Predicting the stock market

[skuzie]

There are there many products out there like the following that use statistics, neural networks, and genetic algorithms to increase your chances in the stock market by being able to predict outcomes by looking at trends.

Neural Networks - NeuroSolutions, Neural Network Software

Imagine a set of virtual traders, they have no idea what the stock market is but their task is to make the most money out of it. They try several techniques using mathematics and statistics and observing trends to try to make best estimates, they build a neural network that helps them decide what worked more in the past and what worked less as they make sense of all the indicators that are being feed to them. Now throw in the most powerful concept, evolution, have the virtual traders compete against each other, the driving force of 'genetic prolification' is the maximum profit that virtual traders can make, if a certain treader can make more then use genetic algorithms for prolification to make their neuralnet traits more frequent in the virtual trader world. After a while you can imagine that these traders will become very skilled at what they do, which is stock market prediction.

Do you think this is possible? are there any better than average products out there that promise this?

Also, was wondering whats the general view out there on stock market these days, are there many traders out there and is it worth it? and is there a way to apply the sciences to increase your stats, or your earnings in trading.

[Turtle]

Originally Posted by skuzie

After a while you can imagine that [these] traders will become very skilled at what they do, which is stock market prediction.

Do you think this is possible? are there any better than average products out there that promise this?

Also, was wondering whats the general view out there on stock market these days, are there many traders out there and is it worth it? and is there a way to apply the sciences to increase your stats, or your earnings in trading.

I have little knowledge of all this in regards to how many traders etcetera, save for occasionaly watching the PBS Business Report with Paul Kangas(sp).

As to making 'predictions', I have a tasty tidbit for you other than neural nets specifically, and that is chaos theory. No doubt to me at least, that neural nets in nature do employ chaos theory (now popularly called 'complex systems' theory) if only in some sort of fractal wiring.

Here's a link I think will interest you: (pun intended )

Originally Posted by csbrown.edu

Suppose you could discern market trends, speed up time to see where those trends were going, then make a bet on what you discovered. That's how geeks in suits -- known as rocket scientists -- are raking it in now.

Cracking Wall Street

[OzAnt]

Hi,

This could be considered off-topic (just ignore me if it is), but it reminded me of a cool movie I saw a couple of years ago on banks, stock markets and chaos theory in predicting stock market trends.

It's called "The Bank". Here's a link: The Bank (2001)

Cheers,
Ant

----
Mouse movement detected. Restarting windows for changes to take effect.

[Fatstep]

I find the way the stock market works intricate and confusing, but predicting certain companies' stocks quite easy.

I voted Active trader because of my father, but I am an active part in it.

[minna]

Originally Posted by skuzie

Do you think this is possible? are there any better than average products out there that promise this?

Also, was wondering whats the general view out there on stock market these days, are there many traders out there and is it worth it? and is there a way to apply the sciences to increase your stats, or your earnings in trading.

Predicting the stock market? Well many have tried in the past and many are trying now. There are people who make a living out of trying to predict the market, there are also people who loose everything due to wrong predcitions/decisions.
Keeping in mind that most statistical methods use PAST prices to predict FUTURE values. (See the logic gap? ) I generally don't trust any statistical method/software that claim to be able to predict anything.
Stats should only be used in conjunction with a good understanding of the market, the companies, and the current economic scene in general. Stats is not going to be able to predict a take over bid. It will only reflect the price change after it has happened.
Of course you can rely solely on stats if you are just looking for a place to put your money and want a better return than the banks. So all you want to know is if the money you invested will grow. But you can see those stats for free online (ie. ASX) There is really no need to buy expensive software.
If you want to do more than that then software will not get you far. They use historical prices and hence you can probably get a good general trend of the stock. But to make substantial money on the stock market you need to know information such as when the next take-over bids going to be, what big decisions the company will make next etc. No software is going to tell you that. In fact its impossible for anyone to know those information beforehand, unless there is inside-trading involved.
What you need is good intuition, lots of research, good understanding, lots of time dedicated to monitoring the market and companies and of course luck.

PS. What I have said are my own personal views.

[Zythryn]

Predicting the overall activity of the market over a long period of time is not too dificult. Predicting day to day activity is not easy at all. I chose 'active trader' although I mainly buy and hold. So it really isn't that active, but I am heavily invested.

[TheBigDog]

A few years ago I went on a mental bender and wrote a huge program for trading stocks. It did fantastically well, but I just don't have the capital to put it into practice. What I did was identify motivating factors in an attempt to isolate "good buys". To that end I came up with a formula to look for stocks that were in trouble, almost to the point of being delisted (removed from the exchange) which is something that companies will move heaven and earth to prevent. The idea was to cash in on that heaven and earth moving working at least one time. That most companies when the plummet to the bottom make one recovery through some sort of drastic action that builds hope in the investors. Then they either live or dies. Either way you have taken your profit and moved on. The strategy required persistent purchasing of stocks in a certain price situation, buying a fixed amount each day, then holding the stock until it had recovered to a certain profit marker and selling. In its worst simulated year it made 26%, and as in the 80% range most other years (I did 6 in all). And the stocks purchased in that 26% years may have recovered for a profit at a future date, but I did not play long enough to find out. I calculated that I needed 4.5 million dollars to fund the system. It required a bunch of purchasing and was WAY out of my ballpark.

Someday...

Bill

[skuzie]

When i said predicting the stock market in a literal sense i didn't mean exactly predict it, but learn to be able to spot the patterns when is a good time to buy and sell to maximize your chances and your growth. The software i mentioned above claims to have average annual returns of 29.68% compared to other products having 15.65%:

Neural Network Investment and Trading Software, stock prices, asset allocation, portfolio management

There is even a testimonial of a treader having 80% to 120% per month returns on futures contracts which is pretty damn impressive:

TradingSolutions - Customer Interviews, Yuri Shramenko

I like the way he created a strategy with the use of intelligent systems, i would like to create something similar myself with regard to options trading. There are lots of trends out there to soak in, and the general rule of thumb is that 50% of stock movement is the market itself and not the individual stock so generalizing isn't such a bad idea.

'Buying and Holding' is a very weak strategy as i believe, if you have indications that the price will fall and your 90% sure it will because it did in the past with the same indications then why take your chances, its all about maximizing your returns and minimizing loss.

I like how you formulated your strategy BigDog, its exactly something like that i want to create, but like i said before with options, since the return can be much much greater and you have more control over your risk and you can gain money if the trend is up or down depending what your actions are.

Heres another movie to add to the list OzAnt its called Pi:

Pi (1998)

ps. Thanks for the link Turtle

.. and remember 10K can turn to over 1 million in under 15 years if you only manage to have a constant 3% increase per month.

[Zythryn]

Hmm, average returns of 30%?
They are feeding off greed.
Ask them if they will give you your money back if you loose money
Buy and hold is FAR superior to market timing which you are proposing. Someone last year on another board I visit did exactly that. They figured indicators were that the market was due for a downturn. They missed. Lost out on a 10% gain because they kept telling themselves that the market was going down.
If I can gain an average of 8% a year, doing absolutely nothing (beyond the initial research), I'll take it. I don't to take the risk in shooting for 30%.

How long have they had this system? Ask yourself, if it really works, why aren't more people using it?

Have the back tested it? I would be curious about how that works out. Did it predict the market crash after 9/11 or any of the other market crashes in the last 50 years?

Walk very carefully before you give these people any money.

[Qfwfq]

Originally Posted by Zythryn

Ask yourself, if it really works, why aren't more people using it?

I would ask Skuzie a quite different question: If it really works, why are they selling it, instead of closely guarding their valuable secrets?

Remember that the profitability from actual productivity is dividends (and capital gain due to anticipation thereof). Larger gains are reaped from speculation on fluctuations; this component of the profitability is a zero-sums game and every extra cent gained by someone is a cent less gained by others. If you find you are adept at it, it means you are beating others to it, just like winning at a poker table. Would you imagine a poker sharp ever giving his best of advice to his opponents?

P. S. I just remembered an old adage of stock traders I once read:

Bulls make money, bears make money, pigs get slaughtered.