Will Algo Trading Take Over?
Was just talking with a few people from work, who have been in S&T for a fair amount of time. We seem to agree that algo trading is the way to go in the near future. Computerization appears to be the trend, since it's cheaper and faster. Banks already spend huge amount of money on building their algo teams.
However, is it possible that algo trading is really going to "take over"? Are regulations likely to backlash?
it already has 70% of trading is Algo.
Wolverine19x89 will "take over" trading in the future. Watch your back.
You're a computer?
What credentials do people have? The ones that administer the algorithmic trading. PhD Mathematics; i.e. quants?
They don't need to, they're not interested in fundamentals, just analyzing relationships / trends.
Algo trading is intrinsically reactive. So it'll get rid of really basic market making, assuming that hasn't happened already. But any semi-smart quant will admit that there's a lot of trading that a) algorithms can't replace, but that b) algo trading has made much cheaper, by narrowing spreads.
http://www.youtube.com/embed/ed2FWNWwE3I
man these thread are all same
yo algo takin over herp derp derp
herpa derp in your yerp moderator- "PHEW THANK GOD SOME POSTED THE WEEKLEY ALGO TRADER THREAD" herp derp derp *front page sticky
less then 20 posts
TO THE FRONT PAGE!
what kind of sales and trading you guys do? used car?
There is no system you can do without an algorithm that I can't do with, as long as you are using only information from markets (i.e. insider stuff or un quantified factors).
Feel free to think of one though.
Not strictly true. A couple years ago, Jim Simons claimed that no single signal Rentec used was profitable after transaction costs. You need to use multiple signals at once to narrow things down to net profitable trades.
Plus, there's data quality: a ton of backtested strategies look good because someone made a typo in 1992.
Cleaning data is a fixed cost, testing strategies in the real world is cheaper if you're already running multiple strategies at once, and most good quants have great instincts about what to investigate.
So it's only true in a really limited, technical sense that any algo strategy can be duplicated.
In other news: wall street may be developing a less than stellar reputation amongst the masses...
Algo trading will only grow. It merely depends on how one defines "take over".
If 70% of all trading is currently 'algo' as was said earlier then I believe it could approach 90%.
Quants are not programmers. I should know. While quants for the most part can indeed do a lot of specialized programming, what they really tend to be are highly analytical/research-oriented people, mostly with higher analytical degrees. Trading at it's heart is highly quantitative (derivatives, pricing of exotic products, rates, options, etc). I don't think it's possible to separate finance from quants, not anymore. Everyone has some sort of trading model/strategy.
Algo trading is simply codifying the decision making process that lots of traders use, cutting away the middle-man.
Yes, I think it will take over - manual traders are simply too slow to respond to market movements nowadays.
I agree with this somewhat. But every quant job i've seen posted, requires pretty strong coding skills in C++, C#, Java. Of course, I imagine that any masters or phd student in the hard sciences/math/econ have picked up some programming. But at most hedge funds and prop shops, there isn't much of a distinction between a quant and a programmer. Software developers, on the other hand, are their own separate group, but nowadays all quants must be able to write their own algorithms after doing the proper research and backtesting.
I see your point. It all depends I guess. There's a new "job description" called Quant-Dev, where you're a programmer who plays with algorithms and stuff, but not quite. What we now refer to as Strats (strategists) are the core financial Quants, and most of them have reasonable exposure to finance.
You're quite right - these guys indeed have a lot of exposure to core programming, unsurprising since the entire business is meant to take the core decision-making out of human hands/minds.
In the Algo-houses, my experience has been:
Algo-trader (formerly a quant) >> Quant/ Strategist >> Software programmer.
Algo Traders tend to make the decisions, Quants/Strats basically analyse data and present their findings to the Trader, then the programmers.... well, they program.
You still tend to have one or two traders as well as risk guys in the team to monitor PNL and stuff.
cash equities trading is boring anyways, all its good for is executing derivs hedges. Trading at a bank isnt about execution, its about risk management, and until someone can build a machine that can manage a derivatives book better than a human traders are safe. Unless of course you trade cash...
Computers will never take market share away from the physical trading shops.
I would love to see Getco try to write an algo for this trade
http://www.ft.com/intl/cms/s/0/93aecc44-d6f3-11e0-bc73-00144feabdc0.htm…
If that link doesn't work then try this:
https://www.evernote.com/shard/s1/sh/4d816088-8f1b-4d2b-827e-fc59e5b17a25/9f597e8e3a0673cbe9986caec51af902
Algos may take over trading, but they will never take over investing
Fantastic video. Wilmott seems like kind of a tool, tho obviously drumming down the math. Derman knows what's up. That software-gone-oyster guy, yikes, broken man. Was it his role in the crisis or just the culture? Perhaps the former blamed on the latter. Very well done tho.
I saw it ultimately as all three coming to terms with their purpose. Which is difficult b/c they're all more scientifically driven than monetarily like the rest of the street. But must accept that their talents lead away from their roots. Derman went to say that the models too quickly detach from reality. In the eyes of the beholder. We are all here for the money.
Derman pointed that modeling and quant roles are transitioning to algo. Do you understand the distinction? Flat numbers to more processes. Thus we are approaching limits of artificial intelligence, which won't advance as rapidly as science previously has in finance (don't see how the gov won't have the edge). Human creativity remains an edge, not to mention the current algo backlash. I think that future trading will indeed require more engineering conscious minds than before, but the heart will still rely on DEALMAKING.
As long as liquid markets use a FIFO (first-in-first out) fill allocator, the biggest advantage algo trading offers is execution speed. Certain shops have implemented microwave radio towers that offer ~8 microsecond execution, which is basically the physical limit (think the speed of light). This is not a sustainable competitive advantage.
I believe that as this speed becomes more widely available, discretionary traders (non-algo) will regain a big edge. Half of my office are algo guys, the other half discretionary - the algo guys are smart and can talk statistical edge, but they really can never explain the inside-outs of the market as well as the fundamental guys. You have to understand that the whole point of algo trading is to be able to cover a wide range of markets with as little effort is possible - I don't believe algo guys will be able to keep an edge over guys who follow a few markets full-time.
I'll concede that John Arnold closing Centarus discredits a lot of what I've said.
Algo trading (Originally Posted: 09/21/2011)
Hi,
I recently attended a seminar on trading where they claimed that 'algo trading' will soon be replacing the entire trading desk throughout wall street and that there'll be no requirment for an actual trader to do all the buying and selling of stocks.
How true is this statement? I was always planning to complete my CFA and become a trader etc etc (currently in ops). But will that be unnecessary as all these IBs will be looking for in the future are programmers etc to do all the trading?
Thanks, KJ
Extremely unlikely. We've been doing more and more "old fashioned" block trades where traders are willing to commit capital in order to avoid being seen in the market and gamed by HFT algos. Traders also provide market insight that computers can't provide. At least not yet.
So, becoming a trader is still a viable career option. That said, you don't need a CFA to become a trader.
Thanks.
what other options do i have to move from BO other than a CFA or an MBA?
algo vs discretionary (Originally Posted: 11/30/2011)
Hey fella monkeys~
Correct me if I am wrong; but algo trading is mostly done in prop trading these days? If not, is this a big trend in trading these days?
If so, who / what industry still do discretionary trading?
Thank you all for the info in advance.
Surivorship bias/news is reported, old news is not.
Plenty of conventional shops around, just the search engines will find the algo ones. Algo is hip, but unless you love it, is rapidly become a very saturated market. Lots of big names losing big money in it as well as others making it. Not hard to guess who they're making money off.
Main ones i've seen are energy and futures markets, but there will be ones everywhere.
Algo vs. Human (Originally Posted: 12/11/2014)
A genuine question:
Do Algo trading out-perform humans nowadays?
Or simply put: do markets trade more on short term momentum/ TA rather than fundamentals?
Can anyone share their experiences or back-testing using either approach?
Your clarification question is pretty different than the first one. Not all algos trade on momentum/TA/things that violate weak-form EMH, and not all humans trade on fundamentals.
Okay, apologies for the confusion.
Would you say most of the algos pick up signals and trade without a person's decision or emotional influence is more effective?
Since this is a really broad question I would recommend picking up a book on what algo trading is really all about. "Algorithmic Trading and DMA" is a great overview of the different applications of algo trading, I would highly recommend this book to anyone interested in the field to get started with.
Trader jobs taken over by computers (Originally Posted: 10/23/2010)
How long before d-day?
this question and like-minded questions are being asked with ever-higher frequency on wsoasis. the question itself is to general to answer and you'd have to know more about financial computing to ask a reasonable question. But to answer the general question: don't know. but as of yet, the technology to do so isn't known to exist (and with philosophical evidence some people don't expect it to ever exist).
Modi nemo iure eos consequatur. Cum nesciunt sed qui reiciendis voluptatem eaque.
Quis aut iste non ut perspiciatis. Et pariatur dicta ad quisquam sunt iste. Quod esse aliquam non enim doloremque libero. Quae voluptatem natus in ipsum hic dolor. Iste molestiae voluptatum voluptatem et quia officiis necessitatibus.
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...
Accusantium a ea adipisci sed praesentium rem et aspernatur. Nobis possimus et sequi vitae. Quaerat quam aut eos. Excepturi nobis excepturi ullam rem. Recusandae eaque earum aut quam tenetur qui excepturi.