% Accuracy

GIG · Jul 29, 2003

OddTrader,

I logged 25 hours, about 10 weeks ago. And yes, this is an undergrad research project. This was deemed as a proportionally reasonable amount of time based on the course timeframe. I haven't done any other WWW searching since that time.

Regards,

Brandon

damir00 · Jul 29, 2003

so how much money did the trades make?

Quote from GIG:

3 out of the 4 algorithms, over a long term period, don't do all that well (~47% accuracy). One algorithm with a given parameter set does well (~52% accuracy) over a long term period.

More...

OddTrader · Jul 29, 2003

Quote from GIG:

OddTrader,

I logged 25 hours, about 10 weeks ago. And yes, this is an undergrad research project. This was deemed as a proportionally reasonable amount of time based on the course timeframe. I haven't done any other WWW searching since that time.

Regards,

Brandon
More...

Try again searching some commercial sites, besides acadamic sites.

Many practical and informative studies have been done by some commercial/industrial researchers who are earning very good money from customers.

bubba7 · Jul 29, 2003

Quote from GIG:

snip

Yes, the two primary factors used in analysis are price (close) and time (or the manipulation thereof). These are the only two parameters which can trigger a buy signal. I also log additional information, such as the next day's open (after a buy signal), the %change between the intraday open and low, and %change between the intraday open and close. I hope to draw some type of conclusions by observing this data.

Check out Welles Wilder and the RSI original he came up with. You have a bias in your approach he does not.

"The very difficult challenge you created was choosing 500 stocks and making that sample a random one. There are no financial incentives for ever doing that any time."

That's a very good point. My primary reasoning for choosing companies at random was to avoid the potential for overfitting results, tailored to a particular sector, for example. Also, I didn't want to target companies who share a characteristic, not found in other groups, and get results (for better or worse) that could potentially innapropriately reflect the algorithm's worth as a predictive tool.

You will find over many years that sectors rotate and the sector is, in effect, actually controlled by people rather than the sector corporate focus. Strangely, you will also find that the variable most rewarding to use for making money is not price. The pockets you re looking for are more easily discovered by two other direct market functions. Fisher at Harvard is a good contact. with a three variable system where you deal with direct variables over time, you wind up in a very good place for picking off market performance signals with great significance.

Optimally you can look at 8 places because you have three variables. Now you are constrained to two places and you are dealing with one of them.

I understand your point, however, regarding the use of other factors as predictive measures. Using only price/time won't be the 'end all' of market analysis. I really enjoy doing this kind of stuff, so I imagine I will have many years left to try a variety of techniques. What I have done is built a core framework (from a software point of view), which can be extended to an n-degree of complexity as needed. Ultimatley, I wanted to start small and simple, understand the relationship (or lack thereof) of primitive/raw metrics, and from there continue to exploit market behaviour with more complex techniques. It's been a pretty good learning experience, which will be invaluable for future work. At this point, I wouldn't want to have a solution so complex (even if it worked well), as to not have a reasonable explanation for the solution's behaviour.

The macro you are working with does have, as you say, behavior. You have gone far enough to know that the solutions you hope for are, at best, woven into something that is momentus and on the surface does not yield results that are significant. The entity displays mostly whiteness. I hope you get to a conclusion sooner or later that there are no fruits to be had. the literature is stubborn. Failure is not documented in our social mileux. I always look for failure in each endeavor I document when I am working the cutting edge. In your case, you recognize at this point that you are going to change the problem to succeed and it will not be required to complete your present work.

The micro analysis is definitley starting to formulate, (at least in my head right now). It's just a matter of time before I develop a more focused solution.

You can't imagine how fuitful it will be. There is the "too good to be true" part of every distribution. I have seen guys have to take brief vacations as a consequence of having such a hit. It literally changes career directions. In the financial industry, it is not common for sea changes in mostly any effort; but it is possible for sure. The institutionalization of ideas is a neat phenomena. You may look forward knocking off a couple.
I am laughing here thinking about a couple. this place doesn't like stories though

Quick question: If 50% is a classically poor vantage point for making money, what would you consider to be a good vantage point (in your own personal opininon)?

I am so glad you asked. Making money is a "threshold thing". you start to observe from below (before) and you measure the "advent" and the "extent". Furthermore, each of these two phenomena are directly related to our very well developed reasoning processes and understanding of how everything works here on Earth. The markets operate, it turns out, in more than one mode as is expected of a "threshold" functionality. Thank God it is a on/off or go/no go phenomena, which precludes complexity.

by measuring the market and determining when threshold are arriving, you secure your vantage point to then measure the "advent" (the metrics required to turn on and function for making money) then you cruise along monitiring he "extent" of the phenomena with, it turns out about the same metrics.

Here is the punch line. You will find the market "moves". You metrics will become definitive of the operating points (which can be expressed two or three dimensionally). Fortunately, and as in nature on this Earth, the movement has a defining characteristic. You can think of the possibilites for movement and easily recognize that random is not the nicest to encounter. In fact the opposite occurs. The movement is the most orderly that is possible. This makes what you are looking for as well defined as possible mathematically.

To reflect for a moment. Visualize that most effort is looking at the random movement as the key possibility. The greatest bunch of capital that backed that vision went down the tubes in Greenwich, Conn after the "club" bailed them out at 300 million a pop each.

Your nonsymmetric accumulation of price data is an aid here in looking at your stuff. It takes guts to not collect balanced data. but you get to look at things others do not. As you consider the assessment of trends beginning with a "threshhold" to go to "advent" to go to "extent" ,you wind up closing the loop by looking at the end effects of trends. The task therefore is to know how the market operating point (metrically defined) movement is forced to an ending. The forces occur as blocking alternative possibilities it urns out. After passing the threshhold, you have all possibilites; the possibilites close down sequentially.

This is a modeller's dream come true.

Skipthe 50%. Go to the place before thresholds and do your metrics on the occurance of activity by crossing the threshold.[/color]

Regards,

Brandon
More...

Log in or Sign up

% Accuracy

GIG

damir00 Guest

OddTrader

bubba7