Search
× Search
Thursday, March 28, 2024

Archived Discussions

Recent member discussions

The Algorithmic Traders' Association prides itself on providing a forum for the publication and dissemination of its members' white papers, research, reflections, works in progress, and other contributions. Please Note that archive searches and some of our members' publications are reserved for members only, so please log in or sign up to gain the most from our members' contributions.

Securities Analysts and Economists.Will Big Data make those jobs obsolete?

photo

 Tim Heaton, Trading Technologist | Several US Patents | In the market for opportunity

 Monday, August 11, 2014

I think "yes", here's a counterpoint http://www.greenbookblog.org/2014/05/19/why-big-data-will-never-replace-market-research/


Print

5 comments on article "Securities Analysts and Economists.Will Big Data make those jobs obsolete?"

photo

 Muhammad A., Independent Day Trader at Equity Day-Trader

 Thursday, September 11, 2014



The machines can't think for themselves. There are two very important things this discussion have not considered so far. The markets are adaptive meaning once everyone's action is aligned with the predicted outcome the market's "big" players will flip the outcome. For example for simplicity lets assume that a large bar on heavy volume after 10:00 am results in a move in the same direction of the bar 70% of the time. Once market participants realize that and start trading it heavily, the big guys will flip the odds and take the other side of the trade. At that point this trade will no longer work. Finding the next trade that work needs humans intellect to isolate it and program it and the cycle repeats itself. Economics and economists have a similar thing going. Economics is based on humans beliefs, if we believe we are in a recession , we will be more careful hiring and spending. If we believe a bank is going out of business, a run on the bank will take place. Except that for economics its the other way around. Governments try to influence our beliefs about the market and the economy to extract a desired behavior from the masses. At some point in time enough people will start doubting the story and the behavior start changing which then needs to be isolated and reprogrammed.


photo

 WH Chan, Information Technology and Services Professional

 Thursday, September 11, 2014



Sometimes,the machine can 'think'. For example,the Hanoi Tower juggling of circular discs. However, if it gets complicate,we may need AI. The difficulty seems to be the irrational human nature such as greed and fear. They are also changing from time to time. Even humans, ourselves, cannot understand one another. What is the hope? Even Buffet fails in some of his investments..


photo

 Muhammad A., Independent Day Trader at Equity Day-Trader

 Thursday, September 11, 2014



The machine thinking is limited. You have to feed it variable of certain types to guide it to "think" in a certain way. Your ability to do what is called data munging precedes the machine's "thinking".


photo

 Matthias W., Software Designer at Carmeq GmbH

 Thursday, September 11, 2014



Muhammad,

you're right, machines can't think like humans. They execute basic instructions. But the complexity of software built using these instructions is huge. Eventually machines were able to beat most, later even all humans in chess, which requires thinking processes based on the outcome of applying a set of strong rules. Neuronal networks, Bayesian networks, etc. are in most cases not as good as humans because they usually have much less resources (vs. brain size and stored wisdom), time (days vs. tens of years) and training sets (selected, labelled data vs. vast experiences) than humans. Still they are already close if not sometimes better.

Rule based learning or evolutionary computing even can discover new knowledge or find new solutions simply by scaling the computational performance and data sets (big data).

As I wrote above, we humans have our own thinking problems due to our limited brain "interface", processing speed and accuracy, and lots of different biases and dangerously incomplete wisdom, which is to be extrapolated and restored in scary ways. The worst of it: we often aren't even able to recognize that happening.

Adaptivity with low latencies is an important aspect, which could be mitigated by (automatically) identifying any relevant conditions or using more generalized rules, while reducing the exposure to events.


photo

 Asra I., SAP IT Consultant at Oil & Gas

 Thursday, September 11, 2014



Matthais, I must admit that I have not yet had the pleasure of reading all the titles listed end to end. I agree with most of the comments in here; in my humble view, the human specie still reacts with perceptions which entails feelings and even though at given variables and time some predictability is possible, its accuracy is limited to the last assumption until changed. In highly dynamic markets it is given that the last assumption will change before the next analysis can be done and hence my view.

Secondly, the infrastructure and the memory limitations for even the in-memory data call is still limited and improving.

Please login or register to post comments.

TRADING FUTURES AND OPTIONS INVOLVES SUBSTANTIAL RISK OF LOSS AND IS NOT SUITABLE FOR ALL INVESTORS
Terms Of UsePrivacy StatementCopyright 2018 Algorithmic Traders Association