I'm still figuring out the spreadsheet, but I think I tried what you suggested with the curve fitting software.

Originally Posted by
Shakarri
And when keeping player level constant like I was suggesting in the last post did the fitting software give the same equation?
Using the first data set I posted in this thread, if you put each column through a curve fitting tool you get the following
Code:
+----+------------------+------------------+------------------+------------------+------------------+
| L | x3 | x2 | x | fixed | error |
+----+------------------+------------------+------------------+------------------+------------------+
| 10 | 4.895768810·10-3 | 1.235917040·10-7 | 2.570765901·10-6 | 9.999905696·10-1 | 5.764666854·10-8 |
+----+------------------+------------------+------------------+------------------+------------------+
| 15 | 3.653738123·10-3 | 1.009877337·10-7 | 1.199061899·10-7 | 9.999798998·10-1 | 2.657761264·10-8 |
+----+------------------+------------------+------------------+------------------+------------------+
| 20 | 2.609889085·10-3 | 3.926274386·10-7 | 5.792725394·10-6 | 9.999817163·10-1 | 2.503959321·10-8 |
+----+------------------+------------------+------------------+------------------+------------------+
| 25 | 1.756895202·10-3 | 2.493801574·10-7 | 1.986120992·10-6 | 9.999881419·10-1 | 3.201981555·10-8 |
+----+------------------+------------------+------------------+------------------+------------------+
| 30 | 1.086550595·10-3 | 4.567713674·10-8 | 3.012965522·10-7 | 1.000003463 | 2.548545087·10-8 |
+----+------------------+------------------+------------------+------------------+------------------+
So the x3 coefficient basically dwarfs anything else, so lets just consider that for the moment.
If I fit the changing x3 coefficient against character level we get another equation:
y = 3.811971429·10-6 x2 - 3.427850571·10-4 x + 0.007940886
Putting in the different values for the tested levels I get a set of level based coefficients like this:
Code:
+----------+----------+----------+----------+----------+
| 10 | 15 | 20 | 25 | 30 |
+----------+----------+----------+----------+----------+
| 4.89E-03 | 3.66E-03 | 2.61E-03 | 1.75E-03 | 1.09E-03 |
+----------+----------+----------+----------+----------+
Error: 2.382546286e-11
Putting it back together so that I try
y =(level coefficent* ranks ^3)+ 1
I get a table like this:
Code:
+----+----------+-------------+----------+----------+----------+
| | 10 | 15 | 20 | 25 | 30 |
+----+----------+-------------+----------+----------+----------+
| 1 | 1.004894 | 1.003656804 | 1.00261 | 1.001754 | 1.001088 |
+----+----------+-------------+----------+----------+----------+
| 2 | 1.039154 | 1.02925443 | 1.02088 | 1.01403 | 1.008705 |
+----+----------+-------------+----------+----------+----------+
| 3 | 1.132144 | 1.0987337 | 1.070469 | 1.047351 | 1.029379 |
+----+----------+-------------+----------+----------+----------+
| 4 | 1.313231 | 1.234035438 | 1.167038 | 1.112239 | 1.069639 |
+----+----------+-------------+----------+----------+----------+
| 5 | 1.611779 | 1.457100464 | 1.326247 | 1.219218 | 1.136014 |
+----+----------+-------------+----------+----------+----------+
| 6 | 2.057154 | 1.789869602 | 1.563754 | 1.378808 | 1.235031 |
+----+----------+-------------+----------+----------+----------+
| 7 | 2.678722 | 2.254283674 | 1.895221 | 1.601533 | 1.373221 |
+----+----------+-------------+----------+----------+----------+
| 8 | 3.505847 | 2.872283502 | 2.336306 | 1.897916 | 1.557112 |
+----+----------+-------------+----------+----------+----------+
| 9 | 4.567896 | 3.665809908 | 2.902671 | 2.278478 | 1.793231 |
+----+----------+-------------+----------+----------+----------+
| 10 | 5.894233 | 4.656803715 | 3.609973 | 2.753742 | 2.088109 |
+----+----------+-------------+----------+----------+----------+
Which is pretty close to the first table I tried, so using a two stage polynomial equation it's fairly accurate. But it's still missing something with the spike portion. It gets off as you spike up, and the more it spikes the more it gradually gets farther and farther off (by like 0.2% more each spike). This can lead to a couple hundred extra gold at high spikes.