Voodoo Science-Statistical Analysis | Page 3 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

[Book Discussion] Thomas R Demark - New Science of Technical Analysis - Chapter 1 30 replies

Best tools for statistical analysis of tick data and news events? 8 replies

Statistical Analysis of Trading History 9 replies

Statistical analysis fallacy 33 replies

The New Science of Technical Analysis book 3 replies

Attachments: Voodoo Science-Statistical Analysis
Exit Attachments

Voodoo Science-Statistical Analysis

  • #41
  • Jun 11, 2016 5:53pm Jun 11, 2016 5:53pm
Quoting mzvega
{quote} your analysis is only as good as the data being analyzed..............Before you choose a tool, does anyone not care about "data validity"? Mathematically, the market is recognized as a "non-linear, non-homogeneous differential equation with non-constant coefficients", impossible to solve in closed form. For practical purposes (AMVT) we deal with variables we can observe, price, time and volume. From these come estimated value. Nowhere outside of AMVT does anyone discuss or address issues of data validity. Is your data fit for the questions...
great, this what I looking for. thanks
like read this thread
the gambler
the devils apprentice Return Today: na
  • #42
  • Jun 11, 2016 5:54pm Jun 11, 2016 5:54pm
Quoting Tardigrade
It doesn't matter whether an institutional trader-slash-quant-slash-mathematician thinks she can beat the market for years to come, in backtest. A question remians---does she knows the difference between science, and voodoo science with faulted assumptions? e.g. Normal Distribution, Expected Value Any ideas for "Anti-Statistics"? Got an eye candy for you ↓↓↓ {image}
great thread
the gambler
the devils apprentice Return Today: na
  • #43
  • Jun 11, 2016 6:00pm Jun 11, 2016 6:00pm
Quoting medici
{quote} {quote} {quote} OK, but there's more to it. {quote} That's missing an important fact. The market is not a computer game based on some randomly generated time series. So statistical analysis, parametric or not, only goes so far. Behind everything that happens in the markets are human beings, so before you get too deeply into statistics you'd better understand market micro-structure, i.e. how do the guys and few gals who move the market run their business. That'll lead you to study everything from Wyckoff Theory, to Market...
the gambler
the devils apprentice Return Today: na
  • #44
  • Edited 11:33pm Jun 11, 2016 9:58pm | Edited 11:33pm
Quoting mzvega
It would be irrational for a speculator who uses a form of TA to subscribe to any form of EMH. EMH has been proven to be invalid... Rather than discuss the facts that support the invalidity of EMH, let's just "assume" that its valid and spend all day discussing "hypothetical assumptions" ignoring the invalidity of EMH despite the facts. Price in a liquid speculative market reflects all currently available information, which means, is that the current price can never be "over or under valued". That would make using a price based strategy for selecting...
There you go again, making a phantom argument. Congrats on using your own words this time, at least that is an improvement.

I cannot address anything you have said because you have not addressed any of my points. Instead, you have gone into your own extended discourse about EMH and TA when I have already established that the boundaries of the discussion should focus in on particular concepts that may or may not be committed to the major assumptions of those broader ideas.

Sorry to say, I stopped reading when I realized that you ignored or failed to comprehend my argument and chose to effectively paraphrase a textbook. I already said that my views do not commit me to accepting the entirety of the EMH or even the conventional form of TA for that matter. Address me directly or not at all.

Im a patient person, so here are some helpful pointers for internet discussions if you wish to try again:
- Confirm your acceptance or rejection of the assumptions that underlie the discussion by saying something like: "I accept <these assumptions> and will argue on these grounds." Arguing without accepting the terms of the argument is a Non-permitted Practice™.
- Before pasting blocks of text be sure they are relevant to the discussion at hand, otherwise, you would give the appearance of having little understanding of said text (a Non-permitted Practice™).
- Address points directly. E.g. "Statement A is incompatible with Statement B because <reason>". Supporting literature can be linked to, you dont have to paste it all (another Non-permitted Practice™).
- Don't get carried away (getting carried away is a Non-permitted Practice™). Make your argument concise and don't turn it into a full length dissertation. This makes what you are saying more digestible and look less like an attempt to be evasive by bombarding the other with gibberish (you guessed it, Non-permitted Practice™, dont you dare practice this).
Quoting mzvega
This conversation is definitely not for me.
Even though I did not manage to read through your entire post, I just had to read the last bit because it always contains the most passive aggression.

No hard feelings mate, next time at least make an effort to understand the text you dump.
  • #45
  • Edited Jun 13, 2016 2:09am Jun 12, 2016 9:57pm | Edited Jun 13, 2016 2:09am
Quoting FXEZ
However if you subscribe to Taleb's brand of black swans it is important to remember that the deficiency of all these statistical measures is that the large unforeseen event, aka the black swan isn't in the data.

I agree. The unseen data points are the real black swan that cannot be summarized by historical data.

Taleb in "The Black Swan" had pointed out that "we focus on pre-selected segments of the seen and generalize from it to the unseen: the error of confirmation". If we take the lesson of the black swan a step further, the syndrome of projecting-too-much-confidence-into-the-future is not only constrained to the financial market but applicable to all facets of life; The demise of Motorola, Nokia, and BlackBerry were few of the successful-failure business phenomenon that were too successful and too accustomed to the past market conditions yet lacked the ability to extrapolate into more dynamic environment.

As counter-intuitive as it may seemed, it was precisely their success, a.k.a. curve-fitting into past market data hindered their ability to be more successful in other conditions, a.k.a. produced robust results that could withstood the ups and downs in dynamic market environment.

Hence there are only few robust approaches left that can not only past the statistical test but philosophical one as well. The only one that is known to me is to buy OTM Put. Have you found any other methods?

A picture for you to consider whether true robust strategies really exist
Attached Image (click to enlarge)
Click to Enlarge Name: Dow.png Size: 486 KB
No Highs? No Lows? Must be Bose!
  • #46
  • Last Post: Jun 12, 2016 11:10pm Jun 12, 2016 11:10pm
Attached Image (click to enlarge)
Click to Enlarge Name: COV 1.jpg Size: 306 KB


Attached Image (click to enlarge)
Click to Enlarge Name: COV 2.jpg Size: 324 KB
No Highs? No Lows? Must be Bose!
0 traders viewing now