RTS Algorithmic Trading Webinar

Discussion in 'Automated Trading' started by travis, May 16, 2009.

  1. travis



    Click on the right where it says "Algorithmic Trading Webinar".

    It's interesting (and relaxing) even though it doesn't sound like a webinar with people who would write on this forum, or engage in automated trading (the way it is meant on this forum).

    I remember hearing, at the start of the webinar, something about two different definitions of "algorithmic trading". The "New York" definition which means automating the mere execution of big orders. Then you have the "Chicago" definition, which means market making and arbitrage (which means automating decisions as well). I think there's a third definition missing, which is the one I use.

    They talk about my definition of "algorithmic trading" when, at minute 28-29, they speak of "back-testing", which is the only way "algorithmic/automated trading" is meant here on this forum. When they talk about "back-testing" they seem very knowledgeable and precise, but only from a theoretical point of view, in the sense that it seems they never tried it.

    After the first 30 minutes, you should start falling asleep.

    I was reminded of a post by someone a few days ago, speaking of the "big boys" vs retail traders. He was saying "if the big boys can't do it, then why do you expect to be able to do it?". But I think he was wrong. Also other people said he was wrong. But in general I am getting the idea that these "big boys" don't know anything about back-testing and automated trading (the way we mean it here).

    Or maybe the ones who know about it keep quiet? Because on the web all I can read and see from the "big boys" never has anything to do with our concept of trading systems.

    Also, someone was saying in another post - it's natural for a retail trader to make 50% per month, whereas the big boys will only make 20% per year because they have to move around bigger capitals (with all the consequences). It sounds reasonable and it sounds like what actually happens (whether because they have big capitals or because the big boys are also dumb). But do you agree that it doesn't sound reasonable for the average investor to give his few 1000s of dollars to a fund manager (and make 20% a year if he's very lucky) rather than to a retail trader? I realize that maybe retail traders are not even allowed to trade other people's money. I am just wondering. If anyone has ideas let me know.
  2. Having ideas about what? About people giving you their money because you think you understand algorithmic trading?

    Have you heard of alpha and beta? Big money goes after low beta and most of the times stays away from high alpha.

    You sound like a newbie. After 10 years of real trading experience you may be in a position to answer most of your own questions.
  3. travis


    A person can be very knowledgeable about all the technical details and not make any money in the markets. That's the point I was making about some of these people. The point here is to make money and not to show off our knowledge and despise those who don't sound as knowledgeable as we are. There must be something intelligent in what I said above, but you didn't get it, because you're too busy showing off your knowledge to perceive any intelligence in what other people say. I believe that you are "newbie", too, but from an intellectual point of view. I am ignorant and you are stupid. We could make a good team - so let's be friends and join forces. You could teach me some of the many things you know, and I could teach you how to use your knowledge. But you need a friendlier attitude.
  4. nitro



    Imo the reason the big institutions don't believe in this sort of trading is they think it is hooey. The only person of some repute that believes in this style of trading is Andrew Lo from MIT, although he is far more quantitative than most, and he runs a large fund and has the "sheets" to prove or disprove that it can be done.

    If he cannot convince the financial markets of the viability of this style of trading, you stand zero chance of doing so.
  5. travis


    Only one reputable guy? Interesting. And yet I believe it is not "hooey" at all (as long as we're both talking about using back-tested trading systems, and what I am referring to in my post).

    What made you think I was trying to convince the "big institutions" or "financial markets" of anything? Nothing that I wrote. If anything, I want them to keep being this way. I was just reasoning out loud, and asking questions, and wondering things.
  6. nitro


    You misunderstand what I am saying, almost certainly my fault. What I am getting at is that the reason these things aren't given more attention by the people you claim don't seem interested (the keep quiet hypotheses has almost zero delta of being true) is that they think it is hooey. That you feel the need to say that you do not think it is hooey is what I was trying to say. No one knows who travis is, or nitro. But they know who Andrew Lo is. Hence, you might start there because you may have more in common with him than you think.

    If by back testing you mean using the scientific method, of course that is practiced, especially on the sell side. The hooey part is not this "back testing", it is that some picture on a chart can somehow divine edge through the perception of patterns in that picture.
  7. Introducing "who does what" is a good way to stimulate discussion. I also like your refernces to the comparison of how the amount of money involved affects the performance relative to the offer (the alpha and beta).

    Order execution is your NYC.

    Your CHI is market oriented with emphasis on orders and imbalances.

    Personally you are approaching it from what the market offers and using iterative refinement of sorts to improve performance.

    Would you be willing to consider an approach that emphasized science and logic as part of the mix and discussion?
  8. travis


    Your language is a little too complicated for me, but do feel free and welcome to take this discussion into any direction you want to. As I said earlier, I am only asking asking questions, wondering things. So far I've learned something more, so thank you all.
  9. Someone suggested you check out Andrew Lo at Tech. His homepage has a brief list of books and publications (papers written for publication).

    I attached a few of these references (abstracts) to show a little of his path.

    He has landed in Financial Engineering at Tech.

    As time has passed at ET, several people have referenced his work (studies) as they apply to TA and to traders.

    Automating trading is something that has not gotten under way, significantly speaking. Using TA in funds is uncommon but it is happening.

    My posts are difficult to understand and follow but, it is true that I follow a lot of what may be called cutting edge stuff in the financial industry.

    As you see writing for publication has it's challenges. Andrew Lo has to get his stuff approved by the powers that be and in academia there is a lot of pressure to "publish or perish".

    In the future, a lot of how the financial industry and the academic community which is subsidized by the financial industry, will be changing significantly.

    Often the players in academia are approached to deal with problems the financial industry is having. Sometimes the "business" of these financial corporations, banks or partnerships comes to involve things that serve as "money attractants" to those who have money that needs to be grown.

    Doing automated trading, in the sense of making money, is, sooner or later, going to show up. Who is there to facilitate this opportunity is a good question.

    It will not be either your NYC or CHI example and it won't be Andrew Lo either. The reason is that there is no "hook" to attract the moneyed people or institutions.

    Last week I reviewed a 200 million dollar and 300 million dollar fund (A and B series) that was devoted to trading commodities (financial and commodity) and that used TA in an automated way. Money management was one of the prominent features.

    Since inception (2003) the return has been 164.51% and it ws stated as an annual 16.60% geometric.

    You have been posting about automation and how it might apply as part of managing capital.

    Maybe, by considering how some of the tendencies to get from "Before" to "After" in these sorts of operations could weed out how and why some financial industry operators do "not get it" and others do "get it". Maybe if a competitor is not having any downturn and is pulling in a nice steady capital appreciation then some of the "hooey" may go away for those who are declining in competitiveness.

    I know you do not want to tax yourself too much, but it may be of some value. See if you can make anything of the Lo abstracts. If so, then you may want to engage to get up to speed on what automation can bring to trading any size that is out there.
  10. This whole discussion is useless. "Big Institutions" backtests there models and systems. To give one specific example, I am very sure that one of BGI's Quant Team "backtests" their models. End of story.

    As nitro mentions, they're much more quantitative than most of ET. The most important difference is the "purpose" of backtesting. From my personal point of view, what they call backtesting and what you call backtesting are 2 different things.

    Seriously... most of the things you read on the internet or books are only 10-20% of what goes on inside the funds. The goodies and neat stuff rarely reaches the retail level.

    Thing that people need to get over is:

    1. Fully automating a trade is not goal or step. It's a pre-requisite for a systematic trader. If you don't have the ability to do so, you're still a newbie. ATM (Automated Trading Models) Machines fail. We have the ability to do so but we choose not to do it.

    2. % return doesn't say shit about a model. The fact that people are comparing and having a "Us against Them" discussion, just tells how newbie some people are.

    3. Stop using "Big Institutions". Give a specific instance or a fund. If you can't do so during a discussion, don't even start. It's like saying, "Retail traders are losers".
    #10     May 17, 2009