How do you know that I have an accent did you put a microphone on my machine I could speak very bad english but with a very good accent: ok you guessed it my accent is also bad, you guessed it because the two is generally correlated statistically For the rest I intend to write a 5 page article on statistical use and abuse in Stock Market like I did write in the past a 5 page article on Chaos Theory and Stock Market (but it is a french article so you won't be able to read it). For short there is two parts in statistical field: the descriptive part and the inferential part. The descriptive part is just sumarising datas it doesn't prove anything. Many of what I saw is just descriptive. And without any inferential study the authors jumped straight away to the conclusion so without any proof. Also for doing inferential studies the datas must be carefully prepared for example good sampling and partioning are essential and difficult to realise and it take also most part of the time (that's why I don't feel like beginning the study of moon effect ) that's why some claimed probability can be highly doubtful when there is no knowledge of such prepration. To illustrate the danger of simplistic use of statistics let take the famous correlation coefficient. When its value is 0 simplistic thinking would be to conclude that there is no dependancy between the variables. Well as an obvious counterproof just calculate such a coefficient on a circle it will be 0 but obviously there is a dependancy relationship. You could think that it is not an error that professionals in stock market make. Well how many build a portfolio by looking at near 0 correlation between assets thinking that they will be immuned against risk because they think that there is no relation between these assets ? It is just this simplistic and dangerous thinking and that's why some funds just blow out when the relation they thought was inexistant then manifests.
The truth is: Every theory could be valid for say 30% of the whole truth. There are too many available and valid theories. All these theories together may cover not more than say 60% of the whole truth. The conculsion is: There is no ever-lasting Holy Grail in financial markets.
A truth is a truth and cannot be x% true or you change the definition of the dictionary . It is rather that a model approach the truth (ultimate reality) with some degree of perfection. Einstein's law approaches it more than Newton's law but nobody will say that Newton's law is not true. It is like mesuring a table with some precision you won't never get the "true" value but will only approach it nevertheless th true value exists. What you want to refer is rather what is used in decision process: scenarios. Each scenario has a x% chance to realise or becoming true, but when it realises it is true at 100% . A scenario can be based on a model that is deterministic at 100% what is not determined at 100% is the initinial conditions for example.
I am afraid the truth is unless all of us has already defined what is THE thuth we want, we could never find out the TRUE truth (Holy Grail?).
Harry, you bring up the best questions. I recall the one with the hat and the three cards. All these idiots were posting their formulas to explain the obvious and they were getting it wrong as well!! I never had such a laugh. You have helped me to understand that the competition in trading are not really intellectual giants, energetic maybe, but breaking new ground.... Doesn't look like it. Max
>>but it is a french article so you won't be able to read it<< I bet you I can. In fact because you are sucha nice fellow I will give you odds of 2/1. >>Also for doing inferential studies the datas must be carefully prepared for example good sampling and partioning are essential and difficult to realise<< Harry, I haven't got a clue what this has got to do with closing prices and the moon. And where does 'preparation' come into the equation in this kind of situation ? And please give us an instance of 'bad sampling' in relation to closing prices. If one were to question a number of people then yes, of course one's sample needs to be representative, but closing prices are closing prices i.e. they aren't subjective and biased like human beings. You may counter that the public when placing buying and selling orders may be influenced by other dominating factors governing at that moment in time. That is true but ..................... given enough data to work with (i.e. a long enough time period) would enable one to determine what the odds are that the results are caused by chance. (which by the way is the only way we can assess its value as absolute proof can NEVER be obtained). freealways
>>but it is a french article so you won't be able to read it<< >I bet you I can. Really then here it is http://harrytrader.membres.jexiste.org/articles/p_1.htm It's rather old since I wrote it a few months before the crack of 1998 and at that time I was just beginning to work on my non-linear model so that I didn't know what I now know. Nevertheless it's still useful for people wanting to discover the field - since it is an introductury article for everybody not for experts. I'm not an expert of chaos theory I'm only a normal user of some concepts coming from that field. Of course you can try with Babelfish whaouarf whaouarf whaouarf http://babelfish.altavista.com/babe....jexiste.org/articles/p_1.htm&lp=fr_en&tt=url >Harry, I haven't got a clue what this has got to do with closing >prices and the moon. And where does 'preparation' come into >the equation in this kind of situation ? Preparation of datas and protocol of a statistical study is the real work for calculation is just mechanical part. A human is subject ... to subjectivity so as not to introduce bias he must use some sort of a seat belt like random sampling, repetition of experiences, controlled factors - although for theses later part it is difficult in stock market since it is not a laboratory - etc. As trivial example someone who is in favor of a positive moon effect will tend to chose a period that advantages this whereas someone who is against will chose a period where it doesn't work. In both cases the experimenter is cheating. Sometimes it is not obvious so that he "sincerely" cheats himself . I can't detail here since if I had all the details I would have already begun the study .
Thank you Henry. I take it that by publishing the web address where one can find your article that you have accepted my bet. The minimum bet between trading gentlemen is at least $ 100-00 so I have in fact risked $ 200-00 against your $ 100-00. There is a saying in english which says : 'He who last best. Do you know what that saying means Henry ? O.K. then Henry. let us proceed : the bet was that I would be able to READ the article, not necessarily UNDERSTAND it. As I am not blind I certainly was able to read it even if I didn't quite understood the meaning of the words. In fact virtually everyone on this site would have been able to win that bet. Cheers Henry freealways