These forums have been archived and are now read-only.

The new forums are live and can be found at https://forums.eveonline.com/

Science & Industry

 
  • Topic is locked indefinitely.
 

Blueprint data adjustments thread

First post First post
Author
Previous Topic Next Topic
Seith Kali
Caldari Provisions
Caldari State
#421 - 2014-06-10 00:19:46 UTC
Jon Lucien wrote:

The assumptions people make sometimes astound me. I, too, can pull numbers out of my backside. Look, Statistics!


Good postings Jon Lucian, good postings.

I have some statistics too.

Sabriz Adoudel wrote:
I am assuming that 20% of producers produce 80% of the manufactured goods in EVE, and that those people seldom get it wrong.


15% of the stuff you say is incoherent drivel, and the other 85% of the time I'm not listening.

Apprentice Goonswarm Economic Warfare Consultant - Drowning in entitlement and privilege. 
Ranamar
Nobody in Local
Deepwater Hooligans
#422 - 2014-06-10 05:15:56 UTC
We know that about 10% of the EVE population installs a manufacturing job each year, and all the manufacturing for the year could be handled by 1-2% of the population of EVE Online. (by characters)

So... 80/20 may in fact not be far off.

However, anyone who claims that *everything* is going to come to a screeching halt as soon as these changes go through because prices will go up is living in a silo. Sure, some things are going to have a pretty severe cost overhangs the way some BCs cruisers and battleships currently do. However, some other things are going to sell out within the week, because I've seen what production overhangs turn into big spikes in price because supply doesn't catch up immediately when the backlog runs out. There *will* be things to produce. They *will* have profit. And, with all the people entering the system, I'm expecting most of them to have their profits driven down to nearly 0 while the new experimenters figure out what to build.

Also, how much of that massive increase in assets per year is supercaps, I wonder?
Seith Kali
Caldari Provisions
Caldari State
#423 - 2014-06-10 06:11:53 UTC
50/72 may not be far off either. How the hell would you know.

Apprentice Goonswarm Economic Warfare Consultant - Drowning in entitlement and privilege. 
Aluka 7th
#424 - 2014-06-10 06:19:30 UTC
Soldarius,

My calculation starting point, 11000 materials and 1250minutes, is for TE/ME0 like it says and it said when you wrote response.
That means 10% waste included so..... That IS why I chose those numbers because they are 10000 material base and 1000minutes base.
Sometimes I go back to post and edit for clarification but this was written from the get go so please read more carefully.
Soldarius
Dreddit
Test Alliance Please Ignore
#425 - 2014-06-10 15:05:37 UTC
Aluka 7th wrote:
Soldarius,

My calculation starting point, 11000 materials and 1250minutes, is for TE/ME0 like it says and it said when you wrote response.
That means 10% waste included so..... That IS why I chose those numbers because they are 10000 material base and 1000minutes base.
Sometimes I go back to post and edit for clarification but this was written from the get go so please read more carefully.


And for the purposes of actual numbers that we can work with in the formulas, your numbers were wrong. You cannot start at final numbers when its the base BP numbers and the formulas that are getting changed. Your post had nothing about base numbers. So I found it both proper and necessary to do those calculations to find them.

http://youtu.be/YVkUvmDQ3HY
CCP Greyscale
C C P
C C P Alliance
#426 - 2014-06-10 17:20:27 UTC
Ok, sorry for the absence, holidays and business kept me away. Let's get back on topic!

CCP Greyscale wrote:
probag Bear wrote:
Qoi wrote:
Oh, and please make sure that extra materials only get the normal 1/0.9 modifier Blink


Are there actually any inventables that use significant amounts of extra materials? If so, this is actually a very significant point I'm surprised no one's noticed yet. If not, I'm okay with making jobs take up a couple of extra R.A.M. for the sake of consistency.

And while I'm at it, just to make sure,
Greyscale, you already know this, but just as a reminder: the Tech column in the invBlueprintTypes table is a bit misleading. A significant number of blueprints are marked as Tech 2, but are not actually inventables; they're basically just T1 blueprints. T2 components are an example off the top of my head.
Naturally, those blueprints should not get the same material requirements change as actual T2 inventables.


Will double-check this on Monday, there are various conflicting sets of data to do with what is and isn't T2, and I think the metric the spreadsheet is using is accurate.


Looks like my sheet is doing OK, it reports RAM and both normal and capital T2 components as T1, so that should all wash out nicely.

Sabriz Adoudel wrote:
CCP Greyscale wrote:
Hey, I'm not upset with anyone - I meant literally what I said, which was that when Theng says "this is what I was worried about", I was not clear what the "this" was referring to. If it's about dealing with some invention issues as part of a follow-on package of changes to invention, then I totally understand the skepticism. Normally I wouldn't even be suggesting that we're waiting for later patches in public, but I'm sufficiently confident of these ones happening that I'm making an exception in this caes. I totally take the point that planning for a follow-up is risky, but the counterpoint in my mind is that the changes that were shelved were only being considered because they *were* a band-aid that would be ripped off shortly anyway. Normalizing to 24h with the current queuing mechanics is not a good long-term fix, in our eyes, so if we plan on the assumption that there is no follow-up, we still wouldn't be implementing that change :)

Sabriz Adoudel wrote:
Concrete suggestion.

Module baseline materials post Crius = 145% of present baseline.

- This reflects that most modules are produced without decryptors (153%) but a significant minority use Symmetry.


T2 Frigates and Destroyers baseline materials post Crius = 135% of present

- This reflects the widespread use of Symmetry decryptors on these.


T2 cruisers and BCs baseline materials post Crius =125% of present

These are usually invented with Accelerant or Process, sometimes Parity or Symmetry in unusual market conditions.


T2 battleships and larger baseline materials post Crius = 120% of present

These are always invented with Process.


I'll have another look at the math here next week. This looks like it lines up very nicely with current numbers, but it's not totally clear yet that we do want to match them exactly. The current balance is to a large degree arbitrary, and if it weren't for the fact that we had done some very specific balance changes to moon minerals to deal with bottlenecks that we don't want to revisit, I wouldn't be attempting to get any closer than ballpark numbers anyway, because trying to tune to match arbitrary legacy balance is generally a poor use of time. If it looks like these changes will significantly impact moon mineral values then we'll probably do some more tuning, but if we'd just be changing everything in the same ratios with no impact on economic balance beyond the price of ships, that's probably not something that merits fine-tuning simply to match existing numbers. If there's a cost-balance problem with T2 ships then that would be much more sensibly addressed by a targeted balance pass on those costs; simply preserving the status quo has no value if the status quo itself has no specific value.

Again, I'm happy to be argued around on this sort of thing :)



The reason I suggest you try to keep somewhat close to in line with the current numbers is that if you do not, it will cause a sharp divergence between market costs and build costs for some items.

For example, if effective Ishtar build costs (assuming a sensible decryptor, which is Process or Accelerant) go from 180m to 230m after Crius, this will cause a period where Ishtars are no longer profitable to produce at all, as existing stockpiles produced at 180m will still exist.

Now this isn't much of an issue for a ship as widely used as the Ishtar - stockpiles will run out quickly. Where it is a much bigger issue is on less frequently destroyed ships, such as the Damnation, the Sin, the Broadsword, or the Lachesis - the market on these ships may not recover for 12 months or more.

Concrete example: The Brutix has not recovered from the increase in build cost back in tiericide. It is *still* under production cost due to existing stockpiles that were built prior to the material cost increase.


I believe one of Crius's goals is to get more people into production, to get more people producing weapons of war. If it gets these people to try production once, lose a lot of ISK, then quit it, it has failed.


Some buffer is actually a positive from our point of view, as it makes it more likely that the market will ride out launch-month disruption without serious shortages. Obviously it's a bigger issue for some products than others, but if specific addressable problems appear post-release we'll deal with them. In markets that are already chronically broken, though, we will likely just let them work themselves through naturally rather than trying to intervene to fix what are at least somewhat player-created problems.

[quote=Aluka 7th]Did anyone notice much...
Dom Roland
Butlerian Crusade
#427 - 2014-06-10 17:26:32 UTC
mynnna wrote:

Looking at the increase in build time only is flawed and only relevant for Tech II BPOs. For invention, the number that matters is total time to market, copying, invention, and building.


Actually its probably MAX(build time, copy/invent time) as there are two different slot resources that are a bottleneck, and the work slot usage can be parellelised.

Consider item A with a 30 day build time and a 1 day copy/invent time and item B with a 1 day build time and a 30 day copy/invent time. These both have the same TTM and we will see them produced in roughly the same quantity (assuming equal demand). Now consider item C with a 15 day build time and a 16 day copy/invent time. This as the same TTM as A and B but we will see roughly double the quantity produced (assuming the same demand) as 2 batches of C will finish for every single batch of A or B. This shows that TTM is a flawed metric. It best represents the inertia of the market - how quickly manufacturers can react to a price spike.
CCP Greyscale
C C P
C C P Alliance
#428 - 2014-06-10 17:28:01 UTC
Dom Roland wrote:
mynnna wrote:

Looking at the increase in build time only is flawed and only relevant for Tech II BPOs. For invention, the number that matters is total time to market, copying, invention, and building.


Actually its probably MAX(build time, copy/invent time) as there are two different slot resources that are a bottleneck, and the work slot usage can be parellelised.

Consider item A with a 30 day build time and a 1 day copy/invent time and item B with a 1 day build time and a 30 day copy/invent time. These both have the same TTM and we will see them produced in roughly the same quantity (assuming equal demand). Now consider item C with a 15 day build time and a 16 day copy/invent time. This as the same TTM as A and B but we will see roughly double the quantity produced (assuming the same demand) as 2 batches of C will finish for every single batch of A or B. This shows that TTM is a flawed metric. It best represents the inertia of the market - how quickly manufacturers can react to a price spike.


Yup. This is why we're trying to match both sides of that equation, so the load is balanced equally for all invention types. Both because it's a nice efficiency, and because by having everything in the center we don't have to worry about these sorts of issues while balancing!
mynnna
State War Academy
Caldari State
#429 - 2014-06-10 18:22:01 UTC
Dom Roland wrote:
mynnna wrote:

Looking at the increase in build time only is flawed and only relevant for Tech II BPOs. For invention, the number that matters is total time to market, copying, invention, and building.


Actually its probably MAX(build time, copy/invent time) as there are two different slot resources that are a bottleneck, and the work slot usage can be parellelised.

Consider item A with a 30 day build time and a 1 day copy/invent time and item B with a 1 day build time and a 30 day copy/invent time. These both have the same TTM and we will see them produced in roughly the same quantity (assuming equal demand). Now consider item C with a 15 day build time and a 16 day copy/invent time. This as the same TTM as A and B but we will see roughly double the quantity produced (assuming the same demand) as 2 batches of C will finish for every single batch of A or B. This shows that TTM is a flawed metric. It best represents the inertia of the market - how quickly manufacturers can react to a price spike.


Think of TTM as an expression of how much of a character's available job time is consumed for a given item. On a typical research and production alt you've got 7200 hours of research time and 7200 hours of build time per month per character, and if you copy ten runs of a blueprint at 72h each, then build from them at 72h each, you've used 10% of your total build time. That's true whether the copies are one runs in ten slots or ten runs in one slot. So when looking at overall throughout game-wide - which I am - it's a very useful metric, not flawed at all.

Member of the Goonswarm Economic Warfare Cabal
Aryth
University of Caille
Gallente Federation
#430 - 2014-06-10 18:51:00 UTC
I am really hoping -20% ME teams is a bug/typo. Or you can throw all these calcs out the window.

Leader of the Goonswarm Economic Warfare Cabal.

Creator of Burn Jita

Vile Rat: You're the greatest sociopath that has ever played eve.
Theng Hofses
State War Academy
Caldari State
#431 - 2014-06-10 20:39:04 UTC
Is this BPO translation as intended?

Providence ME 7/1 > 9/1

Propulsion 20 > 22
Armor 15 > 17
Cargo 76 > 84
Construction 49 > 54

On a different note: The current Sisi build is completely overwhelmed by large number of BPO/BPCs.... I have like 20,000+ in one system and ~1000 or more in other systems and the industry module becomes basically unusable as you can't find anything.
Steve Ronuken
Fuzzwork Enterprises
Vote Steve Ronuken for CSM
#432 - 2014-06-10 22:28:04 UTC  |  Edited by: Steve Ronuken
Wasn't it decided to have the T1 materials for things not being multiplied up?

I'm seeing T2 copies needing multiple of the t1 items per run. (T2 1MN afterburner needing 15 T1 versions, for 10 runs)

It's not so important for some, as it's a fairly low cost, but for a T2 siege module, that's another 64 million on top of the 94 million they currently cost.


In addition, the numbers don't appear to be working right.

I have a light ion blaster I blueprint.
It's ME 10, displaying a wastage factor of 0.9% (should be 0?)

The bill of materials and the industry pages have slightly different numbers on them, neither of which matches the kind of numbers on TQ at the moment.

22 isogen
83 mexallon
144 pyerite
1667 tritanium.

So that's 10% more isogen (same as an unresearched blueprint on tq)
83 Mexallon
144 pyerite (1 more)
1667 trit (27 more)

(now I'm vague remembering that this was said it'd be like this. I think it needs to be a little more obvious that these aren't the actual production numbers.)


Sticking it into an array to manufacture a single run, I'm seeing
15 isogen (5 below current perfect)
57 mexallon (18 below)
98 pyerite (32 below)
1126 tritanium (374 below)

so that's around a 25% reduction with no team. (5 other arrays in that pos, if that's important)


In a station, they're coming out at around the numbers expected. (one more trit. I can see that being a rounding artefact, and not important)

Woo! CSM XI!

Fuzzwork Enterprises

Twitter: @fuzzysteve on Twitter
Kenneth Feld
Habitual Euthanasia
Pandemic Legion
#433 - 2014-06-11 00:32:21 UTC
Man, this is FUBAR

2 brutix to make an astarte
Most t1 items for T2 production are hosed
Multiple run BPC cut down to 1 run bpc

You have got to do something about load times……

I only have like 32 containers all with the max number of BPC before I gave up on invention - takes like hours for them all to scroll thru.
stoxxine
OLVI industries
#434 - 2014-06-11 05:17:02 UTC
I believe stuff like this; T1 ship to build T2 ship, was the original intent of the "extra materials" (ram as its called in database). It was just abused as part of rebalances, it was not broken fundamentally.

confusing for some, yes.

Disclaimer: The above was probably written drunk or by a friend on my pc or a hacker. No warranty for any misinformation provided.
probag Bear
Xiong Offices
#435 - 2014-06-11 09:39:46 UTC
CCP Greyscale, whenever things calm down, may I request the following for modeling purposes:

  1. The rounding function (round() / ceil() / floor() ) used when applying the global 1/0.9 material requirements modifier.
  2. If you do this in a separate pass, the rounding function used when applying the T2 material requirements modifier. If it's in the same pass as 1, 1 would already answer this.
  3. The rounding function used when applying ME reduction to manufacturing jobs. I believe this was stated to be ceil(), but that was two threads back and several weeks ago.



(Reason I'm asking for these: take a T2 blueprint that requires 1 of a particular material. After all these changes, 10 runs may end up requiring anywhere between 10 and 30 of that material, depending on what the rounding functions are.)
CCP Greyscale
C C P
C C P Alliance
#436 - 2014-06-11 10:24:52 UTC
General request: UI/display issues to the Manufacturing/UI feedback thread in Test Server Discussions, please, they're more likely to be seen there :)

Aryth wrote:
I am really hoping -20% ME teams is a bug/typo. Or you can throw all these calcs out the window.


Bug.

Steve Ronuken wrote:
Wasn't it decided to have the T1 materials for things not being multiplied up?

I'm seeing T2 copies needing multiple of the t1 items per run. (T2 1MN afterburner needing 15 T1 versions, for 10 runs)

It's not so important for some, as it's a fairly low cost, but for a T2 siege module, that's another 64 million on top of the 94 million they currently cost.


In addition, the numbers don't appear to be working right.

I have a light ion blaster I blueprint.
It's ME 10, displaying a wastage factor of 0.9% (should be 0?)

The bill of materials and the industry pages have slightly different numbers on them, neither of which matches the kind of numbers on TQ at the moment.

22 isogen
83 mexallon
144 pyerite
1667 tritanium.

So that's 10% more isogen (same as an unresearched blueprint on tq)
83 Mexallon
144 pyerite (1 more)
1667 trit (27 more)

(now I'm vague remembering that this was said it'd be like this. I think it needs to be a little more obvious that these aren't the actual production numbers.)


Sticking it into an array to manufacture a single run, I'm seeing
15 isogen (5 below current perfect)
57 mexallon (18 below)
98 pyerite (32 below)
1126 tritanium (374 below)

so that's around a 25% reduction with no team. (5 other arrays in that pos, if that's important)


In a station, they're coming out at around the numbers expected. (one more trit. I can see that being a rounding artefact, and not important)


Yup, T1 in T2 is a bug as noted in the feedback thread :)

Showinfo hasn't been updated yet, all numbers in there can be assumed to be wrong. There's a lot of reported oddness in various numbers in the implementation, I'm waiting for a general review of that sort of thing before worrying about specifics.

Theng Hofses wrote:
Is this BPO translation as intended?

Providence ME 7/1 > 9/1

Propulsion 20 > 22
Armor 15 > 17
Cargo 76 > 84
Construction 49 > 54

On a different note: The current Sisi build is completely overwhelmed by large number of BPO/BPCs.... I have like 20,000+ in one system and ~1000 or more in other systems and the industry module becomes basically unusable as you can't find anything.


Not 100% sure, but see above; lots of slightly odd numbers floating around right now.

probag Bear wrote:
CCP Greyscale, whenever things calm down, may I request the following for modeling purposes:

  1. The rounding function (round() / ceil() / floor() ) used when applying the global 1/0.9 material requirements modifier.
  2. If you do this in a separate pass, the rounding function used when applying the T2 material requirements modifier. If it's in the same pass as 1, 1 would already answer this.
  3. The rounding function used when applying ME reduction to manufacturing jobs. I believe this was stated to be ceil(), but that was two threads back and several weeks ago.



(Reason I'm asking for these: take a T2 blueprint that requires 1 of a particular material. After all these changes, 10 runs may end up requiring anywhere between 10 and 30 of that material, depending on what the rounding functions are.)


=ROUND({TQ value}*IF({tech level}=2,1.5,1)/0.9,0)

Material fractions are always rounded up in the manufacturing code.
Apelacja
Sad Najwyzszy
#437 - 2014-06-11 12:07:20 UTC  |  Edited by: Apelacja
I logged on sisi and i see some weird costs for me/pe ? was it implemented yet?

the same with some bpos production cost like 1 b per run O.o:?

Also noticed differences between new invention output and already invented bpc after conversion - is it intended?
Steve Ronuken
Fuzzwork Enterprises
Vote Steve Ronuken for CSM
#438 - 2014-06-11 12:10:12 UTC
oops. missed the T2/T1 bug mentioned (blush)

Woo! CSM XI!

Fuzzwork Enterprises

Twitter: @fuzzysteve on Twitter
TheSmokingHertog
Julia's Interstellar Trade Emperium
#439 - 2014-06-11 17:18:43 UTC
Link to the test forum thread.

"Dogma is kind of like quantum physics, observing the dogma state will change it." ~ CCP Prism X

"Schrödinger's Missile. I dig it." ~ Makari Aeron

-= "Brain in a Box on Singularity" - April 2015 =-
Soldarius
Dreddit
Test Alliance Please Ignore
#440 - 2014-06-11 19:55:11 UTC
CCP Greyscale wrote:

=ROUND({TQ value}*IF({tech level}=2,1.5,1)/0.9,0)

Material fractions are always rounded up in the manufacturing code.


My apologies in advance. I really dislike directly contradicting a dev. But while the above formula is correct (as far as rounding is concerned), the follow-on statement is most definitely not.

I can confirm beyond a shadow of a doubt that material wastage does not always round up. Rigs make excellent examples. Almost all small rigs have 0 material waste even at ME0. But certain ones like Small Ancillary Current Routers do not. Those become perfect at ME1. This would not be true if all materials calculations rounded up.

Small Ancillary Current Router I
ME0 10% waste:

  • Burned Logic Circuits: 10
  • Fried Interface Circuits: 2
  • Tripped Power Circuits: 10


ME1 5% waste (and also perfect):

  • Burned Logic Circuits: 9
  • Fried Interface Circuits: 2
  • Tripped Power Circuits: 9


Clearly, the BPC wastage is rounded to the nearest whole number, and not always up.

http://youtu.be/YVkUvmDQ3HY