General Betting
There is currently 1 person viewing this thread.
I'm trying to calculate a set of 'best fit' values for some non-linear data. When I add a trendline line to an Excel chart, I find that a 6th order polynomial is a really good fit for the outputs I have. However when I feed the x values into the equation, it does not match either the data or the trendline.
Presumably this is down to lack of precision in the way the equation is displayed, e.g + 0.7E-05x^3.
Is it possible to have Excel output the y value for each value of x on the trendline? Alternatively, are there any pieces of software out there that are able to offer solutions whilst still being omprehensible to a layman?
Presumably this is down to lack of precision in the way the equation is displayed, e.g + 0.7E-05x^3.
Is it possible to have Excel output the y value for each value of x on the trendline? Alternatively, are there any pieces of software out there that are able to offer solutions whilst still being omprehensible to a layman?
sort by:
Show
per page
Replies: 10
|
By:
Maybe try using SPSS. Think there's a free trial here: http://www.spss.com/uk/statistics/
|
|
By:
Thanks Mac, I'll check it out.
|
|
By:
Increase the number of decimal places. If you have the equation on the chart, right click it, select "format trendline label", select "number" under catagory, then alter the number of decimal places from 2 to 10 (or whatever).
|
|
By:
When I read threads like this, I really feel inadequate.
Is this all gambling related ? If so, the mind boggles. |
|
By:
And dogrelata calls himself a layman !
Demoralizing to say the least. |
|
By:
Cheers Cosmic.
This is starting to look a lot better. First try with 10 decimal places wasn't enough, but at least the output is starting to bear some resemblance to the actual data now. |
|
By:
I use curveexpert. It has a better curve finder utility and it's free!
|
|
By:
Hint: A 6th order polynomial best fit means you are fooooking way off..Try again.
|
|
By:
Thanks only223.
That's a nice peice of software - easy to use and produces exactly what I was looking for. |
|
By:
Thanks for the hint speedypro.
Heres the thing though. In answer to Frog Hairs (rhetorical?) question earlier, the query is gambling related and concerns data gathered from 1,484 races over the last 10 weeks. The data produces an output value for each race favourite, which determines whether to back or lay. Using the raw data, 775 backs produce 20.95 points profit after commission, 709 lays produce 13.92 points profit. By using a 6th order polynomial to produce the output value, 780 backs produce 36.76 points and 704 lays produce 30.76 points. Thats almost double the profit. I hold my hand up, Im no mathematician, so may well be |