Pages: 1 [2] :: one page |
|
Author |
Thread Statistics | Show CCP posts - 0 post(s) |
Qoi
Exert Force
91
|
Posted - 2014.08.23 16:06:00 -
[31] - Quote
Medalyn Isis wrote:Just to confirm, it is now possible for a T2 ship to use for instance, 2 T1 ships to make 1 T2 finished product. Is that correct? As using the given formula that seems to be the case.
Also, I haven't tried inventing any T2 BPCs yet, although I thought they were going to be ME 0 by default, although according to the pdf they are -2 ME. Is this correct?
Thanks very much for the good pdf also, nice info in there.
When i say ME -2% TE -4% i mean that the blueprint reduces the materials by 2% and the build time by 4%. This has nothing to do with the old negative ML/PL values. With the definition of "level" from the PDF this would be a ME level 2, TE level 2 blueprint (each ME level reduces the material requirements by 1%, each TE level reduces the build time by 2%). Blueprints can only have levels from 0 to 10, including copies produced by invention. http://eve-industry.org |
Qoi
Exert Force
91
|
Posted - 2014.08.23 18:17:00 -
[32] - Quote
I added a few things to the reprocessing section, including a sentence about rounding:
The reprocessing output is obtained by multiplying the reprocessing rate with the base material amounts and then rounding down (POS) or rounding to next integer (Station).
I did some quick testing on Singularity and that is what i found, if someone has more accurate information please tell :) http://eve-industry.org |
Creepin
Yellow Flamingo
13
|
Posted - 2014.08.30 14:24:00 -
[33] - Quote
First of all, thanks for your excellent pdf, this is pretty much the only source on Crius manufacture around.
However, I'm a bit confused by your formula for materials required and would appreciate some clarification. Here's what I'm confused about in your formula required=max(runs,ceil(round(runs*baseQuantity*materialModifier,2)): - it seem to me that ",2" part is redundant as there's no way the result of runs*baseQuantity*materialModifier could have more than 2 digits after comma anyway; - why ceil(round())? It seem to me that simple ceil() (or roundup if in excel) will yield totally same result (within the range of what's possible in eve manufacturing, obviously: it might work differently for negative values, but this is not the case here).
Or did I miss something? |
Steve Ronuken
Fuzzwork Enterprises Vote Steve Ronuken for CSM
3762
|
Posted - 2014.08.30 17:31:00 -
[34] - Quote
Creepin wrote:First of all, thanks for your excellent pdf, this is pretty much the only source on Crius manufacture around.
However, I'm a bit confused by your formula for materials required and would appreciate some clarification. Here's what I'm confused about in your formula required=max(runs,ceil(round(runs*baseQuantity*materialModifier,2)): - it seem to me that ",2" part is redundant as there's no way the result of runs*baseQuantity*materialModifier could have more than 2 digits after comma anyway; - why ceil(round(X,2))? It seem to me that simple ceil(X) (or roundup if in excel) will yield totally same result (within the range of what's possible in eve manufacturing, obviously: it might work differently for negative values, but this is not the case here).
Or did I miss something?
What you missed was floating point math. Because when you have fractions involved, there's a small tendency to have results which end in .000000000000000001. Which is enough to make the ceil() take it to the next integer.
An example:
http://www.wolframalpha.com/input/?i=ceiling%28104*0.9*60%29
http://www.wolframalpha.com/input/?i=ceiling%28104*60*0.9%29
(bah. The links don't work right. if you copy and paste, they should do)
For an explanation: http://www.youtube.com/watch?v=PZRI1IfStY0 Woo! CSM 9! http://fuzzwork.enterprises/ Twitter: @fuzzysteve on Twitter |
Creepin
Yellow Flamingo
13
|
Posted - 2014.08.30 20:21:00 -
[35] - Quote
Steve Ronuken, thanks a lot! I'm trying to understand your explanation now, fruitless so far
1. Am I right interpreting your answer as the statement that when eve engine coming up with manufacturing quotes, it uses raw binary results for an output without rounding it up for a correct decimal value beforehead, so that when binary math tells that 2*7 tritanium is 14,000000000000000000001, it rounds it up to 15? 2. Your examples are quite unsettling :) Does this binary maths voids the rule that changing order of multipliers should provide the same results? 3. Now, even if 1 & 2 are true, I've checked ceil function implementation in both excel & openoffice, and there were no hints that it takes into consideration this floating point (that occasional 0.00000001 from your answer), same true for roundmax. I also tried to simulate you example of ceiling(104*0.9*60)=5617 in Excel via both ceiling(104*0.9*60) & ceiling.math(104*0.9*60) (because I didn't understand the difference), but both formulas provided me with 5616, as it should in decimal math. Should I interpret this result as a sign that while I need to take floating point into consideration, Excel doesn't give me such opportunity, or that I used wrong formula? If latter, which is the correct one? |
Steve Ronuken
Fuzzwork Enterprises Vote Steve Ronuken for CSM
3763
|
Posted - 2014.08.30 20:33:00 -
[36] - Quote
You won't find it happening, when you're dealing with integers. Only when you're dealing with floating point numbers. It's because representing them is more difficult than it seems.
A metaphor:
you know that one third, multiplied by 3, is 1.
But representing one third is difficult. If you try to write it, you'll write 0.333333333333 and on and on. But if you stop writing it, then multiply it by 3, you'll get 0.99999999999999999999
If you try to get round that with 0.333333333333334, you'll get a number that's just above 1.
That's pretty much the problem that the computer has.
You see it far more often with programming languages, because they don't try and make things easy on you. Applications like Excel add in logic to 'deal' with it. Probably by just checking for that 0.0000000000000000000001 Woo! CSM 9! http://fuzzwork.enterprises/ Twitter: @fuzzysteve on Twitter |
Creepin
Yellow Flamingo
13
|
Posted - 2014.08.30 20:44:00 -
[37] - Quote
Steve Ronuken wrote:You see it far more often with programming languages, because they don't try and make things easy on you. Applications like Excel add in logic to 'deal' with it. Probably by just checking for that 0.0000000000000000000001 Yep, I watched the video you kindly provided, so I figured this much. But then, if Excel applies additional logic to revert binary results to what we expect in decimal maths, is there a way to correctly simulate in Excel Eve engine calculations that use raw binary maths as results? |
Steve Ronuken
Fuzzwork Enterprises Vote Steve Ronuken for CSM
3763
|
Posted - 2014.08.30 20:50:00 -
[38] - Quote
Creepin wrote:Steve Ronuken wrote:You see it far more often with programming languages, because they don't try and make things easy on you. Applications like Excel add in logic to 'deal' with it. Probably by just checking for that 0.0000000000000000000001 Yep, I watched the video you kindly provided, so I figured this much. But then, if Excel applies additional logic to revert binary results to what we expect in decimal maths, is there a way to correctly simulate in Excel Eve engine calculations that use raw binary maths as results?
That's what the round(,2) is there for. It just eliminates the 0.00000001 Woo! CSM 9! http://fuzzwork.enterprises/ Twitter: @fuzzysteve on Twitter |
Creepin
Yellow Flamingo
13
|
Posted - 2014.08.30 21:04:00 -
[39] - Quote
Steve Ronuken wrote:That's what the round(,2) is there for. It just eliminates the 0.00000001 Gah! So, all this time you were explaning to me why are there Round(,2), while I was thinking you're explaining why are there Ceiling! God I feel stupid! Well thanks a lot for bearing with me :) |
Qoi
Exert Force
95
|
Posted - 2014.09.11 17:02:00 -
[40] - Quote
I added the information about the new invention system to the PDF, i will update it as more information becomes available. http://eve-industry.org |
|
Zifrian
Licentia Ex Vereor Black Core Alliance
1555
|
Posted - 2014.09.11 19:58:00 -
[41] - Quote
Thanks for this. Well done and much better than the old web versions.
One question, where did you get the success rates for reverse engineering? I've been using 20%, 30% and 40% for Wrecked, Malfunctioning, Intact for years. You list 25%, 35% and 40% in Table 2. Did this change and I missed it? Can you provide a link to your source?
Also, your success equation for T3 is slightly different than the one I've used in the past (and others seem to use too):
basechance * 1 + 0.01 * ReverseEngineeringLevel + 0.1 * (Datacore1Level + Datacore2Level)
is what you have but I've used this one for some time:
Success rate = base ate * (1 + reverse engineering skill level * 0.01) * (1 + (datacore 1 skill level + datacore 2 skill level) * 0.1)
They are similar but the rates are not equal.
Thanks GÇ£Any fool can criticize, condemn, and complain - and most fools do. GÇ¥ - Dale Carnegie
Maximze your Industry Potential! - Download EVE Isk per Hour! |
Zifrian
Licentia Ex Vereor Northern Associates.
1580
|
Posted - 2014.09.11 19:58:07 -
[42] - Quote
Thanks for this. Well done and much better than the old web versions.
Edit: [nevermind all fine]
GÇ£Any fool can criticize, condemn, and complain - and most fools do. GÇ¥ - Dale Carnegie
Maximze your Industry Potential! - Download EVE Isk per Hour!
|
BooomBox
Federal Navy Academy Gallente Federation
0
|
Posted - 2014.09.11 21:15:00 -
[43] - Quote
The new RE formula is from dev blog. It confirms the changed in crius chances of success, shown in industry interface for intact relics (decrease from 84% to 82% with perfect skills). For malfunctioned relics the chance shown in industry interface is 71,7% with perfect skills (was 63% pre crius), which uses base chance of 35% (according to formula). Probably the same calculation is valid for wrecked relics, but I have not made calculations myself. I suppose Qoi used this info from industry interface to determine new base chance for malfuncioned and wrecked relics. |
BooomBox
Federal Navy Academy Gallente Federation
0
|
Posted - 2014.09.11 21:15:29 -
[44] - Quote
The new RE formula is from dev blog. It confirms the changed in crius chances of success, shown in industry interface for intact relics (decrease from 84% to 82% with perfect skills). For malfunctioned relics the chance shown in industry interface is 71,7% with perfect skills (was 63% pre crius), which uses base chance of 35% (according to formula). Probably the same calculation is valid for wrecked relics, but I have not made calculations myself. I suppose Qoi used this info from industry interface to determine new base chance for malfuncioned and wrecked relics. |
Zifrian
Licentia Ex Vereor Black Core Alliance
1555
|
Posted - 2014.09.11 22:42:00 -
[45] - Quote
BooomBox wrote:The new RE formula is from dev blog. It confirms the changed in crius chances of success, shown in industry interface for intact relics (decrease from 84% to 82% with perfect skills). For malfunctioned relics the chance shown in industry interface is 71,7% with perfect skills (was 63% pre crius), which uses base chance of 35% (according to formula). Probably the same calculation is valid for wrecked relics, but I have not made calculations myself. I suppose Qoi used this info from industry interface to determine new base chance for malfuncioned and wrecked relics.
OK, the base chance is listed in the info of the relics and 25%, 35% and 40% are correct (Helps if one plays the game more I guess).
As far as the formula, today's dev blog has it here: http://cdn1.eveonline.com/www/newssystem/media/66476/1/REformula.png
I just never saw it before now. Qoi's guide is updated with the current. So no worries!
Thanks GÇ£Any fool can criticize, condemn, and complain - and most fools do. GÇ¥ - Dale Carnegie
Maximze your Industry Potential! - Download EVE Isk per Hour! |
Zifrian
Licentia Ex Vereor Northern Associates.
1580
|
Posted - 2014.09.11 22:42:47 -
[46] - Quote
BooomBox wrote:The new RE formula is from dev blog. It confirms the changed in crius chances of success, shown in industry interface for intact relics (decrease from 84% to 82% with perfect skills). For malfunctioned relics the chance shown in industry interface is 71,7% with perfect skills (was 63% pre crius), which uses base chance of 35% (according to formula). Probably the same calculation is valid for wrecked relics, but I have not made calculations myself. I suppose Qoi used this info from industry interface to determine new base chance for malfuncioned and wrecked relics.
OK, the base chance is listed in the info of the relics and 25%, 35% and 40% are correct (Helps if one plays the game more I guess).
As far as the formula, today's dev blog has it here: http://cdn1.eveonline.com/www/newssystem/media/66476/1/REformula.png
I just never saw it before now. Qoi's guide is updated with the current. So no worries!
Thanks
GÇ£Any fool can criticize, condemn, and complain - and most fools do. GÇ¥ - Dale Carnegie
Maximze your Industry Potential! - Download EVE Isk per Hour!
|
Lateralus
War Supplies Inc
3
|
Posted - 2014.09.26 22:53:00 -
[47] - Quote
Ok, I've done a monumental thing - to me anyhow. I've created my own spreadsheet, which pulls information on crazy amounts of data from Fuzzwork's site, with a few things manually put in from me (such as BPO ME/TE levels) but it does almost everything else for me. It's great I love it.
However, the Job Fees are making me insane.
On page 3 of your PDF you have this formula: jobFee = baseJobCost * systemCostIndex * 0.02 * runs I can do this using your website api by pulling the system cost index...
However... on page 4, section 7, there's a section about how to calculate the baseJobCost... I have no clue now to do this: baseJobCost = SUM ( baseQuantity * adjustedPrice ) for all materials
So, I have a section in my spreadsheet that calculates the T1 manfuacturing cost, and the T2 manufacturing cost, which pulls information using IMPORTXML in a google doc for the prices. I figured I would use those prices to approximate the 'adjustedPrice', but it seems to be way off.
Where do I get the 'adjustedPrice' data? Is that a whole set of different prices for all the materials required to make something that is hidden away somewhere?
|
Lateralus
War Supplies Inc
3
|
Posted - 2014.09.26 22:53:19 -
[48] - Quote
Ok, I've done a monumental thing - to me anyhow. I've created my own spreadsheet, which pulls information on crazy amounts of data from Fuzzwork's site, with a few things manually put in from me (such as BPO ME/TE levels) but it does almost everything else for me. It's great I love it.
However, the Job Fees are making me insane.
On page 3 of your PDF you have this formula: jobFee = baseJobCost * systemCostIndex * 0.02 * runs I can do this using your website api by pulling the system cost index...
However... on page 4, section 7, there's a section about how to calculate the baseJobCost... I have no clue now to do this: baseJobCost = SUM ( baseQuantity * adjustedPrice ) for all materials
So, I have a section in my spreadsheet that calculates the T1 manfuacturing cost, and the T2 manufacturing cost, which pulls information using IMPORTXML in a google doc for the prices. I figured I would use those prices to approximate the 'adjustedPrice', but it seems to be way off.
Where do I get the 'adjustedPrice' data? Is that a whole set of different prices for all the materials required to make something that is hidden away somewhere?
|
Qoi
Exert Force
103
|
Posted - 2014.11.04 19:43:28 -
[49] - Quote
I updated the formulas for Phoebe.
Lateralus wrote: Where do I get the 'adjustedPrice' data? Is that a whole set of different prices for all the materials required to make something that is hidden away somewhere?
The canonical source is http://public-crest.eveonline.com/market/types/ and if you are lazy you can get the baseJobCost from http://api.eve-industry.org/
http://eve-industry.org
|
Cilegon
Volicorp
0
|
Posted - 2014.11.05 14:29:12 -
[50] - Quote
Hi Qoi,
I don't see the production build time forumla taking into the account the reduction based on the science skills... Unless i'm isunderstanding something?
You mention TE modifier, but i assume thats just the TE mod from the bp itself with no mention of the 1% from the science skills
from here: http://community.eveonline.com/news/dev-blogs/invention-updates/
To compensate for this change, all those skills will now give a 1% Time Efficiency bonus for the Tech II manufacturing job they are required for, which is still going to give an incentive for players to train those up, or give an edge for players that already trained them. |
|
Qoi
Exert Force
103
|
Posted - 2014.11.05 17:01:51 -
[51] - Quote
Hey, you are right, i forgot to add that. I updated the PDF now.
http://eve-industry.org
|
Luminocity
The Dark Revenants PLEASE NOT VIOLENCE OUR BOATS
20
|
Posted - 2014.11.06 22:50:00 -
[52] - Quote
Lateralus wrote:Where do I get the 'adjustedPrice' data? Is that a whole set of different prices for all the materials required to make something that is hidden away somewhere? From here: http://public-crest.eveonline.com/market/prices/ These are magical values that CCP uses to mitigate any attempts of players to impact the installation cost by manipulating the market
Qoi great work on compiling this. Is a good single-point-resource for detailed overview of inner workings for post-crius industry |
Qoi
Exert Force
105
|
Posted - 2014.11.07 18:18:38 -
[53] - Quote
Luminocity wrote:Lateralus wrote:Where do I get the 'adjustedPrice' data? Is that a whole set of different prices for all the materials required to make something that is hidden away somewhere? From here: http://public-crest.eveonline.com/market/prices/ These are magical values that CCP uses to mitigate any attempts of players to impact the installation cost by manipulating the market Qoi great work on compiling this. Is a good single-point-resource for detailed overview of inner workings for post-crius industry. One thing that caught my eye is regarding the "Total Job Installation Costs" section of paragraph 5. I think "teamCostModifier" should instead be "teamCostRate" to be consistent with how percentage/decimal values are expressed in the rest of the document Thanks :-) I silently updated the identifier to teamCostRate. You have a keen eye!
http://eve-industry.org
|
|
|
|
Pages: 1 [2] :: one page |
First page | Previous page | Next page | Last page |