Pages: [1] :: one page |
|
Author |
Thread Statistics | Show CCP posts - 0 post(s) |
El'essar Viocragh
Minmatar FSK23
|
Posted - 2007.11.26 12:02:00 -
[1]
So, couple of days ago I fiddled with POS assemblies and other manufacturing stuff on Sisi just to check if everything works out, bugreported the oddities, some of which are already fixed.
Well, today I got the reply to a bugreport about production waste changes on POS assemblies. It is an intended change and will apply to NPC station manufacturing too (I couldn't test that when I created the BR as an UI error prevented manufacturing in stations). Just checked stations, aye, now there too. So, what's happening?
To remove some waste calculation inconsistencies now all rounding is after multiplying the material requirements with the amount of production runs, not before.
What does this mean?
A ME20 Thrasher BPO lists 12 Megacyte as material requirement (0.47% waste listed on the print). Build 1, and this are right. Build 10, and you need 121. Build 100, and you need 1206. Talking about inconsistences, eh?!
Or in other words, fire up your Excel now. Without calculating with fractional material requirements all the way, you effectively can't say how much you need for a production run anymore.
I don't like it, but that could be because I'm one of that oldschool pocket calculator types. My most useful tool in Eve just got nerfed - the calculator! Dang, I totally did not see that happening -- [17:47] <Mephysto> its dead, jim |
Leishent
Caldari Captured Souls Inc.
|
Posted - 2007.11.26 12:11:00 -
[2]
So what you are saying is, what now is calculated using the following formula
Required_Amount = Round(Base_Amount * ((1 + (Default_Blueprint_Waste_Factor / (1 + Blueprint_Material_Level))) + (0.25 - (0.05 * Production_Efficiency_Skill_Level))), 0) * NumberOfItemsToBuild
Will now (after the EVE:Trinity patch is deployed) instead be
Required_Amount = (Base_Amount * ((1 + (Default_Blueprint_Waste_Factor / (1 + Blueprint_Material_Level))) + (0.25 - (0.05 * Production_Efficiency_Skill_Level)))) * NumberOfItemsToBuild
Correct?
|
El'essar Viocragh
Minmatar FSK23
|
Posted - 2007.11.26 12:19:00 -
[3]
They do one rounding at the end of everything so you don't use up half a megacyte, but basically yes.
I want that in my calculator -- [17:47] <Mephysto> its dead, jim |
Leishent
Caldari Captured Souls Inc.
|
Posted - 2007.11.26 12:23:00 -
[4]
So it need to be updated to the following to use the upcoming change
Required_Amount = Round((Base_Amount * ((1 + (Default_Blueprint_Waste_Factor / (1 + Blueprint_Material_Level)))+ (0.25 - (0.05 * Production_Efficiency_Skill_Level)))) * NumberOfItemsToBuild , 0)
Thanks for the heads-up
|
Matthew
Caldari BloodStar Technologies
|
Posted - 2007.11.26 12:39:00 -
[5]
He was probably talking about the inconsistency between the roundings on blueprint waste and skill waste.
On TQ, the blueprint waste gets rounded at the per-run level, but the skill waste is rounded at the total level (i.e. after being multiplied by the number of runs).
i.e. current TQ: ROUND(ROUND(Base * BPWastefactor) * Batches * SkillWastefactor)
Where BPWastefactor and skillwastefactor are expressed as multipliers (e.g. 5% waste is multiplier of 1.05)
The new sisi implementation is: ROUND(base*BPWastefactor*batches*skillwastefactor)
Which treats the two types of waste equally. It's also a simpler calculation, both to understand and compute.
Of course, as nearly all producers serious enough to use a spreadsheet tool have PE5 (and hence a skillwastefactor of 1), most people probably hadn't noticed this.
I've not tested what effect this has on the "skill waste bug" - where skill waste does not get applied if blueprint waste is zero.
Originally by: El'essar Viocragh A ME20 Thrasher BPO lists 12 Megacyte as material requirement (0.47% waste listed on the print). Build 1, and this are right. Build 10, and you need 121. Build 100, and you need 1206. Talking about inconsistences, eh?!
Not sure if this is new on sisi, but I don't remember seeing this on TQ: There is a disclaimer on the sisi blueprint BOM tab now that states that requirements are rounded for a single run and that multiple run jobs may differ.
With 0.47% waste, you really need 12.0564 units per run.
The old calculation effectively gave you a "free" 0.0564 units per run, regardless of how many you made.
While still not completely realistic, the new method is far closer to a realistic calculation.
If they wanted to go all the way, they'd change to always rounding up, so you'd need 13 units for a single run, and 1207 for 100 runs, working on the basis that you can't really pull partial units out of thin air, so you need to put in the full extra unit even if you don't use all of it. ------- There is no magic Wand of Fixing, and it is not powered by forum whines. |
El'essar Viocragh
Minmatar FSK23
|
Posted - 2007.11.26 12:48:00 -
[6]
Originally by: Matthew Not sure if this is new on sisi, but I don't remember seeing this on TQ: There is a disclaimer on the sisi blueprint BOM tab now that states that requirements are rounded for a single run and that multiple run jobs may differ.
LOL
I totally missed that one. Indeed, that is new.
And you are ofc right about the 12.05 units and the calculation being a bit simpler on the server level. But on the client and player level, not using a spreadsheet just became infeasible. -- [17:47] <Mephysto> its dead, jim |
Matthew
Caldari BloodStar Technologies
|
Posted - 2007.11.26 13:10:00 -
[7]
Edited by: Matthew on 26/11/2007 13:11:39 Edited by: Matthew on 26/11/2007 13:11:21
Originally by: El'essar Viocragh
Originally by: Matthew Not sure if this is new on sisi, but I don't remember seeing this on TQ: There is a disclaimer on the sisi blueprint BOM tab now that states that requirements are rounded for a single run and that multiple run jobs may differ.
LOL
I totally missed that one. Indeed, that is new.
And you are ofc right about the 12.05 units and the calculation being a bit simpler on the server level. But on the client and player level, not using a spreadsheet just became infeasible.
Yeah, I can understand that might be a bit frustrating if you're not used to it - I got way beyond the point where hand calculations were feasible a couple of years ago.
If you add 1 to all the BOM values, that would mean you can still get a quick calculation of the maximum needed by hand (it will overestimate), so if you get that much together, do the run, and subtract back off whatever is not taken by the job, could be workable.
Of course, just using excel isn't going to be easy either if you want true precision. The rounding logic excel employs isn't the same as that used by the eve factories (at least it isn't on TQ, I've not checked to see if sisi changed the rounding logic too).
The Excel ROUND function uses what I call "normal" rounding, i.e. 0.5 rounds up. So 1.5 rounds to 2, 2.5 rounds to 3, 3.5 rounds to 4.
The Eve factories use "round to even" rounding, i.e. 0.5 rounds to the nearest even whole number. So 1.5 rounds to 2, 2.5 rounds to 2, 3.5 rounds to 4.
(note that both methods round >0 to <0.5 down and >0.5 to <1 up)
To get "round to even", you'd need to use the VBA ROUND function, rather than Excel's own ROUND function. Alternatively, the ROUND function in Access also performs round to even.
While "round to even" does seem a bit of a pain, from a purely mathematical point of view, there is a reason for it - always rounding 0.5 up introduces a bias into the rounding. Round to even eliminates that bias assuming you have an even distribution of fractional values in your dataset.
The up-side of the new calculation is that if you do use "normal" rounding (either by mistake or as a "close enough" option), you'll only ever be one unit out. On the old calculation, you could potentially be out by 1 unit per run. ------- There is no magic Wand of Fixing, and it is not powered by forum whines. |
Sofitia Mourtos
GALAXIAN RAZOR Alliance
|
Posted - 2007.11.26 14:27:00 -
[8]
I am really annoyed about this change...
If I get it correct then when you build 1 unit you save a few minerals and when building e.g 1000units then you get that saving carried on.
with the new system you pay extra for producing high volumes :-( that really sucks - especially if you are a micromanager, then the temptation to sit infront the computer and do 1000x 1units instead of just 1000units will be very tempting....
---------------------------------------- WTB: Guardian BPO |
Matthew
Caldari BloodStar Technologies
|
Posted - 2007.11.26 14:47:00 -
[9]
Originally by: Sofitia Mourtos I am really annoyed about this change...
If I get it correct then when you build 1 unit you save a few minerals and when building e.g 1000units then you get that saving carried on.
with the new system you pay extra for producing high volumes :-( that really sucks - especially if you are a micromanager, then the temptation to sit infront the computer and do 1000x 1units instead of just 1000units will be very tempting....
Not really that tempting once you realise that you could probably make more running level 1 missions in a frigate for the same time that it would take you to place 1000 1 unit jobs (and that's just the clicking to do the jobs, not waiting for each to complete).
As I've mentioned though, I do agree that bulk should be cheaper, and ideally they would use a round up instead of round to even to keep that 1 run more expensive. ------- There is no magic Wand of Fixing, and it is not powered by forum whines. |
Jurgen Cartis
Caldari Interstellar Corporation of Exploration
|
Posted - 2007.11.26 19:58:00 -
[10]
If I'm understanding this right, Perfect ME just got slightly nerfed. Grr. . . now I want some other approximation of the economies of scale. . .
So did my calculator. -------------------- ICE Blueprint Sales FIRST!! -Yipsilanti Pfft. Never such a thing as a "last chance". ;) -Rauth |
|
Saladin
Minmatar Eternity INC. Mercenary Coalition
|
Posted - 2007.11.26 20:21:00 -
[11]
I think CCP needs to re-think this change. I could see this being an issue if my product was continuous (i.e. a single continuous bar of steel for instance) as opposed to being discrete units. Because I am producing discrete units, the requirement has to be the sum of the individual units, not some post-sum round off.
|
Saladin
Minmatar Eternity INC. Mercenary Coalition
|
Posted - 2007.11.27 00:57:00 -
[12]
Originally by: Matthew
Of course, as nearly all producers serious enough to use a spreadsheet tool have PE5 (and hence a skillwastefactor of 1), most people probably hadn't noticed this.
What about people who make t2 items at a POS where the advanced array has a factor of 1.1? How is the waste rounded then? For example, right now with my PE 5 it would take 20 condors to make 20 crows. After this change will it take 22 condors to make 20 crows?
|
El'essar Viocragh
Minmatar FSK23
|
Posted - 2007.11.27 02:31:00 -
[13]
It will take 20 the same way that it takes only 20 even without the PE skill at all.
Some materials on t2 blueprints are not affected by waste. RAMs, ships on t2 shipprints or advanced materials on t2 module prints for example.
In the Crow example it will increase the amount of advanced materials though.
-- [17:47] <Mephysto> its dead, jim |
Kylar Renpurs
Dusk Blade
|
Posted - 2007.11.27 02:47:00 -
[14]
This makes sense IMO. One thing that frustrates me is how I can reprocess 10 of an item and get 100% return on it's materials, but I reprocess 10000 of the same and I lose out. It's CCP bringing things inline with each other.
Improve Market Competition!
|
Captain Agemman
Minmatar Legio Ultra
|
Posted - 2007.11.27 08:04:00 -
[15]
You have the option to train the skills for perfect refining though.
Where are the skills that let you produce every print at 0% waste?
|
Jameroz
Art of War
|
Posted - 2007.11.27 08:52:00 -
[16]
I hope this change doesn't go through... will be pain in the ass to produce items one at a time. All this does is increases demand for ME research even further.
|
Matthew
Caldari BloodStar Technologies
|
Posted - 2007.11.27 09:00:00 -
[17]
Originally by: Jurgen Cartis If I'm understanding this right, Perfect ME just got slightly nerfed. Grr. . . now I want some other approximation of the economies of scale. . .
So did my calculator.
Yeah, with the new calculation there is no universal ME level that is "perfect". Instead, perfect ME now becomes a function of the number of runs in the job you want to do as well as of the largest individual material requirement. So a print could still be researched to perfect ME for a 10 run job, but you'd need to research it a lot further to be perfect on a 100 run job.
Originally by: Saladin I think CCP needs to re-think this change. I could see this being an issue if my product was continuous (i.e. a single continuous bar of steel for instance) as opposed to being discrete units. Because I am producing discrete units, the requirement has to be the sum of the individual units, not some post-sum round off.
Which works fine in the real world, where there's nothing stopping you subdividing materials down to the smallest denominator required (assuming it's above the molecular level). In eve, you have to deal with the fact that fractional units of items aren't possible to store.
The only thing they need to change is to make the waste round up. Then they're modelling the discrete unit requirements properly by forcing you to supply all the material required, rather than letting you magic partial units out of thin air. It also means that economies of scale would work properly. ------- There is no magic Wand of Fixing, and it is not powered by forum whines. |
Benn Helmsman
Caldari Dark Prophecy Inc.
|
Posted - 2007.11.27 12:53:00 -
[18]
Originally by: Jameroz I hope this change doesn't go through... will be pain in the ass to produce items one at a time. All this does is increases demand for ME research even further.
Dont forget to calculate the installation costs for jobs in empire, that could make installing single run jobs kinda foolish in some cases.
|
Ogul
Caldari ZiTek Deepspace Explorations Interstellar Alcohol Conglomerate
|
Posted - 2007.11.27 16:46:00 -
[19]
Edited by: Ogul on 27/11/2007 16:47:07
Originally by: Kylar Renpurs This makes sense IMO. One thing that frustrates me is how I can reprocess 10 of an item and get 100% return on it's materials, but I reprocess 10000 of the same and I lose out. It's CCP bringing things inline with each other.
Except they break the working stuff to be in line with the already broken stuff. Smart move.
And I fail to see how any of the server side calculations get any simpler by changing the order of mupltiplying and rounding.
--- This is a war declaration, issued from your alt corp. It is used to gank people in high sec. |
Matthew
Caldari BloodStar Technologies
|
Posted - 2007.11.27 21:34:00 -
[20]
Originally by: Ogul And I fail to see how any of the server side calculations get any simpler by changing the order of mupltiplying and rounding.
It now only rounds once instead of twice. Previously it would round after calculating the per-run BP waste, then again after taking into account the number of runs and the players level of Production Efficiency. Now it just rounds once at the end. ------- There is no magic Wand of Fixing, and it is not powered by forum whines. |
|
Kylar Renpurs
Dusk Blade
|
Posted - 2007.11.28 00:19:00 -
[21]
Quote: Except they break the working stuff to be in line with the already broken stuff. Smart move.
What's working about it in the first place? I find it pretty unusual that manufacturing *doesn't* round up in every scenario.
For example, something takes 10.4 Armor Plates to produce. Under the current system, That'd take 10 armor plates to produce 1, and 1000 armor plates to produce 100
What I've gotta ask is how the heck do you make something without that extra 0.4? You break up a unit of the stock and throw the rest away as waste, thats how. You can't just say "well, this component needs to be 1.4 meters long, lets make it 1 meter then"
If the system weren't *broken* at the moment, it'd take you 11 Armor Plates to make that item, and 1040 to make a hundred. I would *much* prefer a system where my external calculations mean I over-estimate my costs, rather than under-estimate.
Improve Market Competition!
|
Ogul
Caldari ZiTek Deepspace Explorations Interstellar Alcohol Conglomerate
|
Posted - 2007.11.28 01:57:00 -
[22]
Edited by: Ogul on 28/11/2007 02:00:18 Edited by: Ogul on 28/11/2007 01:57:07
Originally by: Matthew
It now only rounds once instead of twice. Previously it would round after calculating the per-run BP waste, then again after taking into account the number of runs and the players level of Production Efficiency. Now it just rounds once at the end.
In that case the smart thing to do would be to include the production efficiency multiplier in the single run build cost, round and then multiply by the number of runs.
Originally by: Kylar Renpurs
What's working about it in the first place? I find it pretty unusual that manufacturing *doesn't* round up in every scenario.
For example, something takes 10.4 Armor Plates to produce. Under the current system, That'd take 10 armor plates to produce 1, and 1000 armor plates to produce 100
At least that's logical behaviour concerning the number of runs, i. e. if you need 1 A to build 1 B you should need 100 A to build 100 B.
With the new rounding in place you might just get a situation where 1 A is still needed to build 1 B but building 100 B requires 149 A.
And THAT is idiotic.
P. S. You won't find any item in EVE that needs 10.4 armor plates, it's either 10 or 11 and increasing blueprint ME levels might or might not decrease that number. But it will always be an integer.
--- This is a war declaration, issued from your alt corp. It is used to gank people in high sec. |
Kylar Renpurs
Dusk Blade
|
Posted - 2007.11.28 03:49:00 -
[23]
Quote: ...And THAT is idiotic.
P. S. You won't find any item in EVE that needs 10.4 armor plates, it's either 10 or 11 and increasing blueprint ME levels might or might not decrease that number. But it will always be an integer.
Oh yeah, I agree totally with all that. My argument was more that they are replacing a broken system with a broken system, not a false-logical system with a true-illogical system.
Personally I'd prefer material requirements to be read as, ie Armor Plates 11 (10.4) or somesuch (running with the last example), at least then it's *true* and *logical*.
Information is lost because they're trying to hide floating point numbers.
Improve Market Competition!
Originally by: Shadarle Exce |
Pwett
Minmatar QUANT Corp. QUANT Hegemony
|
Posted - 2007.11.28 05:12:00 -
[24]
My solution to this problem is to charge everyone 10% more for everything.
Sorry little guy! _______________ Pwett CEO, Founder, & Executor <Q> QUANT Hegemony
|
Ogul
Caldari ZiTek Deepspace Explorations Interstellar Alcohol Conglomerate
|
Posted - 2007.11.28 07:51:00 -
[25]
Originally by: Kylar Renpurs
Personally I'd prefer material requirements to be read as, ie Armor Plates 11 (10.4) or somesuch (running with the last example), at least then it's *true* and *logical*.
Information is lost because they're trying to hide floating point numbers.
If you could have a fraction of an armor plate in your hangar, there wouldn't be a problem. But as it is, that would still implicate the possibility of cheaper manufacturing by splitting your activities into smaller jobs.
I'd hate that.
--- This is a war declaration, issued from your alt corp. It is used to gank people in high sec. |
Matthew
Caldari BloodStar Technologies
|
Posted - 2007.11.28 09:33:00 -
[26]
Originally by: Ogul Edited by: Ogul on 28/11/2007 02:00:18 Edited by: Ogul on 28/11/2007 01:57:07
Originally by: Matthew
It now only rounds once instead of twice. Previously it would round after calculating the per-run BP waste, then again after taking into account the number of runs and the players level of Production Efficiency. Now it just rounds once at the end.
In that case the smart thing to do would be to include the production efficiency multiplier in the single run build cost, round and then multiply by the number of runs.
So what you're advocating is:
round(base*bpwastefactor*skillwastefactor)*runs
That could be an option. Though doing it that way, you then lose the ability to implement a simple economies of scale model by making it a round up instead of a round to even.
Originally by: Ogul At least that's logical behaviour concerning the number of runs, i. e. if you need 1 A to build 1 B you should need 100 A to build 100 B.
With the new rounding in place you might just get a situation where 1 A is still needed to build 1 B but building 100 B requires 149 A.
And THAT is idiotic.
Well in the first case, you're likely to be pulling a partial unit out of thin air, which isn't a good solution either. But I do agree the current situation needs improving.
If the base usage is 1.49 per unit, you should need 2 to build 1 unit and 149 to build 100 units.
Originally by: Ogul P. S. You won't find any item in EVE that needs 10.4 armor plates, it's either 10 or 11 and increasing blueprint ME levels might or might not decrease that number. But it will always be an integer.
All mineral requirements exist as a double (i.e. a double precision floating point number) until the point of actually taking the minerals. That CCP has chosen to set the base requirements as having zero decimal component is probably a convenience for them, no more. The only reason rounding is applied at all is because the inventory system cannot store fractions of a unit of something (though they did kinda do a hack to implement that with RAMs by storing the fraction as damage on the item).
Originally by: Ogul If you could have a fraction of an armor plate in your hangar, there wouldn't be a problem. But as it is, that would still implicate the possibility of cheaper manufacturing by splitting your activities into smaller jobs.
I'd hate that.
If the rounding is round to even, then you'd be right. However, his example specifically rounded 10.4 to 11 - a round up.
Using a round up, a 1 run job will always be an inefficient option. Lets say you have a base requirement of 1.2 units per run. Some examples using a round up:
1 run: 2 units (50% batch size waste (2-1.2)/1.2) 10 runs: 12 units (0% batch size waste) 57 runs: 69 units (0.88% batch size waste (69-68.4)/68.4) 589 runs: 707 units (0.03% batch size waste (707-706.8)/706.8)
Basically, you'd have two effects. Firstly, there will be certain optimal run sizes where you can get minimal waste (i.e. there are local minima in the cost vs batch size graph). In reality these minima won't be zero because of having to try and satisfy multiple materials at the same time. However, because the batch size waste could only ever be <1 unit, larger runs share that 1 unit of waste over more items. Hence larger batches will, barring local minima, always be more efficient.
Items with larger requirements will also be more efficient than smaller ones, as again the <1 unit of waste would be shared over a larger total mineral requirement.
The new sisi calculation only needs to be changed to rounding up to be both logical and create a reasonable economies of scale model. ------- There is no magic Wand of Fixing, and it is not powered by forum whines. |
Dario Kaelenter
|
Posted - 2007.11.29 06:07:00 -
[27]
Originally by: Pwett My solution to this problem is to charge everyone 10% more for everything.
Sorry little guy!
rofl ... best and most logical solution to CCP's new problem. Helps cover the cost of all those updates to production systems and tools that will be needed too
|
Nhaz
THE FINAL STAND Divine Retribution Alliance
|
Posted - 2007.12.16 18:18:00 -
[28]
While im no production guru. Setting 1 production run i get the correct amount with no added waste and no rounding up.
setting 2 production runs you start to see waste.
why? if its rounding down on the first run. why the heck isn't it rounding down on the second run.
in effect you can do two single production runs and see no rounding up. If i can do this then the way it calculates waste isn't correct, and the way its being done now is bugged.
especially if you are dealing with rigs. where some of the items are fairly expensive parts.
with tight competition the difference between 1 part matters.
my opinion on this is if then intend to keep this way of doing the math then they need to exclude the rigs from the wastage calc. _____________________________________________
It's NOT paranoia, If they REALLY ARE out to get you! |
|
|
|
Pages: [1] :: one page |
First page | Previous page | Next page | Last page |