Yes, I think that academics who actually find something end up, more often than not, keeping the info to themselves and trading based on it, so that the public information is essentially what they've decided won't work. That is the paradox. When I took Finance in b-school, the professors taught us the EMH, but they themselves often traded on the side. If a PhD in Finance looked at my model, it would probably drive him to despair. I don't use anything more complicated than the basic mathematical functions of addition, subtraction, multiplication and division. And the key concept behind the system I learned from a guy who never even went to college. Yes, by "P&L" I basically meant the broader category of performance metrics in general. If it performs, you have to run with it even if, in theory, it "shouldn't" work.
this is a cool thread.. i've gotton alot out of reading (taleb) fooled by randomness and as well by the misbehavior of markets by Mandelbrot... i also like Soros's idea of fallibility.. it seems that everyone starts trading makes a few dollars and then thinks themselves deserving of a nobel prize in mathematics! i as well personallty think that the people that don't make it in trading go to selling systems, and other such propoganda! mentoring etc.. i believe that mostly its a sum zero game and we are all trying to game the largest number of people out there.. so we can figure out what the largest misperception in the market is and exploit it! weither it be over price volalitity or undervalued fundamentals in good companies! Randomness isn't random if there aren't patterns.. We will make a story out of anything its partly the only reason we can remember so much.. Read "The Bed of Procustes" or blink, or outliers, or the black swan... no one can ever really know whats going to happen.. and there is no absolute way to make money in the market, and efficient market theory is a general line of crap! i personally believe that we all will believe whatever makes the most sense related to what we wanna see.. what does Efficient mean anyway surely nothing related to the theory! I've never seen a rational state out of a person that gets cut in half nor one that doubles his month..
If you actually took a class in chaos theory you would realize there are patterns in random walks because of the way most computers generate random numbers. There are structures and rules to most things we think of as "chaotic".
thats exactly what Mandlebrot says! have you read any of his books? fractals and scaling in finance? or the Misbehavior of markets?
Stuborness is not a good quality. I just told you that Einstein had no conneciton with academics when he developed relativity. You say it does not matter. You cannot change the past to fit your misconceptions in it. This is crap. The guy who invented the transistor had no idea of quantum mechanics. This was common sense what he used obtained from experimentation. There is not a single industrial implementation of quantum mechanics contrary to what you hear to justify the grant inflow. Everything that is claimed to be related to QM could exist without that theory. No, you are dissonant. You have mixed science and academics. I attacked academics publishing papers trying to scare people that markets are random. You generalized this to everything. You think you can explain integrated circuits with some QM but what about making them? There is no science to guide you how to make ICs. It's all empirical knowledge. It is obvious that you have little connections with technology and science. In what science book can you find a detailed description of a clean room operation? Most Ph.Ds get their diploma and have never been to a clean room. 99.99% of Ph.D theses published is crap. This not only hurts true science but also taxpayer pockets. We pay for people to have a good University life and publish crap. This has to stop. We need more factory workers and more technicians. Too many Universities means the end is near. Everyone wants to become a scientist because of the wrong ideas you have.
Not all academics are money oriented , some of them are after knowledge .. Even Simons admits that / some of his associates refused to work with him any more as they found making money in the markets extremely boring /..So there is a chance that they publish a paper or two which might be worth reading This thread is taking a prospectively dangerous turn .. The mere fact that intradaybill seems to be put off by academics and the number of universities doesn't change a thing.. If there is science behind your trading / that is it is as objective as a human can possibly get coming up with his edge / your chances of survival are the greatest and Mr. Simons is the best example in the history of trading..
Read any of the Market Wizards books and you will discover that most, not all, are highly intelligent but not highly educated.
Mathematics took a bad turn when they were doing Math for the sake of Math.. As if it wasn't really that interesting if it actually had an application.. I think Academia is a totally different world then the one a practitioner operators in. Like a few smart guys that come up with the most used derivatives formula just to participate in a record blow out! LTCM or the fact that your first touch with Statistics is to remove outliers in a distribution.. The most important and significant things in life are removed from the data just to make things smooth and Coherent! Going to College does more harm them good! I've took a few classes recently and watched the students constantly try to figure out what to regurgitate back to the teacher to get a good grade.. Meanwhile right when they walk out the door they could no practically apply anything that just learned! All they do is game the education system.. and that has nothing to do with the strength of street knowledge! haha
Given any time series (could be stock/indices/fx price, sales volume, etc), what are the existing mathematical tests to quantify its level of randomness in a fixed time frame? http://www.wessa.net/tsa.wasp gives graphical representation of cyclical, trend, seasonal components but no quantified value of randomness...i.e. "To what confidence level is it forecastable"? If a time series is oscillatory, range bound(trendless) and follows closely to normal distribution, are there methods besides NeuralNetworks, singular spectrum-recurrence analysis to forecast a short term prediction interval(no need exact value)?
I used to love this thread ! Full of science and math, a couple of elephants seem to have been introduced. So let me bring forth a couple of ideas. Cause I do not even believe in the word random. Lets look at a coin toss again. 50% of the time heads and 50% of the time tails. Why ? Cause apparently randomness follows a pattern 50-50 is a pattern. There is no reason randomness needs revert to a percentage. There is no reason for numbers to have to fit in to your bell curve. heads tails tails tails tails heads tails tails tails infinity followed by heads There ! I broke the laws of randomness! I have created randomness 2 ! If markets are random, and so is a coin toss, or betting on red and green, or maybe to a lesser degree sports, than why are we trading ? Can any of you consistently make money off betting on a coin toss? But the markets ? yes ? Well if they are both random, either they both have a way to make money off of, or they both are a tax on the poor. To say that both are random and you can only make money on one is a lie! If I remember, this was the tread where we were actually asking if a coin has memory and remembers its last result. We can't look so deep in to this without bringing in science. Its an insult to logical people to avoid it. We have peeled away the skin of the future, we have been in a place that the future can be seen without its results written. We have learnt that the results are dependant on what society as a whole agrees it to be. The world of quantum physics ! The single slit theory shows us all possible outcomes. The double slit theory shows us what happens when our physics is applied to possibilities. But in the end we only see what we want to see. If all possibilities exist, and our observation results in a conclusion (that we can see). Then what does that say about the markets? Is the market not a range of possibilities that society as a whole creates, Is the market not the most honest survey of human thought ? Randomness does not exits, we only see it when choose too. what we know exists is all possibilities. They are always there. If you think other choices disappear just because we only see one possibility, then you must agree that fridges and elephants can disappear as easily as a cloud. As for our coin toss. Well I am working on a device that flips a coin once in the air less then half an inch of the ground. If I can get it to fall on heads all the time (or more then 50%) then I have broken your laws of randomness and you can buy my device with all your trading winnings.--Or is that once we can calculate the parameter it is no longer random ? Doesn't that just mean that things that are too hard to calculate are called random but in reality they represent our laziness? Is it possible that by making an indicator that is beyond the means of our capability our markets will be predictable ? Sorry, kinda confusing kinda controversial. Had to resurrect in style Long time no see ET !