Snippet A nice math function...

[emjlr3]

the webpage version

 

[AceHart]

What did you plug into that poor thing?
And what were you expecting instead?
What's the problem?

 

[Renendaru]

I had problems with the Webpage version too, when I pressed the button to 'do the work' nothing happened at all.

 

[AceHart]

You need JavaScript turned on.

 

[Builder Bob]

Wow! That formula finder script works really well. Nice job!

Most useful when giving 3 values. With 4/5 it could make some crazy long formulas.

About reports about the script not working, I can say that it works perfectly on firefox.

 

[AceHart]

> With 4/5 it could make some crazy long formulas

That is to be expected.
And also shows that your values have no "simple" system to them

The usual GIGO Law applies (Garbage in, garbage out).

 

[emjlr3]

must be my issue here at work

 

[monoVertex]

(Garbage in, garbage out)

Ha ha .

It's funny, I inserted:

3
-5
0
0
1234

And the expected thing with the formula is:

Level 1: 3
Level 2: -58.125
Level 3: -1.5
Level 4: 154.875
Level 5: 1650

Levels 4 and 5 are way off, lol.

But anyway, good job on this, kind of late-comer to have any use for me.

 

[Siretu]

    local real a = 0.5 * (y3 - 2 * y2 + y1)
    local real b = y2 - y1 - 3 * a
    local real c = y1 - a - b

Can you explain those 3 lines? I know it works but I want to know WHY!

I get the a * x * x + b * x + c line, I almost get the a = line.

 

[AceHart]

Problem:
Determine a, b and c from "a * x² + b * x + c",
where the result gives y1 when x = 1, y2 with x = 2 and y3 with x = 3.
(I.e. the expected result on levels 1 to 3)


(1) a * 1 + b * 1 + c = y1
(2) a * 4 + b * 2 + c = y2
(3) a * 9 + b * 3 + c = y3

(4): (2) - (1): 3a + b = y2 - y1
(5): (3) - (2): 5a + b = y3 - y2

(6): (5) - (4): 2a = (y3 - y2) - (y2 - y1)
(6): a = 0.5 * (y3 - 2 * y2 + y1)

(7): b from (4): b = y2 - y1 - 3 * a
(8): c from (1): c = y1 - a - b

 

[Siretu]

Thanks! I think I got it now. +rep

Edit: I have to spread... Accept my thanks though.

 

[Chocobo]

bump

 

[AceHart]

Uhm... yes?

 

[Sooda]

I don't have nothing constructive to say, but impressive work as usual.

 

[Blackrage]

For Me, The webpage version doesnt work with over three numbers.

 

[SerraAvenger]

What did you plug into that poor thing?
And what were you expecting instead?
What's the problem?

Input: 9 - 8 - 7 - 9 - 7
Output:
Level 1: 9
Level 2: 8.539
Level 3: 7.232
Level 4: 8.061
Level 5: 4
Level 6: -21.985
Level 7: -96.936
Level 8: -257.903
Level 9: -551.944
Level 10: -1036.125

 

[Chocobo]

over 3 numbers it seems to have problem.

but it's normal when it comes to a higher sequence, there can't be a perfect formula that would fit what you want.

 

[AceHart]

Not really.
It just depends on what you plug in as values.

Those from post #36 for example are not representable with a polynomial of degree 4.
As such, while you do get a formula, it's only ever a "nearly there".

 

[SerraAvenger]

Not really.
It just depends on what you plug in as values.

Those from post #36 for example are not representable with a polynomial of degree 4.
As such, while you do get a formula, it's only ever a "nearly there".

Shouldn't a polynomial of degree 5 do it?

 

[AceHart]

Maybe. But it only has 5 values to work with.
Adding any more levels to calculate with would just lead to even more complicated formulas.