No, I was talking about something different. Not BB. Later I was demonstrating how BB in it's usage also show great volatility of far in excess of 1 S.D. and how wrcmad was WRONG by posting a chart as proof and making the comment. wrcmad doesn't know much except how to read a bit and muck around. I don't think he still has any idea of what I meant.
Also go back to wrcmad chart and see the deception and what he wrote. You should really stop buying into his nonsense. EDIT: Read this post: http://forums.silverstackers.com/message-856274.html#p856274 Then read his: http://forums.silverstackers.com/message-855135.html#p855135 You should be able to make sense of it.
Well, if you've read my links above and followed the posts. Then you have a few options, I'll try and make this as easy to understand as I can, sorry if I am not clear enough... 1) He chose to use BB to misrepresent what I said. 2) He chose to put up a chart on it By all means... but he when further... 3) The chart shows 100 day moving day average which isn't even indicative of what I was saying. Also it creates wide bands which isn't even the point of them, go learn about their usage, (I want you to learn and don't take anyone's word for it). 4) He used 2 Standard deviations, not 1. Again either knowingly or deceptively. BUT then he claims that it's only ONE standard deviation from the mean. THIS IS WRONG, go look it up, here's a link... http://stockcharts.com/school/doku.php?id=chart_school:technical_indicators:bollinger_bands So just with that... what do you think? Deceptive or ignorant? Or both? EDIT: Spelling correct.
Here's his post... Notice the 100 day and the 2 standard deviation he claims were just one? See it? It's in bold for you.
I don't want to go into the specifics of the discussions simply because I hate statistics (on average every human being has one testicle) and I'm not knowledgeable enough to draw judgements on TA debates. I just wanted to suggest to double check and then triple check before making absolute statements about his lack of knowledge on the subject and then, also, avoid generalizations. If you stick to the subject and provide sound proof of your claims, you'll be fine. Everybody is wrong sooner or later, the important thing is to aknowledge it when it happens.
What exactly did you say? You have a lot to say, but 13 posts and counting, and you still haven't hinted. :| You know so much too. Can I tempt you with another oz of silver as a sweetener? I know your methods will be innovative, if not ingenious. If you can explain your methods, rather than bash or modify mine, I'll send you an oz free-post. There you go, 4 oz's on the table. Safe as houses I reckon.
Here's the data... I just finished it. I chose to use SLV daily price close for the last 3 months to calculate it 2/7/15 - 2/10/15
Also.... Funny he counted my 13 posts. That guy is good for a laugh. Also I never said 5% move was 1 S.D. I said it was MORE as in higher, my exact words were "statistically significant". I trust anyone following who claims to know stats and can be questioned, will let me know what they think?
Hey Milo. Well done... A for effort... I sincerely give you credit for that. And I apologise for assuming you made it up. However as suspected, your methodology is flawed, and it validates the alarm bells that were ringing re %price used to calculate std dev of a price. Here is why: What you have calculated is not std dev of price movement, but std dev of gain factor. Now, you are probably thinking so what? But the difference is very pronounced, and the results are not suitable to use in the std dev calculation of price moves.. The gain factor is a multiple (multiplication factor), and thus cumulatively does not equal zero around a normal distribution, as it does for price move data. This gives rise to erroneous skew of the distribution. I'll explain this in simple terms: If a price increases from $10 to $15, it is gain of 50%, or a gain factor of 1.5. If it then falls from $15 back to $10, the is a fall of 33.3%, or a gain factor of -1.33. Now. given the same starting and finishing point, the cumulative gain factor for this move should equal 0 in a normal distribution, however 1.5 + (-1.33) = 0.17. So, you can see this skews the distribution, and gives an erroneous std dev. This skew can also be demonstrated in the data you used (SLV), whereby the closing price was 13.87 on 3/8/15, and also closed at 13.87 on 30/9/15. The cumulative gain factor between these two dates was 0.00489. Because gain factor is a multiple, and the multiple effect is dependent (but not proportional) to the starting price of each period and the price move, the skew becomes even worse depending on both the starting price of any given period, and the price move between each period (in your case calculated as a % gain factor). This non-proportional distortion of your independent variable data set carries over to you std dev calculation, and so it doesn't make any sense to use it. I have done the "maths" for your given data set of SLV prices to demonstrate the flaw, with a comparison of std dev of price moves vs sdt dev of gain factor: You can see that at the time of question, your calculation of gain factor std dev was skewd by a factor of 1.5, or one and a half times, and thus seemed smaller than it truly was. Regardless, I am impressed at your efforts, and although your methodology was flawed, you came through with a good attempt. Putting aside all the guff you have written over the last week, and all the unsubstantiated shyte you made up about me lately, I will none-the-less do the decent thing and honour the offer of a free oz. (1 out of 4 a'int bad ) Whether or not you take up the offer will be something you can debate with your conscience and self-respectability. PM me your details, and have a nice day.
Hi All, I have to repost my results after talking to my uncle who's a professor emeritus of mathematics, the method I used was fine but I made an error in not using the absolute values of the % move. Meaning I am looking for the % change plus or minus, so I had to drop the minus sign to get the absolute value. I suspected this but rushed it, my honest error. Method was fine, I am measuring exactly what I stated. Here's the correct results: We can see that the average move of SLV during this 3 month period was +/- 1.128% The Z-Score of the day in question was 3.274 S.D.(distance) from the mean. I am now satisfied with the results. EDIT: "ABSOLUTE" MAGNITUDE OF THE PERCENT OF DAILY RETURNS was the correction.
Cool. I'm out. Unfortunately, I am limited in my ability to ignore facts and rational logic in order to keep arguing. :|
wrcmad, I accept your offer of 1 ounce of silver however have a request, I'd like the rough equivalent value to be donated to the Fred Hollows Foundation. For around the small price of an ounce coin $25, it can make a blind person see. You can also write it off in tax. https://donations.hollows.org/bespo...te/receipt&v=d3c4a294503e86e7c63ebc8519ea9ed1 Don't worry I scrip and save to eventually buy another ounce.
Well done guys. Kudos to both of you. Despite the intensity of the discussion the end result was honour amongst gentlemen. I applaud you. The charity will bless you.
I'm very impressed that this nearly flame post turned into a donation to the blind and a correction. Very well played gentlemen.
After consolidating for some time, we should be about to see some decent volatility (excitement) again
