HardRock's  Instablog

This post has 8 comments:

Register or Login to rate comments »

KtgGroup Comments (107)
 
Just as an aside. An R-squared value of 0.39 denotes that there is an extremely poor fit between the result from the regression study and the data. Silver may do what you say, I sure don't know, but the statistical analysis does not support your conclusion.
22 Jul 2010, 01:53 PM Reply Like

HardRock Comments (47)
 
Hi KTGGroup,

I was beginning to thinks that nobody was reading this. It is just too far fetched; and maybe people think I am a crank. Thanks to you, there is intelligent life, that reads instablogs sometimes.

I understand the R2, but that is the best fit, from relationships supplied by Excel, at least to my eye and the R2.

I suppose, as in life, there may be no mathematical relationship that explains everything; maybe only parts of it.......such as 39% part. Or maybe that could be used as a percentage odds; 39 % chance of getting the big tail whipping up.

With my investing, I surely am looking forward to a wild swing up for precious metals such as Gold and Silver.

Gold Luck to you all!
HardRock
Disclosure: Long GBG.wt.a Great Basin Gold warrants and USA Silver USSIF on the OTC.
22 Jul 2010, 03:34 PM Reply Like

KtgGroup Comments (107)
 
I am not arguing with you or being hostile, just so you know where I am coming from. Back in the day, when I did regression analysis, a result like this would have been discarded as it has no practical meaning. I would not bet a dime to say that result means there is a 39% chance of being accurate, because the portion of the curve you are interested in is not constrained by any data.

The problem as I see it is that you are trying to fit a curve to an essentially a range bound data set and extrapolate from there. The best equation to fit the actual data is likely not going to be much more than the boring y=mx+b. That kind of price forecast won't be very exciting.

Part of the problem is that your data set is too small to make the extrapolation you are looking for. What I suggest doing to make your analysis more interesting is test several scenarios for the future, based on realistic assumptions (with regard to historic volatility, inflation, demand, or some magic set of things), and then see how well the real data can correlate to your hypothetical scenarios. Pick the most reasonable and an unreasonable set of assumptions for a future price, and the answer you want will likely lie in between. Regardless, you need more data to extrapolate an exponential curve from a linear data set. Don't forget, however, that as much as you don't want it to happen, there is nothing to say the price has to go up. There may be other ways to look at this, but that is all I have got.

I make no claim to be evidence of intelligent life and object to the implication.

Thanks
22 Jul 2010, 05:55 PM Reply Like

HardRock Comments (47)
 
Thanks for comments KTGGroup,

Just curious what you might think of regression for Gold Price here:
seekingalpha.com/insta...
This comes with a R-squared of .94.

HardRock
22 Jul 2010, 11:09 PM Reply Like

KtgGroup Comments (107)
 
Well, that curve seems to fit the data better, however, I would not be comfortable using it for an investment decision. It would be great if this type of analysis worked, but I can't comment more than I have about its value as a tool for predicting future price movements. My understanding and personal experience with regression analysis was to statistically evaluate how well existing data fit a trend and/or to quantify that a trend existed as opposed to random scatter. Extrapolation past the data set may be possible, but I would prefer to have some sort of basis for that than just the equation. I admit that this is just my own understanding, so I can't say your idea is wrong.
Thanks
22 Jul 2010, 11:40 PM Reply Like

tymmm Comment (1)
 
...can HardRock repeat the curve fit using data points up to present date [18Oct 2010]? I would like to see the same analysis with the last three months prices added - from my non-math perspective, it seems to align fairly well.
19 Oct 2010, 01:40 AM Reply Like

KtgGroup Comments (107)
 
Hi,
I am also interested in seeing a follow-up of recent data plotted on your previous graphs. I have thought about our discussion recently with gold and silver being higher and the equations you derived. I hope you made some good profit.
19 Oct 2010, 09:06 AM Reply Like

HardRock Comments (47)
 
Hi KtgGroup and tymmm,

Yes, I have make a little profit with investing, thinking that Gold and Silver prices should move higher.

Mainly, I have invested in Great Basin Gold (GBG.wt &GBG) and US Silver Corp (USSIF). Yes they have both gone up!

See my latest chart prompted by you folks:
seekingalpha.com/insta...

HardRock
29 Oct 2010, 03:29 PM Reply Like