Econometrics and practice
- Thread starter DT-waw
- Start date
Interesting paper DT-waw, I skimmed it and will re-read later.
For your trouble and the discussion:
http://econometrics.wiwi.uni-konstanz.de/CoFE/Papers/dp99_18.pdf
I haven't read this yet but it addresses the difficulty expressed on page 7 of the paper you posted, dealing with stationarity.
Although the paper you posted uses just a one day look ahead and is built around that paradigm, I don't know of anything that a three year test would reveal that would alter the results as reported. The reason I say this is because the authors are inching along using the last four days? of data to look one day ahead. Even though the time series of ticks may be non-stationary at this short frame of reference the fact that everything is reset? on a daily basis to conform to the current trading environment negates the non-stationarity issues. Thanks and good trading.
For your trouble and the discussion:
http://econometrics.wiwi.uni-konstanz.de/CoFE/Papers/dp99_18.pdf
I haven't read this yet but it addresses the difficulty expressed on page 7 of the paper you posted, dealing with stationarity.
Although the paper you posted uses just a one day look ahead and is built around that paradigm, I don't know of anything that a three year test would reveal that would alter the results as reported. The reason I say this is because the authors are inching along using the last four days? of data to look one day ahead. Even though the time series of ticks may be non-stationary at this short frame of reference the fact that everything is reset? on a daily basis to conform to the current trading environment negates the non-stationarity issues. Thanks and good trading.
... refers to this as counting.
Statistical studies / correlation studies to achieve an edge.
I think it helps if you can connect the statistical thesis to a real world phenomenon.
Otherwise like TA, when you throw in a large number of time series data into a machine, by sheer volume alone, some relationships will be generated that will have great bayesian / correlation etc results .... but don't necessarily mean anything for the future.
Statistical studies / correlation studies to achieve an edge.
I think it helps if you can connect the statistical thesis to a real world phenomenon.
Otherwise like TA, when you throw in a large number of time series data into a machine, by sheer volume alone, some relationships will be generated that will have great bayesian / correlation etc results .... but don't necessarily mean anything for the future.
http://www.fractalwisdom.com/FractalWisdom/strange.html
If, like Vic, you have thought in terms of positive and negative runs of a market then the first picture in the link provided will show you the bi-polar nature of markets, in general. The second picture, is a rough approximation of what a probability tree looks like and can give an insight into the expectancy of a run reversing or continuing after a given number of occurences. Of course, for independent events the probability of a reversal is always the same. But, who says market returns are independent?
bdixon,
you have to go a little slower for this student
I don't get the pictures, but then again, I am nowhere near Niederhoffer's depth of mkt understanding.
The top picture is representativie of the dual nature of market activity, some days Bullish other days Bearish. The bottom picture shows what happens when a single path diverges into two paths and what happens when they diverge. This is not unlike the branching of a tree of probabilities calculated from counting the sequence of events of a repeated experiment. Say, your experiment has three or four different outcomes after completion. Keeping track of the outcomes and how they are reached using a probability tree is a way of increasing your knowledge about each step of the experiment.
I'm still somewhat in the dark.
But its probably because of perspective - there are a thousand ways to look at the market and because I'm locked into one I have trouble seeing others.
I don't get why the probability tree becomes so funky in the second picture.
But its probably because of perspective - there are a thousand ways to look at the market and because I'm locked into one I have trouble seeing others.
I don't get why the probability tree becomes so funky in the second picture.
That is a bifurcation diagram, but it resembles a probability tree. Probability trees are kind of like family trees...sometimes they get all weird depending upon what events occur.
http://www.emporia.edu/math-cs/yanikjoe/Chaos/Bifurcation.htm
http://www.emporia.edu/math-cs/yanikjoe/Chaos/Bifurcation.htm
Sorry I will read the article in one or two weeks or maybe this we - 'cause we are very busy with all the improvment of our trading platform - and I will comment upon it.
Quote from DT-waw:
"I have been struck recently by the disconnect between the worldview expressed by these economic and finance papers, and the view that I was seeing by standing on trading floors and talking with investment professionals"
I have a similar impression. They're using entirely different languages. Harry, for the first time I understand something from your postThanks.
I've found another paper http://www.if5.com/papers/Constructing_A_Managed_Portfolio_Of_High_Frequency.pdf
Analysis uses GARCH and Wavelet Encoding A Priori Orthogonal Network (WEAPON - is this a little marketing trick?
These methods are beyond my understanding. With 4 ticks per round-trip results are good, but for a very short period - only 6 months. How about >3 years...