Fill up ESI data...

[actam]

ppl, some data I did get from right-clicking on Full ESIs in auction…
(starting level 1, the data will show the TT/skill ratio…)
This data , as you will see, it is greatly different from that one of CareBear’s…

btw, feel free to integrate/use/elaborate the data at your wish

[HTML]TT Skills
0,01 10
0,02 19
0,03 32
0,05 49
0,06 53
0,07 66
0,09 77
0,10 85
0,12 100
0,15 122
0,18 143
0,19 149
0,21 162
0,30 216
0,32 227
0,33 233
0,37 255
0,40 268
0,47 299
0,49 308
0,51 316
0,53 324
0,56 335
0,57 339
0,60 350
0,63 361
0,65 367
0,73 393
0,77 404
0,80 412
0,82 418
0,83 420
0,89 435
0,94 447
0,99 458
1,00 461
1,02 465
1,12 486
1,13 488
1,36 535
1,40 543
1,45 554
1,54 573
1,76 626
1,80 637
1,94 677
2,04 707
2,29 783
2,45 828
3,04 946
3,11 957
3,44 1003
3,90 1067
4,01 1084
4,28 1128
4,74 1218
6,98 1522
7,36 1558
8,24 1656
8,77 1728
9,22 1790
9,98 1877
9,99 1878
13,09 2092
13,89 2153
15,94 2339
16,53 2380
16,85 2400
16,89 2402
17,15 2417
20,29 2556
36,74 3285
52,18 3642
65,92 3962
70,14 4018
79,91 4171
99,64 4475
102,95 4506
114,89 4627
188,38 5319
[/HTML]

Graph:
https://www.planetcalypsoforum.com/gallery/files/1/2/2/0/3/EsiTTskillRatio_thumb.gif

[actam]

ops, forgot to had:
value of skill are implanting skills, not explanted!

So the 1,00TT–>461 skill its the one seen in auction (the fill up from 1 to 461 levels!)

[jdegre]

try the following equation (the dots are data gathered from auct, same as you did):

https://www.planetcalypsoforum.com/gallery/files/4/5/8/1/skills.jpg

you must go for a much higher order poly than 2, if you’re doing min square poly interpolation (highest that excel can go is 6, afaik).

/jdegre.

ps: “x” is skill level, “y” is ESI tt value

[Kyl]

yo, i tried to gather the relation between Skill on chip and chip TT value.

First i thought its an exponential plus some kind of trigonometric function (eg. sin).

The prob with the exponential is: it wont go throw the y=0 and x=0.

I stopped @ a polynomial 4th kind atm. But i still don’t get a usable function.

Then i have some other theories:

• I tried to gather some info about compression functions but my skill in math isn’t high enuf to understand that formulas. Its like calculating the pressure curve if u press a liquid till u cant compress it further. That would be equivalent to filling a 1250 ped chip.

• Or we have to set up a diagram with the esi tt on the y-axis and the skill level on the x-axis. Draw the line and then calculate the surface beneath the curve to get some kind of usable value to go on with…but i am too noob in that one too ..

And there is another thing: I think the ‘value’ of the skills on the chip has to be calculated an other way then the skill in the avatar…

However, ill add my values this evening here (chipping in).

Kyl

[actam]

y = 7E-05x6 - 0,0003x5 - 0,0026x4 + 0,0072x3 - 0,0148x2 + 0,6229x + 6,1467

seems different? :scratch2:

[Kyl]

@jdegre

try the following equation (the dots are data gathered from auct, same as you did):

https://www.planetcalypsoforum.com/gallery/files/4/5/8/1/skills.jpg

you must go for a much higher order poly than 2, if you’re doing min square poly interpolation (highest that excel can go is 6, afaik).

/jdegre.

ah you here too

btw i am pretty sure that a 1250 ped chip can give 8k skills. No proof yet of course.

My excel only does 5th poly and this one dont work with my data. rubbish values..

I think i have to try office 12 soon…

Kyl

[actam]

yep, one thing i did not writh:

my graph is log-log!

so it seems line a line…

[Kyl]

@actam

yep, one thing i did not writh:

my graph is log-log!

so it seems line a line…

but why

[actam]

@Kyl

but why

because log-log fits the data better, imo.
But really everyone can one the graph they want, the only thing important is hte skills

[Witte]

@actam

yep, one thing i did not writh:

my graph is log-log!

so it seems line a line…

Nice work here. Logarithmics seems to make more sense to me then 5th degree polynomials. I will soon add this to wiki. Tnx for sharing.

[actam]

@Witte

Nice work here. Logarithmics seems to make more sense to me then 5th degree polynomials. I will soon add this to wiki. Tnx for sharing.

np, but look that in the graph I had to use a polinomial on top of the log-log!
Hope now that I’ve not confused everybody

[jdegre]

@actam

y = 7E-05x6 - 0,0003x5 - 0,0026x4 + 0,0072x3 - 0,0148x2 + 0,6229x + 6,1467

seems different? :scratch2:

what do you mean? in your graph, the equation is only 2nd degree.

@actam

yep, one thing i did not writh:

my graph is log-log!

so it seems line a line…

that’s interesting…i tried in the past to draw linear vs log, because i think it is more “logical”, but i’ve not tried log-log… maybe i’ll give it a try…

/jdegre.

[actam]

@jdegre

what do you mean? in your graph, the equation is only 2nd degree.
/jdegre.

yep 2nd degree on log-log…
the x and y in the graph is x=LN(TT) and y=LOG(Skills)

[Kyl]

@actam

yep 2nd degree on log-log…
the x and y in the graph is x=LN(TT) and y=LOG(Skills)

head cracking … but will try that with my data too of course

[jdegre]

@Witte

Nice work here. Logarithmics seems to make more sense to me then 5th degree polynomials. I will soon add this to wiki. Tnx for sharing.

the polynomials don’t have to make any sense it is just a generic interpolation function, using the minimum square method. it can interpolate any function, given enough high order to the polynomial.

i agree though that making one (or both) or the axis logarithmic may make the function more “straight” :), making the aproximation simpler.

afaik, it is the same approach used by carebear in his calculator, with the only difference that he used 3-4 polys, from different skill intervals (0-2000, 2000-4000, 4000-8000, and so…)

/jdegre.

[Witte]

@jdegre

the polynomials don’t have to make any sense it is just a generic interpolation function, using the minimum square method. it can interpolate any function, given enough high order to the polynomial.

i agree though that making one (or both) or the axis logarithmic may make the function more “straight” :), making the aproximation simpler.

afaik, it is the same approach used by carebear in his calculator, with the only difference that he used 3-4 polys, from different skill intervals (0-2000, 2000-4000, 4000-8000, and so…)

/jdegre.

Yes i understand, but I even tried 10th degree polynomals and the result was still not very pleasing. And my intuituin (looking at the graph) tells me there is an expontential funtion involved somewhere.

EDIT: The problem with using higherdegree polynomal approximation with a negative coefficient for your higest order (like you have), its that eventually the graph will go down again and become negative.

When I use 10th order, this effect will be allot stronger, making the result even worse. See:

https://www.planetcalypsoforum.com/gallery/files/9/1/6/example_thumb.png[size=1][br]Click to enlarge[/br][/size]

[jdegre]

I have tried with the log-log and, honestly, I’m not sure if we gain anything vs. the linear-linear graph. Sure the graph is “straightened” when you apply the logs, so you need a smaller order poly to approximate it, but this does not mean anything special… it is not more accurate or anything. You still can approximate the log-log function with an order-2 poly or order-6 or however.

I guess the proof would be to extrapolate beyond the gathered data, and then check for correctness of the extrapolated data, for instance for very high skill levels.

/jdegre.

PS: Extrapolated data up to 16k skill level:

https://www.planetcalypsoforum.com/gallery/files/4/5/8/1/skills2.jpg

PS2: Just saw your edit witte. yes, you’re right… the higher the order of the poly, the faster it deviates from the data points. that’s why i think the best approach is to split the skill levels in chunks, as carebear did. other than that, we could involve the maths heavyweights, i.e., JimmyB

[Mathayus]

there’s something confusing me…
if i got lets say 1k of a certain skill,and if u look to a chip that doubles your skill,i find chips worth around 10,11 ped tt…i got 1k,and with this chip i would get 2k..
but if i look a skill that i only have 5,6 or 10 for example and if i look for a chip with 1k skill its only worth around4,5 ped or not even that much…?| weird

[Doer]

Good work collecting all the data actam.

I’m not sure there’s any reason to use log-log plots here: all we need is a good approximation of the curve for the sake of calculations, not to reverse-engineer the original software formula used. Since everyone will be working in normal (vs log) space a polynomial fit in normal space is simpler.

[Kyl]

new list with my data

TT Skills Avatar 0,01 10 Actam 0,02 19 Actam 0,03 32 Actam 0,05 49 Actam 0,06 53 Actam 0,07 66 Actam 0,09 77 Actam 0,1 85 Actam 0,12 100 Actam 0,15 122 Actam 0,18 143 Actam 0,19 149 Actam 0,21 162 Actam 0,3 216 Actam 0,32 227 Actam 0,33 233 Actam 0,37 255 Actam 0,39 264 Kyl 0,4 268 Actam 0,45 291 Kyl 0,46 295 Kyl 0,47 299 Actam 0,49 308 Actam 0,51 316 Actam 0,53 324 Actam 0,56 335 Actam 0,57 339 Actam 0,6 350 Actam 0,63 361 Actam 0,65 367 Actam 0,73 393 Actam 0,77 404 Actam 0,8 412 Actam 0,82 418 Actam 0,83 420 Actam 0,89 435 Actam 0,94 447 Actam 0,95 450 Kyl 0,99 458 Actam 1 461 Actam 1,02 465 Actam 1,12 486 Actam 1,13 488 Actam 1,34 531 Kyl 1,36 535 Actam 1,4 543 Actam 1,45 554 Actam 1,54 573 Actam 1,76 626 Actam 1,8 637 Actam 1,94 677 Actam 1,96 683 Kyl 2,04 707 Actam 2,29 783 Actam 2,44 826 Kyl 2,45 828 Actam 2,68 882 Kyl 2,74 894 Kyl 3,01 941 Kyl 3,03 945 Kyl 3,04 946 Actam 3,11 957 Actam 3,44 1003 Actam 3,59 1024 Kyl 3,9 1067 Actam 4,01 1084 Actam 4,28 1128 Actam 4,43 1156 Kyl 4,74 1218 Actam 5,06 1284 Kyl 5,78 1401 Kyl 5,84 1408 Kyl 6,03 1431 Kyl 6,98 1522 Actam 7,36 1558 Actam 7,95 1620 Kyl 8,24 1656 Actam 8,43 1681 Kyl 8,77 1728 Actam 8,92 1749 Kyl 9,07 1770 Kyl 9,22 1790 Actam 9,26 1795 Kyl 9,68 1846 Kyl 9,76 1854 Kyl 9,98 1877 Actam 9,99 1878 Actam 10,15 1892 Kyl 10,6 1929 Kyl 10,88 1949 Kyl 10,89 1950 Kyl 11,19 1970 Kyl 11,22 1972 Kyl 11,33 1979 Kyl 11,75 2005 Kyl 11,91 2015 Kyl 11,93 2016 Kyl 13,02 2087 Kyl 13,09 2092 Actam 13,31 2109 Kyl 13,89 2153 Actam 14,95 2253 Kyl 15,94 2339 Actam 16,47 2376 Kyl 16,53 2380 Actam 16,85 2400 Actam 16,89 2402 Actam 16,98 2406 Kyl 17,15 2417 Actam 20,29 2556 Actam 21,66 2622 Kyl 23,12 2709 Kyl 25,33 2841 Kyl 27,97 2947 Kyl 36,74 3285 Actam 41,65 3416 Kyl 46,05 3509 Kyl 47,29 3533 Kyl 49,25 3573 Kyl 52,18 3642 Actam 55,03 3723 Kyl 62,57 3909 Kyl 65,59 3957 Kyl 65,92 3962 Actam 70,14 4018 Actam 74,7 4082 Kyl 79,66 4167 Kyl 79,91 4171 Actam 88,94 4346 Kyl 99,64 4475 Actam 101,07 4488 Kyl 102,87 4505 Kyl 102,95 4506 Actam 103,14 4508 Kyl 114,89 4627 Actam 119,66 4689 Kyl 157,91 5049 Kyl 188,38 5319 Actam 190,91 5341 Kyl 201,37 5414 Kyl 226,18 5535 Kyl 261,17 5750 Kyl