Algorithmic Trading

Quote from nitro:

Ok.

And what edge do you think you can have over machines that are basically making markets based on their current baskets/inventories and where the ES/NQ/YM are trading? In other words, even if you knew the exact algorithm by which these things were basing their trading decisions on, how would that make you a better trader?

What I trying to point out is that understanding the machine will only help you in programming another machine maybe try to do the same thing, but I doubt it would help you as a human being at all. The way you trade now is the only reasonable way, outside of technical analysis, that a human could trade.

There are only two ways to trade outside of arbitrage (statistical or otherwise), to take liquidity away from the maket and therefore pay the B/A spread (for whatever reason, breakout, etc), or to add liquidity by making a spread, which is mostly what computers acting as MMs are doing, except when they come in a buy/sell baskets for index arbitrage where they are then taking liquidity. The MMing is almost 100% the domain of well capitilized MMs.

People like Bright traders try to do pseudo MM type strategies like the Opening Strategy, or the Trade Through strategy but that is a huge patience game and is not for everyone.

What I am trying to get at is that, imo, very few of the players can play these games, although three people that I know of on ET claim to be able to trade stocks 100% by computer, and no one in their right mind is going to tell you how they are making money doing that (if they are, which is a big IF)

Most of the reason the markets seem so confused now is that there are so many different ways that stocks get kicked around by their derivatives, whether it be Futs on them, Futs that include them, Options on them, or ETFs that include them.

nitro

Interesting ... I agree that I have yet known one case where a small sized (i.e., not in the top 10 in volume) MM have tried to do full automated MMing. However, although one have no insight into the automated algorithms used, their behavior can be modelled for non-MM systems to be taken advantage of.

In my opinion, stocks have too many possible ATSs, but for single listed products, the automated system can only perform limited information hiding (unlike stocks, where smart splitting and routing is becoming a high art), making it possible to perform pattern analysis. However, such patterns disappear so quickly that the only thing that can take advantage of such a pattern is another automated system, one that must also maintain a sufficiently complex pattern database, complexity that is probably out of reach of most (if not all) individual traders.

Such high frequency data analysis and trading more closely resembles game theory rather than any finance theory or TA.
 
Quote from rufus_4000:

.... However, such patterns disappear so quickly that the only thing that can take advantage of such a pattern is another automated system, one that must also maintain a sufficiently complex pattern database, complexity that is probably out of reach of most (if not all) individual traders.

...

..Or are these short term "patterns" actually valid ?

That is , perhaps they are short lived precisely because they are manufactured by all the ATS's ?

..It really is amazing how a simple and uncomplicated price volume chart still seems to work fairly well in the face of all these ATA's.

..Something to think about ....
 
Quote from Scottsdale:

The edge I can gain over a machine is achieved when I figure out what that machine is reacting to. A simple example was trading the large cap nasdaq stocks off of the S&P and NDX futures back in the old SOES and SelectNet days. Because I knew that a huge amount of index arb was being done, I could trade in front of the arbitrageurs when cash and futures became out of line. By trading a few NDX stock which were highly correlated to the index, and through my superior order routing knowledge I was consistently faster than the arbs who were trying to buy or sell a much larger basket. Also my feel for the momentum allowed me to anticipate when programs would kick in before they actually did. The problem I encounter with today's nasdaq is that it's not clear what variables are triggering automated buy/sell decisions, and there is no advantage to being a highly skilled order router. However, I do believe that with the knowledge of what variables a machine is trading off of, my observation skills would allow me to anticipate the triggering of buys/sells, and I could then get in front of them.

Great response.

It is not difficult to rationally go to work on figuring out how to have an advantage over either others or the market at large.

This thread and your example in it are show how reasoning is a natural part of positioning oneself to make money.

Once any sequence of activity is realized and understood, it is the point of departure for creating an approach for making money.

With an approach in hand, designing a detection system for the precedents of entry is not usually even rocket science.

Your post illustrates how you consider this opportunity and how those who inquire of you have difficulty in understanding this concept. It certainly appears to be a very deep line in the sand at ET.

There is a thread on chat rooms running that is a superb example of a failure to understand this reasoning and analysis concept.

Anyone journalling who is making a strong effort to collect information and data on trading should make a very strong effort to come to understand what precedes any and all market actions he is currently taking.

Eventually each known precedent leads to its predicessor. From this walk backwards in the approach used it will appear that the market follows common sequences.

Gradually a total map of possibilities can begin to appear. This is where a deeper sort of reasoning comes into being.

Cause and effect can be examined as an alternative for analysis. Once is is seen that the paths of the sequences behave as qiute independent entities that seem to only converge and have common end points, two things come to mind:

What allows this isolation to exist? And what happpens after the end points?

The discovery, finally, that sequences are guided as much by events not happening (what wasn't that?) as what appears to be cause and effect, leads a person to construct the pragmatic alternative to the sequences.

It is not anomolies which would be construed as a study of trivial cases which often drives marcro analysis.

If one turns to vector analysis, the space of these sequences begins to appear. At this point the minimum number of variables required for market analysis has been set upon the table and all the work done before, it if lacks all the significant market variables, can be revisited.

So it may be seen that an alternative avenue, the collective consideration of amomoly algorithims does define the intersticies of the pertinent money making space.

The nodes of the sequences form the meat of the vector space; anomolies fill in the vaccum around the nodes.

If one were torecognize the potential of this knowledge and wanted to begin, knowing that sequence endings cluster around just several common nodes, would be the place to walk backwards or forward from these terminals.

What just happens before and after these principle nodes is how to find out how to make the most money in the shortest time.

It is very noticable, also that there are fewer voids in the space around the common endpoint (and subsequent starting point nodes). Voids being a description of the isolated anomolies that started this thread.

Obviously the macro people are determined to mine deeper and deeper into the world of anomolies. You see the convergence here in two streams of thought. the Prop trader focus (Bright) on anomolies and now the mechanical operations of the heavy hitters.

Not obviously, but very very important is what this means to those who operate in the high money velocity trading (this is not a reference to high velocity trading(it does not have a high money vlocity)). What is means is to accept the prop and mechanical stuff as "noise" signal and to filter it OUT. Thing of it as filterable by simply using a filter that is the opposite logically of a momentum filter.

What is important for traders to begin to think about is learning how the market works. when you see the true basis of cause and effect, you have a free ride the rest of your life.
 
Quote from Drock409:


the WORST of them is up 100k on his boxes this year, the others are up many TIMES that

obviously they are not out there tellin eveyone their stragies...quite opposite...2 of them are in dedicated rooms that NO ONE...and imean NO ONE can enter...period

we know they are out there....there is even "box wars" goin on...where there are boxes that look for other boxes to "trick" them..high bid, high bid...smack


1/ 100k on a 1m account isnt much is it. 100k on a 100k account is different.

2/ ever heard of transparency? exchanges go to great lenghts to ensure this - one of the benefits of screen trading. id love to know how they determine which 100 lot orders are from which algo's. hope no one is spinning you a yarn on how great they are.


all this algo paranoia makes me laarf.

there are only a handful of original algos out there - developed by a handful of people. as they go from bank to bank to fund, they just put an adjustment in to reserve the prior companies exclusivity to their intellectual work.

most are only there for efficient order execution of buy side orders. very few are speculative in nature. those that are are probably no better than the average et's system codified - like buy opening b/o & hold/whatever.

if you dont like the crushing volatility of algos, and yearn for the ranges of yesteryear, go trade something else that isnt so liquid. you wont find the efficiency in those markets and when they break - they break.


nobody says you have to trade es, goog, msft, usd/eur whatever.

come on - be original.
 
Quote from rufus_4000:

Such high frequency data analysis and trading more closely resembles game theory rather than any finance theory or TA.

It does not ressemble, it is!


Quote from Grob109:

It is not difficult to rationally go to work on figuring out how to have an advantage over either others or the market at large.

This thread and your example in it are show how reasoning is a natural part of positioning oneself to make money.

Once any sequence of activity is realized and understood, it is the point of departure for creating an approach for making money.

[ Blah blah blah blah... ]

Not obviously, but very very important is what this means to those who operate in the high money velocity trading (this is not a reference to high velocity trading(it does not have a high money vlocity)). What is means is to accept the prop and mechanical stuff as "noise" signal and to filter it OUT. Thing of it as filterable by simply using a filter that is the opposite logically of a momentum filter.

What is important for traders to begin to think about is learning how the market works. when you see the true basis of cause and effect, you have a free ride the rest of your life.


Dude, you should save your arrogant pseudo-scientific vocab for other circumstances. Maybe you should think of writing a book. "How to become rich by trading high velocity money in vector spaces surrounded by vacuums filled with anomalies" would be a big hit.

Filtering noise out of signal, I would start by filtering out your posts.

(Still waiting for your maths and theory, by the way)
 
Quote from Drock409:
we know they are out there....there is even "box wars" goin on...where there are boxes that look for other boxes to "trick" them..high bid, high bid...smack

I used to do that when selling all the time... since they scampered like rats on the offer. Put in a sizeable bid under the market, wait for the bidding war, and then whack-a-mole baby!
 
Not obviously, but very very important is what this means to those who operate in the high money velocity trading (this is not a reference to high velocity trading(it does not have a high money vlocity)). What is means is to accept the prop and mechanical stuff as "noise" signal and to filter it OUT. Thing of it as filterable by simply using a filter that is the opposite logically of a momentum filter.

What is important for traders to begin to think about is learning how the market works. when you see the true basis of cause and effect, you have a free ride the rest of your life.
While this is all good stuff to discuss and cogitate over, actually DOING IT is an entirely different matter.
I've built a few trading systems over the past few years, and one characteristic that was obvious: duration of trade.
The shorter the trade duration, the higher the percentage of winning trades was possible...BUT, the average winning trade-to-losing trade ratio went down. I had systems where I could actually "tune" this trade-off....it was uncanny to see the equity curves change. Unfortunately, the higher percentage winning systems also suffered less-than-stellar performance stats because the number of opportunities was much lower (less trades).
The other thing I learned from this frustrating experience: position sizing is a huge factor. However, knowing WHEN to use large size vs. average size was always a conundrum. Using the equity curve sometimes helped, but wow, the drawdowns would get fierce as size was applied.

Based on the above, it's most likely these high frequency trading systems are using a combination of large size as well as short duration to gain a consistent tradeable advantage.
Huge demand for those that have the know-how AND the degree..only PhD's need apply:
http://www.hedgefundcenter.com/employment/job_list_detail.cfm?job_id=4523
http://www.hedgefundcenter.com/employment/job_list_detail.cfm?job_id=4533
 
I wonder if anybody has set up an algo to track the single contract traders under the assumption that they will be wrong most of the time?
 
Quote from syswizard:

While this is all good stuff to discuss and cogitate over, actually DOING IT is an entirely different matter.
I've built a few trading systems over the past few years, and one characteristic that was obvious: duration of trade.
The shorter the trade duration, the higher the percentage of winning trades was possible...BUT, the average winning trade-to-losing trade ratio went down. I had systems where I could actually "tune" this trade-off....it was uncanny to see the equity curves change. Unfortunately, the higher percentage winning systems also suffered less-than-stellar performance stats because the number of opportunities was much lower (less trades).

Now that's interesting...

Are you saying that periods with shorter trade duration were more noisy? or markets with shorter trade duration were more noisy?

If the former, you should look into autoregressive conditional duration (ACD) models as a way to identify periods where more noise trading occurs.

I'm also trying to find this paper:

http://ideas.repec.org/p/fth/louvco/9789.html

It seems to investigate something similar to what you just mentionned. I've seen a presentation made by the author on the subject, and it seemed quite promising...
 
Back
Top