Dear Dr. Vemic

"yes sometimes but not always"

Does it means that we know when to use AI for stock price prediction?

in fact, by starting this discussion, I am really interested in hearing the scientist experiences on this field.

have you ever had a successful experience on predicting the stock price.

what is your idea about dedicating our time and science on this subject in order to earn money?

do you see it as a way of making money for an AI expert?

let me ask you this question differently,

personally, do you risk your money and let the AI to sell and buy for you ?

Dear Dr. Vemic

of course, it is my pleasure to work with you on the subject, right now,

I am trying to do some field researches in order to further clarify the financial potential of the subject for myself.

Dear Dr. Namadchian,

I believe it is possible to make a profit on stocks based on Artificial Intelligence. In my conception and experience as an amateur investor, programmer, MSc and author of an article on Artificial Intelligence, I can make a few brief comments on the subject.

Of course, mathematicians believe in technical analysis, which in itself would give an affirmative answer to your question. On the other hand, on fundamentalist analyzes, we can create basic models composed of "perceptors or sensors" and "actuators" that can result in positive results based on modeling from easily identified objectives.

Finally, the most complicated and influential factor in investments, I believe must be human impulsivity. This would be reduced to zero with models that establish safe margins of daily profit, considering the daily average change of an asset, with minimally safe percentages of profit but with an efficiency in the performance of the artificial intelligence model.

If you are interested a little "perceptors, sensors and actuators," follows the link of my article published here.

Best regards,

Alexandre Villarmosa

It is only possible to make money if there is an inefficency in the market of interest, ie, an opportunity that is not generally know to the public. Every prediction in the market that is acted on becomes a "self-cancelling" prophacy. That means if there is money to be made by using any type of algorithm, including AI or machine learning in general, the very act of taking the "excess" profit out of the market restores the efficiency of that market segment and makes it no longer available. Note, there are often small predictable variations in most markets (stocks, options, futures, etc.), but after discounting for transaction costs there is no actual profit opportunity.

Thousands of people have developed their pet algorithms for making money in the market and most of them can make some "profit" when the market is generally trending upward - no algorithm is required to do that. However, only a handfull of individuals make money under all market conditions, and their skills would be challenging to model using AI. This would require full access to a number of these experts' practices and though processes and considerable expertise in machine learning methods.

Dear Dr. Wilber

it is a really an interesting fact that I have was not aware of.

do you mean that there is some kind of balance in the market and any prediction strategy will be successful for a while and it will be temporary?

Exactly. If you alone have discovered an inefficiency (ie, opportunity), you may take advantage of it as long as you want, provided 1) you do not get greedy and start to tip the balance yourself, or 2) others don't catch on and incorporate this previously unseen opportunity into their trading strategy until it is no more, or 3) underlying forces have not caused a temporary loophole and measures are taken to remove it - usually when someone notices money is being lost and they "fix" it.

The efficiency of the market in general is caused by a large number of investors trying to gain an advantage over everyone else by making a profit. Since the market is for all intents a zero-sum game, everyone who makes a profit gets it by taking from a larger collective of other investors. It is the effect of the large number of investors trying to get ahead that makes this give and take into a balance with small fluctuations that appear nearly random.

Of course, it is not quite a zero-sum game because of transaction costs (makes everyone pay a little extra premium) and other more slowly moving changes that typically (usually, but not always) increase the value of the underlying assets over time. This latter effect, in terms of real value, is negated by inflation, which just tends to keep pace in a more or less stable -- that is, exponentially growing -- population.

As a side, but related note, these primal balances tend to get out of wack when fundamental variables change, such as population growth diverging from the historical exponential rate that had existed for a very, very long time. The growth rate is now dropping much below the historical levels, especially in first world countries. But, the consequences of this are another discussion.