Science & Industry

Yes, of course I meant to multiply.

And, yes you are right, the math does not fit this way, because it's like

ROUND(raw) + ROUND(extra) and not ROUND(raw + extra) :( sorry for that - I should've checked my code the first way

You sure, that you get 5 with Prod Eff 1?
I ran the numbers and got this quote:
[pre]
|============================================================|
| NAME | REQUIRED | MISSING | DMG/JOB | WASTE |
|============================================================|
| Raw Material (4x) | | | | |
|------------------------------------------------------------|
| Isogen | 2 | 3 | 100.00% | 50.0% |
| Mexallon | 3 | 4 | 100.00% | 33.3% |
| Pyerite | 3,644 | 4,214 | 100.00% | 37.5% |
| Tritanium | 4,394 | 5,081 | 100.00% | 37.5% |
|------------------------------------------------------------|
| Extra Material (4x) | | | | |
|------------------------------------------------------------|
| Isogen | 1 | 3 | 100.00% | |
| Mexallon | 1 | 4 | 100.00% | |
| Pyerite | 570 | 4,214 | 100.00% | |
| Tritanium | 687 | 5,081 | 100.00% | |
|------------------------------------------------------------|
| Skill (1x) | | | | |
|------------------------------------------------------------|
| Industry | 1 | 1 | | |
|------------------------------------------------------------|

Quote for: Tracking Link I, Material Efficiency Level: 0
Production Efficiency Skill: I, Production Slot Modifier: 1.2[/pre]

edit: DAMN YOU NEW BOARD!!!! for not having a PRE!!! - just quote and copy n' paste the bill of materials into a notepad ...

can you provide a screen shot of your BOM?

The quote comes from my private programs.
If you are sure about the 5, then I'll take a look into it on Singularity myself and erect a POS there and test this issue.
I spent quite a while on the code, so I'd like it to be right :)

Also note that if you round the raw and extra separately then the materials of lockbreaker blueprint are wrong. Only if you round them together you get correct results.
So maybe the problem with tracking link is not mexallon, but isogen which should be 2 instead of 3. (using rapid assembly with a level 5 PE skill).

Finally I got it, but I'm not pleased with the outcome :(

As a preface and reminder. There are three types of waste, which have to be applied to the base material quantity given in the database:

Material Waste: The base waste factor of every item is modified by the ME level of the blueprint. The formula for ME < 0 is MW = wasteFactor * (1 – ME), for all other it is MW = wasteFactor * (1 / (ME + 1)). With the "perfect ME level" there is no additional waste.
The wasteFactor is found in the invBlueprintTypes table of the database dumb. The number provided there should be considered in percent. So a waste factor of 10 equals 10%.

Skill Waste: The base skill waste for lack of knowledge is 25%. Each level of Production Efficiency reduces the waste by 5%. The formula is SW = 1.25 – (0.05 * PE level). With perfect PE, there is no additional waste.

Production Waste: Each assembly line has a material multiplier. Usually it is 1 (one), but for some assembly lines (Rapid Equipment, Advanced Small/Medium/Large Ship Assembly) it is higher. The formula is easy: PW = baseMaterialMultiplier
The baseMaterialMultiplier can be found in the ramAssemblyLineTypes of the data dump.


I’ve already showed in [1], that both raw and extra materials are subject to waste. Raw materials are affected by all three types of waste. Extra materials only by skill and production waste.
To determine the effective material quantity the formula looks like this:

Quantity = ROUND(rawMaterialQuantity * (SW * PW + MW) + baseExtraMaterialQuantity * SW * PW)

The rawMaterialQuantity is derived from the quantity to be found in the invTypeMaterials table, the baseExtraMaterialQuantity is the quantity from the ramTypeRequirements table from the database dump.
Mathew and Fubar explained in [2] how to calculate the material quantity (if an item in the ramTypeRequirements table is marked as recycle = 1, then the materials for this component have to be subtracted by the quantities of the actual item: e.g. the Gyrostab II has amongst other things 890 units of Tritanium as raw materials requirement and a recycle flagged Gyrostab I module as an extra material. Thus the 696 units of Tritanium for the Gyrostab I have to be removed from the raw quantity of 890, making 194 (= 890 – 696) units of Tritanium the new rawMaterialQuantity.

The problem: The formula above sometimes didn't work out as the OP already mentioned with the Tracking Link I.
Until this post, I was sure the rounding was done like in accounting: ROUND DOWN on values < 0.5, otherwise ROUND UP. Well, this is not true for every value: CCP rounds in favor for the user, when it comes to the number 2.5! Then the value is rounded down to 2!!!

Take a look at this quote. It is the quote of a ME 0 Gyrostab I Blueprint (MW = 0.1), my char had a PE level of 3 (SW = 1.15) and I was building at a normal equipment array (PW = 1). The database dump states the raw material quantity of 2 units of Zydrine. So the above formula would give a final Quantity = ROUND(2 * (1.15 * 1 + 0.1) + 0 * 1.15 * 1) = ROUND(2.5) ... which would be normally 3, but the actual bill of material says 2 (and also uses up 2 units, when built!).

The same thing happend here (Tracking Link I, ME 0, PE 1, standard array) and here (Tracking Link I, ME 0, PE 5, rapid array)

As far as I can tell, all values above 2.5 are rounded correctly, like in accounting. Values between 1 and 1.499999999 are rounded down too. Above 1.5 and below 2 is rounded up. But I cannot test the behaviour of 1.5 (no such base quantity exists in the db!)


Peace Math is a lie!


[1] http://community.eveonline.com/ingameboard.asp?a=topic&threadID=1561558
[2] http://community.eveonline.com/ingameboard.asp?a=topic&threadID=1220024 (posts #30 and #31)

It /may/ be a floating point issue. I know javascript throws them sometimes with some of the code I've been using. (nothing fancy. just a *100, floor, then /100. I shouldn't end up with 6 decimals, but sometimes it happens.

Yeah, I was thinking the same, but in the end, how can (1+1)*(1*1.2)+1*0.1 = 2.5 (raw+extra)*(SW+PW)+raw*MW generate a floating point defect? No dividends, No more than one digit after the comma.
Unless CCP does something really, really wired, ofc ;)

Well round to even is not correct either... After checking eccm projector ii, only with normal round you can get the correct materials. I dont understand whats so special with that tracking link blueprint...