[X] My Assistant

The third difference of the total EXP required is constant (2) over a small part of the range, which is why I initially thought it was of the order of n cubed, but that doesn't hold for higher values of n. Or indeed lower values. Back to the drawing board...

here's some more numbers,

<lots of numbers>

I'm pretty sure the step always increases, so there can't be two 1395s in a row, nor 1495 followed by 1489. It looks like you have two numbers missing earlier on. In fact the ones near the beginning look a bit odd anyway.

Let's see... if instead of 4 8 14 21 41 64 83 it went 4 8 14 21 30 41 53 67 83, that would all work out nicely.

I can also confirm the following numbers:
42 => 43 : 2150
45 => 46 : 2552
46 => 47 : 2698
47 => 48 : 2849

The third difference of the total EXP required is constant (2) over a small part of the range, which is why I initially thought it was of the order of n cubed, but that doesn't hold for higher values of n. Or indeed lower values. Back to the drawing board...
I'm pretty sure the step always increases, so there can't be two 1395s in a row, nor 1495 followed by 1489. It looks like you have two numbers missing earlier on. In fact the ones near the beginning look a bit odd anyway.

Let's see... if instead of 4 8 14 21 41 64 83 it went 4 8 14 21 30 41 53 67 83, that would all work out nicely.

The "first ten" numbers XMike poster are off, just like when he first posted them. I plugged them into my spreadsheet and estimated those same exact places as where there were discrepancies.

As of now, I've got a pretty decent looking polynomial curve. I'm still trying to determine whether or not its 4th, 5th, or 6th order, since I've been eyeballing the drawn trendline rather than plugging values in, but my forecasts are pretty close (within ten points, IIRC). The only thing I'm a bit hesitant about is that we've got very small coefficients in the first few terms, but that'll sort itself out once I get enough data and get assed to do a more thorough analysis than just eyeballing it. I might just throw out XMike's numbers because their accuracy is questionable. Maybe Tenboro will feel gracious, throw us another bone, and give them to us so we can scurry about trying to put a method to his madness for his amusement.

pay no attention to the scrap totals for the figuring out cost per new stat purposes (the 2nd set of costs and sums i lumped together for total costs), just these numbers for 1 to 61 (minus the values that were off the first time i posted them)
"Let's see... if instead of 4 8 14 21 41 64 83 it went 4 8 14 21 30 41 53 67 83, that would all work out nicely." (if someone decides to pump a stat from the beginning and confirm this that would probably be great (IMG:[invalid] style_emoticons/default/tongue.gif))

where my numbers get inaccurate are during the beginning for 1-34 (missing some values):
4 8 14 21 41 64 83 101 120 141 164 189 216 245 277 310 346 384 425 468 514 562 614 668 725 786 841 916 987 1061 1138 1220 1305 (there are 33 numbers here, yet it's for 1-34 so 1 number is missing)

throw us another bone

Alright. It's not a standard n-order polynomial. (IMG:[invalid] style_emoticons/default/rolleyes.gif)

I'm suspecting the formula looks something similar to this:
(baseEXP + EXPincrease * level)^(1 + exponent * level)

Na, that's the training formula, remember. (IMG:[invalid] style_emoticons/default/rolleyes.gif)

where my numbers get inaccurate are during the beginning for 1-34 (missing some values):
4 8 14 21 41 64 83 101 120 141 164 189 216 245 277 310 346 384 425 468 514 562 614 668 725 786 841 916 987 1061 1138 1220 1305 (there are 33 numbers here, yet it's for 1-34 so 1 number is missing)

No, the 1305 is for 34 => 35, so there are two missing, or rather one missing (30) and one made up (64) where there should be two (53, 67).

This is what I get for the total EXP required, along with the second and third differences.
Note: the 841 for level 29 turns out not to fit either, so I "corrected" it to 850. (Guesswork alert)

CODE

Level   Total   Step    Step'   Step''

1       4       4
2       12      8       4
3       26      14      6       2
4       47      21      7       1
5       77      30      9       2
6       118     41      11      2
7       171     53      12      1
8       238     67      14      2
9       321     83      16      2
10      422     101     18      2
11      542     120     19      1
12      683     141     21      2
13      847     164     23      2
14      1036    189     25      2
15      1252    216     27      2
16      1497    245     29      2
17      1774    277     32      3
18      2084    310     33      1
19      2430    346     36      3
20      2814    384     38      2
21      3239    425     41      3
22      3707    468     43      2
23      4221    514     46      3
24      4783    562     48      2
25      5397    614     52      4
26      6065    668     54      2
27      6790    725     57      3
28      7576    786     61      4
29      8426    850     64      3
30      9342    916     66      2
31      10329   987     71      5
32      11390   1061    74      3
33      12528   1138    77      3
34      13748   1220    82      5
35      15053   1305    85      3
36      16448   1395    90      5
37      17937   1489    94      4
38      19524   1587    98      4
39      21214   1690    103     5
40      23012   1798    108     5
41      24922   1910    112     4
42      26950   2028    118     6
43      29100   2150    122     4
44      31379   2279    129     7
45      33791   2412    133     4
46      36343   2552    140     7
47      39041   2698    146     6
48      41890   2849    151     5
49      44898   3008    159     8
50      48071   3173    165     6
51      51415   3344    171     6
52      54939   3524    180     9
53      58649   3710    186     6
54      62553   3904    194     8
55      66659   4106    202     8
56      70975   4316    210     8
57      75510   4535    219     9
58      80272   4762    227     8
59      85271   4999    237     10
60      90516   5245    246     9
61      96017   5501    256     10

Judging by the "noise" in the third difference (and that includes values I'm certain are correct), I'd say there's rounding being applied to more than one place in the formula. Which doesn't help if you're trying to work out what it is...

No, the 1305 is for 34 => 35, so there are two missing, or rather one missing (30) and one made up (64) where there should be two (53, 67).

This is what I get for the total EXP required, along with the second and third differences.
Note: the 841 for level 29 turns out not to fit either, so I "corrected" it to 850. (Guesswork alert)

Judging by the "noise" in the third difference (and that includes values I'm certain are correct), I'd say there's rounding being applied to more than one place in the formula. Which doesn't help if you're trying to work out what it is...

The value for 29 is 849, I have it recorded.

And Tenboro just said its NOT a standard nth-order poly, so we're back to square one as far as figuring out the form of this equation. =/

There's no rounding outside of normal floating point inaccuracies. Which has about 14 significant decimal digits.

Na, that's the training formula, remember. (IMG:[invalid] style_emoticons/default/rolleyes.gif)

Yea but I recall you saying how you moved beyond polynomial. (IMG:[invalid] style_emoticons/default/tongue.gif)

5767 6043 6330 6628 6938 7260 7594 7941 8302 some more numbers (62-70)

Here. I'm not sure if the equation has changed, but if it hasn't...

I've not touched the numbers in a month... go figure. I tried figuring out Origin 8, but I didn't get far. (IMG:[invalid] style_emoticons/default/rolleyes.gif) Kudo's to Sonic, Rights, and other people that I've asked for numbers. Second and third sheets are plots of the data, the firstbeing 0-100, the second being the full range of data.

 exp.zip ( 13.18k ) Number of downloads: 31
(pssssssst: it's an excel sheet in a zip file)

That's everything up to 100ish, as well as a Kruboro supplied 199->200 cost. (IMG:[invalid] style_emoticons/default/tongue.gif)

If someone figures it out, good for them, I helped. With 100+ points I hope someone can figure it out. (IMG:[invalid] style_emoticons/default/rolleyes.gif)

This post has been edited by Spectre: Aug 29 2009, 18:14

Here. I'm not sure if the equation has changed, but if it hasn't...

I've not touched the numbers in a month... go figure. I tried figuring out Origin 8, but I didn't get far. (IMG:[invalid] style_emoticons/default/rolleyes.gif) Kudo's to Sonic, Rights, and other people that I've asked for numbers. Second and third sheets are plots of the data, the firstbeing 0-100, the second being the full range of data.

 exp.zip ( 13.18k ) Number of downloads: 31
(pssssssst: it's an excel sheet in a zip file)

That's everything up to 100ish, as well as a Kruboro supplied 199->200 cost. (IMG:[invalid] style_emoticons/default/tongue.gif)

If someone figures it out, good for them, I helped. With 100+ points I hope someone can figure it out. (IMG:[invalid] style_emoticons/default/rolleyes.gif)

Oh wow, that's a lot of good data. I guess I can go ahead and start doing a real analysis rather than LOLExcel. Better go look for my MATLAB discs.

Took me a while to find my MATLAB discs, but its now installed on my old laptop. And then I had to go back and re-read some of my notes on MATLAB. Finally got around to plugging the numbers in, and its not looking pretty. There are some big gaps in values in the 100-200 range, so I can only use the continuous values without doing some mathmagic to smooth some stuff out. Otherwise, MATLAB's basic regression analysis functions shove straight lines between the points and that ain't gonna work. I'll have to find my MATLAB book and read the chapter on regression analysis before I can do any more.

I'll keep adding to this as I gain levels, hope it helps (IMG:[invalid] style_emoticons/default/smile.gif)

120 > 121 : 61855
121 > 122 : 64147
122 > 123 : 66521
123 > 124 : 68978
124 > 125 : 71521
125 > 126 : 74154
126 > 127 : 76880
127 > 128 : 79701
128 > 129 : 82621
129 > 130 : 85644
130 > 131 : 88773
131 > 132 : 92012
132 > 133 : 95365
133 > 134 : 98835
134 > 135 : 102427
135 > 136 : 106145
136 > 137 : 109994
137 > 138 : 113977
138 > 139 : 118101
139 > 140 : 122369
140 > 141 : 126786
141 > 142 : 131359
142 > 143 : 136092
143 > 144 : 140991
144 > 145 : 146062
145 > 146 : 151311
146 > 147 : 156744
147 > 148 : 162368
148 > 149 : 168189
149 > 150 : 174215
150 > 151 : 180452
151 > 152 : 186908
152 > 153 : 193591
153 > 154 : 200509
154 > 155 : 207669
155 > 156 : 215082
156 > 157 : 222755
157 > 158 : 230698
158 > 159 : 238920
159 > 160 : 247431
160 > 161 : 256242
161 > 162 : 265364
162 > 163 : 274806
163 > 164 : 284581
164 > 165 : 294701
165 > 166 : 305178
166 > 167 : 316024
167 > 168 : 327253
168 > 169 : 338878
169 > 170 : 450914
170 > 171 : 363375
171 > 172 : 376277
172 > 173 : 389635
173 > 174 : 403465
174 > 175 : 417785
175 > 176 : 432613
176 > 177 : 447966
177 > 178 : 463863
178 > 179 : 480324
179 > 180 : 497369
180 > 181 : 515020
181 > 182 : 533298
182 > 183 : 552226
183 > 184 : 571827
184 > 185 : 592126
185 > 186 : 613148
186 > 187 : 634919
187 > 188 : 657466
188 > 189 : 680818
189 > 190 : 705004
190 > 191 : 730055
191 > 192 : 756001
192 > 193 : 782875
193 > 194 : 810711
194 > 195 : 839545
195 > 196 : 869413
196 > 197 : 900352
197 > 198 : 932402
198 > 199 : 965603
199 > 200 : 999999
200 > 201 : 1035632

This post has been edited by RenRenRen: Oct 14 2009, 23:48

(IMG:[invalid] style_emoticons/default/biggrin.gif) Finally!
OK I cheated a bit and looked at the Javascript source.

Spoiler text - Highlight to read...

(level + 1)^(1.8932742^(1 + level / 400)) rounded

The arbitary value is found by assuming that to get from level 199 to level 200 requires 9999999 experience. Then just change y for x (invert the function). The value is the logarithm of said function.

For some reason, Derive gives me this value for 'x' which is different... Not sure why this is wrong.

Spoiler text - Highlight to read...

x=[(ln 9999999)^(400/599)]/[(ln 200)^(400/599)] which is 2.102107538

This post has been edited by lovehcomics: Jul 14 2011, 06:42

I moved your post, because you put it in the wrong thread. What you've found there is the formula for EXP required to raise your primary stats, NOT to raise your level. You might want to edit that to take out the reference to the "arbitary value" you were trying to find, which doesn't in fact exist...

Why did no one find this before? Because it wasn't there, most likely. The numbers were probably worked out by the server before, and 0.6.0 moved the code to the character page to facilitate the stat reallocation thingy.

Because it wasn't there

^

I'm surprised it took you four days. (IMG:[invalid] style_emoticons/default/tongue.gif)

I know. It should have been obvious (given my connection delay) that it was updating the EXP values without communicating with the server. I have no excuse. (IMG:[invalid] style_emoticons/default/sad.gif)

I tried updating the wiki but can't do it so posting here. It looks like there is a slight inaccuracy in the formula posted there:

(current_stat + 1)^(1.8932743^(1 + current_stat / 400))

Ran that through Excel and did not get the correct value. Ran this one:

(current_stat + 2)^(1.8932743^(1 + (current_stat+1) / 400))

and hit it on the nose. Either the original formula was wrong or Tenb just made it a little harder with the new patch.