With respect to all the other posts I am unknowingly ripping off based on the mountains of clickbait articles and listicles that I read daily on the death of trading, the inevitable move to full automation of markets, compression algorithms, etc., it’s hard not to feel like demise is just around the corner. As if the previous harbingers of doom weren’t enough there is also stricter regulation, changing risk preferences, and the fact that beating the market is statistically slim.
Granted I’m merely an analyst paying my dues after bootstrapping out of the insurance industry via an MBA, but paying your dues is tough when the goal you are working toward may be disappearing before your eyes.
Of course I am also crushed by the irony that in trying to spark conversation and opinion on this topic I am soliciting feedback through a click bait type title.
I want to hear from you. I work in an area of commodity trading where the market is shrinking. Though it will be a long slow decline – strike that – should be a long slow decline. What if it isn’t? Five years ago coal producers never dreamed that they would be innovated out of a job, and the market would view billions of tons of coal reserves in the ground as essentially worthless.
So I leave you with this Monkeys, how close is trading’s judgement day? Is a market that primarily consists of automated trading strategies pitted against each other really a market? How does a market function without the human element? Does Moore’s law top out and make this whole thing moot? Does P equal NP? Does any of this matter if it’s all one giant matrix like computer simulation as according to Elon Musk?
-Out
What side of the insurance business did you come from? Im trying to explore MBAs and potential post-MBA opportunities myself
Specialized commercial underwriting.....stay on the hustle and best of luck to you. That's the only way I can explain my move.
Thanks, I am on a team underwriting financial lines. Do MBA programs have any interest in underwriters? Sorry to go off topic but if I transfer I want to be in something that is still stable. Algorithms are used to make a lot more underwriting decisions now adays as well
Depends on the desk. Unfortunately, commodities trading at a bank isn't the best place to be right now, you may consider lateraling to a physical trading shop or supermajor.
Why is commodities trading bad right now? Because of the volatility of oil?
It depends on the sector and type of trading.
Physical Trading in commodities is still dominated by humans. It's all about relationships, trust and creating a partnership. An algo cannot go and talk to refiners/producers etc and write up a contract agreement to supply XYZ. This industry will always have human elements to it. The derivatives side of the commodities business leans more toward the algo side, but only to a certain extent. There is only so much algos can move commodities markets because at some point, suppliers/consumers will just hold their contracts 'til delivery if the prices are too high/low.
Market Making is a combination. The traditional market makers have gone, except for in the fixed income world to some extent. I work as a MM and my job consists of part trading, part risk management, part coding. We use a lot of VBA and Excel to build out models and then have our IT department build our platforms, systems and Algos. In the future, I can see the perfect candidate being primarily a trader but who posses good coding skills too.
When I first started out I had a crisis of confidence too. I asked the head of trading if he still saw himself working in the same field in 10 years and if he thought we would be replaced by computers. He told me that he once thought the same, but that in his prior firm they had run MM algos alongside their traders to measure performance, and surprisingly they found that the humans were able to make more money over time due to their ability to use judgement. If a market is dumping, an experienced trader can feel it, you can sense that squeeze that's coming. An algo cannot, it will just follow it's pre-defined code. In such scenarios the humans can make that little extra.
Saying that I'm sure many will disagree. In my view, you need to learn how to code so that you can be a complete package and adapt to the changing times.
To be fair, a lot of this "judgement that algos just can't do" is going to start happening, machine learning is getting easier & easier all the time d/t massive increases in parallel processing power and new focus on FP16 computation in GPUs. That "judgement" thing your brain does is just an algo in and of itself, one that's very complex albeit, but certainly not impossible to replicate given enough processing power.
Agreed, judgement will definitely be replicable at some point. Strong AI will probably be achieved in our lifetimes. After that point, who knows what will happen? Whole industries may be wiped out, leading to massive unemployment and deflation. I think it's probably further out than the popular press, click-bait articles would presume. With regard to trading, I think that the shorter term the trade, the easier it is to replace with an algorithm. Markets have never been particularly good at discounting events more than 18 months out. Algorithms that are trained and backtested with the same data will probably learn the same sorts of behaviors. There's also special situations, and doing physical legwork to find better data, that could give you a potential leg up on algos as well. Maybe algos will figure out ways to pick up on a lot of these situations. But the point is, algos will have flaws, and there will probably be ways to exploit these flaws. That being said, when strong AI comes, we're all screwed anyway, so what's the point.
Thanks for the perspective. What coding languages have you studied specifically?
Thanks all for the responses, I am in the physical side of the business and for the most part I agree that this niche of trading is still dominated by relationships and judgement.
Although I think physical commodities will also go the way of the AI trader in the future as well once those companies decide to make the investment.
Three comments/questions:
crazy that algos are making 'well-educated' traders/bankers feel they'll be replaced by robots. Can you imagine what an entire town dependent on manufacturing or coal mining feels like; they have no fancy degree or understanding of global economics and finance, let alone ability to code (yet), to fall back on... Makes you think.
You got an MBA and are now an analyst at a bank? I thought you typically come out as an associate. Or are you at a fund that only hires analysts with MBAs? Really not trying to offend, I honestly don't know the structure and am curious.
To any older Apes out there, was there the same or similar fear of your job being taken by robots back in the 80s... 70s???
There has been a lot of talk about machine learning etc, but then don't we reach a point where we basically create algorithms that start to behave too much like a human, at which point they become prone to the same mistakes a human does. What if each machine begins to learn in it's own way, thus building it's own individual view on the market. Surely then it will be like humans, some machines go long whilst others short based on their learned assumptions. The whole point of the markets is to facilitate trade in an auction-like environment. When people talk about machines replacing humans in the markets they always have this belief that it will be some magical world driven by AI which only ever win. Well who is on the other side of the trade then?
Haha - good point! There are people in much more precarious situations than ourselves.Coming from someone with a scientific background and some understanding of quantitative trading (internship with quant fund) I think allot of this AI stuff is hype.
The majority of quant trading revolves around analysing market data to predict if the future price will go up or down. Computers are always going to be able to do this better than man. But algorithms are nowhere near the point of understanding complex concepts like macroeconomic theory, phycology, business strategy, ect, ect, which play a much larger role in predicting prices for medium to long timeframes. This will always leave room for human trader who can navigate these waters that algorithms don't understand.
Also, this AI thing isn't new. Every decade since the 1950's has claimed that AI will become a reality very shortly; yet they have been wrong every time.
And even if we do reach genuine AI in our generation, we should rejoice not cower in fear. AI isn't just going to eliminate the roll of the trader; every profession from doctor to professor, lawyer to CEO will become redundant. Even programers won't be needed. This isn't something to fear, it's something to be excited about.
hey, yo, cool avatar, op.
Et molestiae aut quidem autem. Eaque et repellendus enim doloremque tempore quia. Sunt porro delectus ipsa magnam perferendis sit. Et consequatur nostrum et assumenda et sit deserunt.
Est nulla sed eum aut aliquam facilis. Porro aut vel reprehenderit.
Perspiciatis unde vitae fugiat sint enim. Nam sed enim sunt in nulla voluptatem. Blanditiis ipsam voluptates enim rerum est soluta maiores.
See All Comments - 100% Free
WSO depends on everyone being able to pitch in when they know something. Unlock with your email and get bonus: 6 financial modeling lessons free ($199 value)
or Unlock with your social account...