After umpteen years playing with financial stuff and ...
>And being amused, right?
Yes ... it's all about amusement (for me).
I know with certainty that nobuddy can predict (with any consistent accuracy) stock prices .. say a month into the future.
Aaah ... but what influences your prediction, eh?
>Given what information?
Okay, suppose I gave you the weekly prices for the past 8 months.
I even plotted them for you.
Do you think you could guess the price 4 weeks into the future?
>Sure! I'd give it a try. You got a spreadsheet to play with?
Yes, it's here (just click on the picture to download the Excel spreadsheet):
Here's how it goes:
- You click the button called Another? and five years worth of weekly stock prices are downloaded and a random selection of 32 weeks is plotted.
- You move the slider to Guess the price 4 weeks after the last plotted price. (Your guess is shown as a green dot.)
- When you think you may have it, you click the button called Show me !!.
- The next 4 weeks of the stock prices are plotted and you sit back and pat yourself on the back.
>Or NOT !
The stock that's downloaded is, maybe, one of the DOW stocks or maybe something on the TSX ... and you're told what it is when you ask: Show me !!.
You'll get something like: or maybe this:
>What good is the spreadsheet?
Did I mention the thing about amusement ... with financial stuff?
>So, are you any good at predicting? I assume you use some exotic predictive algorithm, right?
Indubitably!
Can't you tell?
I'm lousy ... but, then, so is everybuddy else !
| |
>But predicting the future ain't easy ... is it?
It's impossible, but the idea is to see what properties of the historical behaviour influence your guess.
Do you say: "Aha! The price has been 10% higher in the past, so I'll guess something like that historical price."
>I'd say: "I see it's goin' up by $1 a week, to I'll guess ..."
Good luck with the speadsheet.
|