Pages: [1] :: one page |
|
Author |
Thread Statistics | Show CCP posts - 0 post(s) |
Salpad
Carebears with Attitude
49
|
Posted - 2012.10.11 06:39:00 -
[1] - Quote
A few days ago, I discovered an MS Excel function that lets me grab the Xth highest (or lowest) value out of a data set. I figured that with this function, I'd be able to build a spreadsheet to brute-force analyze my way, through trial and error, to optimal omni-tank fits for a variety of shield ships. Simply try with or without DC, with 1 Invulns, or 2, 3 or zero, and with various numbers of specific hardeners.
I've spent a couple of hours on that, typing in data, getting the stacking penalties right, and so forth. With a ready-to-test product, I tried it on my current ship, and the values don't match what the in-game Fitting screen shows me. They're off by several percent, meaning it can't be rounding errors.
I'm not too good with spreadsheets, so at this point I'm tempted to give up.
Has anybody else made a spreadsheet like this, that facilitiates the comparison of several different fits on the same screen, showing the result (primarily the omni-tank) of each fit?
Don't refer me to EFT, because EFT can't do this. EFT requires me to manually swap out modules one at a time, and only shows the result for one fit at a time, meaning I have to do manual record-keeping in Notepad, and that does not faciliitate trying a lot of different permutations to, e.g., find the best 3-slot tank for a Cerberus, or the best 4-slot or 2-slot tank, and at what point the DC stops making a serious difference for my shield tank...
I don't know if Pyfa can do this. I downloaded it about a week ago, but could not at all figure out how to import my API data.
Is there a spreadsheet that already does this? Or a program that lets me manually compare a variety of different fit permutations? Or a program that does the analysis of its own, runs through all the permutations for X slots for a given hull, and outputs the optimal fit?
|
Serena Serene
Heretic University Heretic Nation
3513
|
Posted - 2012.10.11 06:47:00 -
[2] - Quote
Which values are off by several percent? If it's ehp, the in-game fitting screen doesn't show ehp for incoming damage evenly distributed over all damage types, but rather for the damage type you have the weakest resists for, calculated separately for shield and armor (edit: and structure of course) and then added up. That's what I've read a while ago, at least.
In eft you can at least precisely set the damage type distribution against which you want to see your tank values. |
Salpad
Carebears with Attitude
49
|
Posted - 2012.10.11 06:50:00 -
[3] - Quote
I don't do EHP, and I don't do in-coming damage, because I want the spreadsheet to be a tool for all kinds of omni-shield-tanking: Active, passive, buffer. All I care about are the four resists, and even more so I care about what the lowest of them is.
The values that are off by several percent are the actual resists. EM resist, kinetic resist, and so forth.
|
Serena Serene
Heretic University Heretic Nation
3513
|
Posted - 2012.10.11 07:00:00 -
[4] - Quote
I see. Then there are only four possible explanations I guess: - Your spreadsheet isn't correct - You made an error with the ship fit - There is some skill influence you didn't consider - Your fit has active hardeners and you're looking at the in-game fitting screen while docked up or with deactivated hardeners.
@your question: I don't know of any spreadsheets attempting what you did, but that doesn't mean there aren't any, of course. |
Riot Girl
Riot Club
145
|
Posted - 2012.10.11 07:03:00 -
[5] - Quote
Is your spreadsheet taking skills into account? For passive hardeners, the numbers will be affected by resistance compensation skills and there are probably other skills that affect the overall resistances such as ship command skills affecting ship bonuses. |
Salpad
Carebears with Attitude
49
|
Posted - 2012.10.11 07:10:00 -
[6] - Quote
Serena Serene wrote:I see. Then there are only four possible explanations I guess: - Your spreadsheet isn't correct - You made an error with the ship fit - There is some skill influence you didn't consider - Your fit has active hardeners and you're looking at the in-game fitting screen while docked up or with deactivated hardeners.
@your question: I don't know of any spreadsheets attempting what you did, but that doesn't mean there aren't any, of course.
I've double-checked the ship fit, and I'm only using active hardeners, so no skill influence there, and the Cerberus doesn't get a resists bonus based on any skillz, the same way e.g. the Nighthawk (and Drake) does. Also, of course I undock and activate the DC and Hardeners before I check the figures. So the fault must be in my spreadsheet.
|
Serena Serene
Heretic University Heretic Nation
3513
|
Posted - 2012.10.11 07:23:00 -
[7] - Quote
Salpad wrote:Serena Serene wrote:I see. Then there are only four possible explanations I guess: - Your spreadsheet isn't correct - You made an error with the ship fit - There is some skill influence you didn't consider - Your fit has active hardeners and you're looking at the in-game fitting screen while docked up or with deactivated hardeners.
@your question: I don't know of any spreadsheets attempting what you did, but that doesn't mean there aren't any, of course. I've double-checked the ship fit, and I'm only using active hardeners, so no skill influence there, and the Cerberus doesn't get a resists bonus based on any skillz, the same way e.g. the Nighthawk (and Drake) does. Also, of course I undock and activate the DC and Hardeners before I check the figures. So the fault must be in my spreadsheet.
At this point I don't see another possiblity either. Unless you didn't consider that your DC isn't affected by stacking penalties caused by your hardeners (afaik)... well, that'd be a fault in your spreadsheet, too. |
Salpad
Carebears with Attitude
49
|
Posted - 2012.10.11 07:27:00 -
[8] - Quote
The spreadsheet specifically exempts the native shield resists of the ship, and the resist effect from a DC (if the DC cell is set to "yes" - otherwise there's no bonus), from the stacking penalty.
I've found and fixed two errors in the spreadsheet, since my last post, but the spreadsheet results still don't match the in-game values.
It's a very error-prone thing to try to make a spreadsheet calculate, though, because of all the multiplication and division involved. Resist bonuses don't really add to resist; they act to reduceincoming damage by a fraction. So a lot of divisiion and multiplication and use of reciprocals...
So I'm still very interested to hear if anybody else has already done this kind of work, and is willing to share.
|
illirdor
The Grey Eagle Society
23
|
Posted - 2012.10.11 08:34:00 -
[9] - Quote
Salpad wrote:
Don't refer me to EFT, because EFT can't do this. EFT requires me to manually swap out modules one at a time, and only shows the result for one fit at a time
that is not true you can have more than one window open at once in eft i use eft alot to compare more fits next to each other withput having to write down results and so in a notepad.
Come to Amamake... The universal answer to everything... |
Salpad
Carebears with Attitude
49
|
Posted - 2012.10.11 13:18:00 -
[10] - Quote
illirdor wrote:Salpad wrote:
Don't refer me to EFT, because EFT can't do this. EFT requires me to manually swap out modules one at a time, and only shows the result for one fit at a time
that is not true you can have more than one window open at once in eft i use eft alot to compare more fits next to each other withput having to write down results and so in a notepad.
In a spreadsheet, I can easily have 15 or 20 rows of different combinations of DC/hardeners (even on a small monitor), for a given ship hull, and I can change these configurations at a rate of several changes per second, using the keyboard (deleting and changing one-digit numerical values) whereas in EFT the rate would be several seconds per one change (because I have to use the mouse and I have to click-and-drag a lot).
|
|
Zhilia Mann
Tide Way Out Productions
699
|
Posted - 2012.10.11 19:48:00 -
[11] - Quote
If you post the sheet in Google Drive and link to it I'll try to look at it over the next day or so. Hard to troubleshoot when you don't know exactly what you're looking at though. |
Salpad
Carebears with Attitude
87
|
Posted - 2012.10.12 14:51:00 -
[12] - Quote
Zhilia Mann wrote:If you post the sheet in Google Drive and link to it I'll try to look at it over the next day or so. Hard to troubleshoot when you don't know exactly what you're looking at though.
Do you prefer MS Excel 2003 or a conversion to Libre Office?
|
Zhilia Mann
Tide Way Out Productions
705
|
Posted - 2012.10.12 15:04:00 -
[13] - Quote
Salpad wrote:Zhilia Mann wrote:If you post the sheet in Google Drive and link to it I'll try to look at it over the next day or so. Hard to troubleshoot when you don't know exactly what you're looking at though. Do you prefer MS Excel 2003 or a conversion to Libre Office?
Libre if you can manage it.
|
Salpad
Carebears with Attitude
87
|
Posted - 2012.10.12 15:14:00 -
[14] - Quote
Zhilia Mann wrote:Libre if you can manage it.
This is a DropBox link, not Google Drive, but as far as I can determine, it should work. Libre Office, and what advisory text there is is in English. Please let me know if there are any problems.
|
Zhilia Mann
Tide Way Out Productions
705
|
Posted - 2012.10.15 20:34:00 -
[15] - Quote
Got it. Wasn't that tough once I started grinding it hard. But of course I started with the formulas, which are pretty much fine. The only change I'd recommend on that would be to use the actual stacking formula like so:
=0.5^(((AG51-1)/2.22292081)^2) where AG51 is a reference cell for being the nth number in the stack (and in your sheet it's actually number 2).
The problem, as it turns out, is that you had the Cerberus' native therm and kin resists backwards. Try flipping those around; suddenly everything rounds properly. So now your Cerberus with DC, 2 EM, and 1 exp reads 79.4, 82.5, 73.8, 80.3 -- which is where it should be. |
Salpad
Carebears with Attitude
91
|
Posted - 2012.10.15 20:36:00 -
[16] - Quote
So it was just a massive brain fart not solved by me triple-checking everything? Thanks. I'll input the correct values, and get back to this thread.
|
|
|
|
Pages: [1] :: one page |
First page | Previous page | Next page | Last page |