SERVING THE QUANTITATIVE FINANCE COMMUNITY

 
User avatar
samudra
Topic Author
Posts: 730
Joined: December 23rd, 2002, 2:48 pm

How to beat machine learning/automation/big data to make money?

August 28th, 2015, 12:50 am

Here in India I am trying to use fundamental analysis and basic common sense to make money for last 7 months or so. Things has worked to certainextent. I am ahead of the index and returns of most good MF. My major background has been technology. Made some money in the goldeninternet era. Left USA in 2007 came back to India. Left work enjoyed life as much as I could, worked hard on gym, travelled length and breadth of India, really came to know India(greatest knowledge which I hope will help me make my money)and now after 8 yr break trying to make money as an investor.But my God what exponential growth in technology I see. I simply dontbelieve to what extent machine learning/analytics and big datahas developed and what lies ahead.If this continues I feel atleast 60% technology jobs will go.I thank God for quiting job in 2007. I was working around 80 hr week then. IfI would have worked last 8 yrs with this technology pace I may haveruined my health big time. I can imagine the stress level in high tech/quant jobs in NYC and silicon valley now a days. Thank God I left USA and came backto India.But now this technology beast is scaring the hell out of me.I feel 80% trading will be completely automated in next 20 years.The human quant as we know today will be very few largely replacedby AI.How at all I can beat these AI beasts and automated tradingsoftwares and make reasonable money in my investments? Will I be able to use common sense, fundamental knowledge and knowledge of India to stay ahead of AI/machine learning?Times are indeed going to be very tough going forward.Welcome to the world of automatons.With so few jobs going forward what will human beings do? PhD, post Doc then what .... I mean how many more tags will you need to get a quant job 10 yrs from now? Technology was supposed to free us from work but it seemswe may end up being slaves. God bless human beings!
 
User avatar
platinum
Posts: 158
Joined: May 6th, 2009, 11:08 am

How to beat machine learning/automation/big data to make money?

August 28th, 2015, 1:13 am

Having watched this scene for a number of years, I would say that there are still areas where qualitative judgment matters and a discretionary investment process can do well informed by quant methods and a systematic approach, so there is a rigorous formalism to it.Examples would include event-driven arb, emerging tech companies/IPO-post-IPO performance, and global macro.In the last area, your personal knowledge of India and more broadly, if you should choose to branch out, certain parts of Asia/Southeast Asia could be useful. If you were interested in the tech/IPO side, the Indus network in the US is quite strong, particularly in Silicon Valley, so you could get connected there.Rather than investing your own money, would you be interested in and able to form a team of similar industry veterans and raise capital?
 
User avatar
Traden4Alpha
Posts: 23951
Joined: September 20th, 2002, 8:30 pm

How to beat machine learning/automation/big data to make money?

August 28th, 2015, 1:44 am

The notion that one must "beat" or "lose to" the machines is a false dichotomy. Although it may be true that the history of technology has been one of workers doing X getting replaced by machines doing X, the real lesson is one of the rise of people doing Y now that they have machines that do X. If one makes the doing of X less expensive, then suddenly people find many new applications for X that weren't possible when it depended on hiring expensive X workers. The challenge is to be the kind of person who can switch from doing X to doing Y. Think of all the tough problems or unmet needs in the world that aren't getting handled because they don't justify hiring an expensive PhD/Post Doc. That's the set of Y! At the very least, the kind of person who could do X may be in a good position to become an X machine maker, manager, or consultant.
 
User avatar
samudra
Topic Author
Posts: 730
Joined: December 23rd, 2002, 2:48 pm

How to beat machine learning/automation/big data to make money?

August 28th, 2015, 4:24 am

Platinum,Thanks a lot for spending your valuable time and providing valuable input. I greatly appreciate it.As you will realize for last 8 years mostly I have had fun and was reading about great investorsmore for curiousity about there psychological aspects and to glean about how there brain worked. One of my hobbies is psychology. When I went back to India I have decidedto retire and spend my time enjoying my hobbies.But off late a guilty feeling is stirring my conscience. Here I have seen good amount of poverty.So it is paramount to do something to help the needy.So I decided to become and investor multiply my wealth and use it for a good cause.Honestly I have no prior experience as an investor. So really don't know whether I willsucceed. Given this backdrop I am not sure if others industry veterans/professionals will like a newbie like me tobe part of there team. I really don't have a track record as an investor. But maybe if I succeed with my moneyI may join some team.What do you feel?Traden4,I agree with you. But as AI keeps on developing the guys who will be coding needto be more and more advanced. I am not sure how you can keep on learning more andmore for all your life, burnout chances are high. A few have the ability surely. But for others how do they manage?
 
User avatar
platinum
Posts: 158
Joined: May 6th, 2009, 11:08 am

How to beat machine learning/automation/big data to make money?

August 28th, 2015, 12:22 pm

It depends on what your expertise is in and if you are interested in managing a team. If you were not on the investing side, that could mean that you were in IT for many years. Your 80-hour work weeks should be worth something.But not everyone likes to manage people, even when it comes to other software engineers, so part of the question would focus on that.In terms of a non-conventional career path, it's not impossible. I know people with very strong skills in certain areas of finance, including quant work, who have worked for large banks, small hedge funds, themselves (offering consulting services), personal trading, and back to firms again, as consultants or founders/key employees once again.This depends to a great extent on network, personal reputation, and motivation though. You joked somewhere about being lazy. I suspect that it is not true, but if it is, then this type of option might require too much time and energy for you. If you tried to work with a headhunter, that could be difficult unless you find a very resourceful one who is open-minded about the more unusual paths in life. **Looking at your post again, you might consider looking into micro finance, if you want to be involved in poverty-alleviation projects.
Last edited by platinum on August 27th, 2015, 10:00 pm, edited 1 time in total.
 
User avatar
samudra
Topic Author
Posts: 730
Joined: December 23rd, 2002, 2:48 pm

How to beat machine learning/automation/big data to make money?

August 28th, 2015, 1:50 pm

Thanks platinum. Greatly value your feedback. One reason I not very keen on doing a job is if I work for a brokerage/investment firm due tocompliance requirement my ability to invest will be restricted. If I can manage 15% CAGR on my investment I will be in good shape and be able to achieve my goal.One advantage of being solo is I can take risk which normally I cant if I work.But if my investment gives me good returns and I have more confidence I may contemplatemanaging portfolio of others.Managing portfolio of others is a big responsibility. What if I am not able to deliverreturns that fulfill expectations.
 
User avatar
trackstar
Posts: 27033
Joined: August 28th, 2008, 1:53 pm

How to beat machine learning/automation/big data to make money?

August 28th, 2015, 2:01 pm

Well, that is why I would also be a proponent of a small expert team rather than a one-man shop.It is difficult to have all the expertise required for a really well-diversified trading entity - you might choose just to be in stock and bonds, or you might want to access currency and commodity markets, for example. Further, it is better to have an experienced risk manager who sits slightly outside of the system that is to be managed.One thing I have noticed in a few places is that a very wise and firm risk manager may have a background in trading from an earlier stage of his career. This means street smarts and a finely tuned sense of the markets, rather than a pure instinct for risk aversion.This type of cross-functionality can be useful in other areas of the team as well.So I'd look back to your IT experience, acknowledging that you are seeking new challenges, and see how that fits with the development of a new entity. (BTW - I do not mean this as a put-down - I have a lot of respect for the IT side of the house.)
Last edited by trackstar on August 27th, 2015, 10:00 pm, edited 1 time in total.
 
User avatar
samudra
Topic Author
Posts: 730
Joined: December 23rd, 2002, 2:48 pm

How to beat machine learning/automation/big data to make money?

August 28th, 2015, 2:19 pm

trackstar,Thank you for valuable feedback and spending your precious time.Boss there are too many super smart people like you who are in trading. And now to addto that we have the automated trading softwares. My last project was on dark pools. Long time back now things must be way moreadvanced.Anyway I don't want to trade I don't have the resource to trade really. So I am trying to be a diligent value investor and get 15% CAGR. Lets see what kind of returns I can manage.Luck has been on my side and I am making about 27-28% as of now. Hit some winners. Butit is only about 7-8 months. So still now not sure if it is fluke or I am makingthe right calls.
 
User avatar
exneratunrisk
Posts: 3559
Joined: April 20th, 2004, 12:25 pm

How to beat machine learning/automation/big data to make money?

August 31st, 2015, 2:09 pm

How to beat the thinking machines?First, we start thinking like machines to enable machines to think like us. And this is the chance: become a solitaire psychologist, behavioral economist---use machines as constructors, apply counterfactual simulation?
 
User avatar
samudra
Topic Author
Posts: 730
Joined: December 23rd, 2002, 2:48 pm

How to beat machine learning/automation/big data to make money?

September 2nd, 2015, 3:50 pm

Ex....unrisk,How do you get this name .... I completely agree with you.Only thinking (in solitude ... without getting affected by the machines thinking )and originality can save us I guess.
 
User avatar
Pat
Posts: 1207
Joined: September 30th, 2001, 2:08 am

How to beat machine learning/automation/big data to make money?

December 10th, 2015, 7:10 pm

As soon as I read about the cars driving 1,000,000 miles through ordinary traffic without drivers, I realized that the world has changed.Despite my dread, I am again re-inventing myself as an automation/machine intelligence expert
ABOUT WILMOTT

PW by JB

Wilmott.com has been "Serving the Quantitative Finance Community" since 2001. Continued...


Twitter LinkedIn Instagram

JOBS BOARD

JOBS BOARD

Looking for a quant job, risk, algo trading,...? Browse jobs here...


GZIP: On