tell you why when you give the answer out OK.

By the way, you can still make money with 20 % winners and 80% losers when trading IF the 2 winners make more than the 8 loss's so what is the point you are trying to make.

Perhaps just put some money down and just do it, you will learn more in a month than all the years of research and procrastination.

PS If i am wrong, just ignore everything i said, ok.

EDIT - How many random charts did you generate to pick the one's you posted ?

(Message edited by msparks on March 08, 2006)

msparks

Member

Post Number: 353
Registered: 10-2004

Rating: N/A
Votes: 0


Wednesday, March 08, 2006 - 12:34 am:

Hi David
Have to change after a second look at the little charts, from E as a stock chart to being random generated and F to being a stock chart.

So A,C anf F are stock charts and BDE are random.
Reasons below.

tony_m

Member

Post Number: 547
Registered: 01-2003

Rating: N/A
Votes: 0


Wednesday, March 08, 2006 - 06:47 am:

Do stocks really trend? I hope so, otherwise I must have made a lot of money over the last few years by accident. This one is trending and filling the coffers and I have 9 others just like it. I think the answer is yes. Tony_M

BOL

chart_rider

Member