These forums have been archived and are now read-only.

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

EVE Information Portal

 
  • Topic is locked indefinitely.
 

Dev blog: Industry & 3rd Party Developers

First post First post
Author
Previous Topic Next Topic
Skalle Pande
Teknisk Forlag
#21 - 2014-07-10 22:42:32 UTC
Quote:
Research Multipliers
Level Multiplier
1 1
2 29 / 21.0
3 23 / 7.0
4 39 / 5.0
5 278 / 15.0
6 928 / 21.0
7 2200 / 21.0
8 5251 / 21.0
9 4163 / 7.0
10 29660 / 21.0

Maybe a silly question but:
Is this to be read as fractions? That is, the multiplier at level 10 is 1412.38095238095....? Not 1400, 1412 or 1412.381 but the full ****? Now, that is what I call finetuning on a truly scientific scale Lol
Or does the number after the slash have a different meaning?
Sir HyperChrist
Persnickety Pilots
#22 - 2014-07-10 23:39:39 UTC  |  Edited by: Sir HyperChrist
Skalle Pande wrote:
Quote:
Research Multipliers
Level Multiplier
1 1
2 29 / 21.0
3 23 / 7.0
4 39 / 5.0
5 278 / 15.0
6 928 / 21.0
7 2200 / 21.0
8 5251 / 21.0
9 4163 / 7.0
10 29660 / 21.0

Maybe a silly question but:
Is this to be read as fractions? That is, the multiplier at level 10 is 1412.38095238095....? Not 1400, 1412 or 1412.381 but the full ****? Now, that is what I call finetuning on a truly scientific scale Lol
Or does the number after the slash have a different meaning?


That table got my attention too :)
I quess it's fractionals. Except for the second all ratios (of one level and the level before it) are about 32^(1/4), It looks like a similar progression as skillpoints, except in 10 levels, not 5. Iirc something like that was mentioned in the devblogs.

I could calculate new "earnback times" for Material research costs, as each level saves 1% extra, the first level "costs" 2% of an item's base price, it takes 2 items to earn the investment back. The total list is:
lvl1: 2
lvl2: 3
lvl3: 7
lvl4: 16
lvl5: 37
lvl6: 88
lvl7: 210
lvl8: 500
lvl9: 1189
lvl10: 2825

Components are usually built in 1000+ quantities, so that research pays back. Not so sure about battleships though :)

Could someone update the eve wiki on numbers like these?

edit: "savings" are slightly simplified: lvl 1 saves 0.01 on 1.10, which is slightly less than 1% :)
(even this depends on the definition of base item costs: is that assuming zero waste? :)
asteroidjas
Rothschild's Sewage and Septic Sucking Services
The Possum Lodge
#23 - 2014-07-11 01:28:15 UTC
With this new reliance on the very inaccurate UI "Estimated Value" number, can you please, please, please use better methods to determine this figure.

There is already data that is used to make a much more accurate picture of the price of an item. The daily average tickers. Calculated EACH DAY. Just grab that from the major trade hubs and average them together. Again, this figure is already calculated in every region, on a daily basis. How is it that hard to use it for the "estimated value"?

Atm, the "Estimated Value" of an item will be very missleading as to whether or not something is worth building. This will also help alleviate 'value' issues for new items that sometimes go a couple weeks before the game says they are worth anything.
Skalle Pande
Teknisk Forlag
#24 - 2014-07-11 10:18:31 UTC
Sir HyperChrist wrote:

That table got my attention too :)
I quess it's fractionals. Except for the second all ratios (of one level and the level before it) are about 32^(1/4), It looks like a similar progression as skillpoints, except in 10 levels, not 5. Iirc something like that was mentioned in the devblogs.

Yeah, it looks likely. But heck, if this is meant to be an exponential function, why then calculate it as fractions? And anyway, why calculate it at all, instead of using a fixed figure? In fact an integer would surely suffice, as the multiplier is applied to something (baseCost) which will vary with at least some percentage points on a monthly timescale.
CCP Nullarbor
C C P
C C P Alliance
#25 - 2014-07-11 10:24:27 UTC
asteroidjas wrote:
With this new reliance on the very inaccurate UI "Estimated Value" number, can you please, please, please use better methods to determine this figure.

There is already data that is used to make a much more accurate picture of the price of an item. The daily average tickers. Calculated EACH DAY. Just grab that from the major trade hubs and average them together. Again, this figure is already calculated in every region, on a daily basis. How is it that hard to use it for the "estimated value"?

Atm, the "Estimated Value" of an item will be very missleading as to whether or not something is worth building. This will also help alleviate 'value' issues for new items that sometimes go a couple weeks before the game says they are worth anything.


You should be using the adjustedPrice from the /market/prices/ endpoint which is both not as volatile as the average price shown in the inventory UI and less manipulatable by payers.

It is also really important to realize that having accurate values for the base price does not actually matter in the grand scheme of things, we only need this value to be correct within an order of magnitude and for it to rise and fall gradually with shifting markets.

CCP Nullarbor // Senior Engineer // Team Game of Drones
CCP Nullarbor
C C P
C C P Alliance
#26 - 2014-07-11 10:28:58 UTC  |  Edited by: CCP Nullarbor
Skalle Pande wrote:
Sir HyperChrist wrote:

That table got my attention too :)
I quess it's fractionals. Except for the second all ratios (of one level and the level before it) are about 32^(1/4), It looks like a similar progression as skillpoints, except in 10 levels, not 5. Iirc something like that was mentioned in the devblogs.

Yeah, it looks likely. But heck, if this is meant to be an exponential function, why then calculate it as fractions? And anyway, why calculate it at all, instead of using a fixed figure? In fact an integer would surely suffice, as the multiplier is applied to something (baseCost) which will vary with at least some percentage points on a monthly timescale.


Yes they are fractions, and the values are based on a curve which fits to nice intervals for the amount of time required to research up to each level, similar to skill training:

Rank 1 Item Seconds

  1. 105
  2. 250
  3. 595
  4. 1414
  5. 3360
  6. 8000
  7. 19000
  8. 45255
  9. 107700
  10. 256000


Rank 4 Item Seconds

  1. 420
  2. 1000
  3. 2380
  4. 5656
  5. 13440
  6. 32000
  7. 76000
  8. 181020
  9. 430800
  10. 1024000

CCP Nullarbor // Senior Engineer // Team Game of Drones
Sir HyperChrist
Persnickety Pilots
#27 - 2014-07-11 10:44:53 UTC  |  Edited by: Sir HyperChrist
CCP Nullarbor wrote:
Skalle Pande wrote:
Sir HyperChrist wrote:

That table got my attention too :)
I quess it's fractionals. Except for the second all ratios (of one level and the level before it) are about 32^(1/4), It looks like a similar progression as skillpoints, except in 10 levels, not 5. Iirc something like that was mentioned in the devblogs.

Yeah, it looks likely. But heck, if this is meant to be an exponential function, why then calculate it as fractions? And anyway, why calculate it at all, instead of using a fixed figure? In fact an integer would surely suffice, as the multiplier is applied to something (baseCost) which will vary with at least some percentage points on a monthly timescale.


Yes they are fractions, and the values are based on a curve which fits to nice intervals for the amount of time required to research up to each level, similar to skill training:

Rank 1 Item Seconds

  1. 105
  2. 250
  3. 595
  4. 1414
  5. 3360
  6. 8000
  7. 19000
  8. 45255
  9. 107700
  10. 256000

...


And those numbers are based on arbitrarily rounded results of the function: 250*32^(n/4), with n from -1 to +8.

Why you are rounding them though, I don't understand. It doesn't make it any more transparent, easier to understand. Also the rounding is really arbitrarily: 45255 is actually rounded best, 19000 differs quite a bit from 19027. Why don't you fix this list to actually have some base?

-SHC

-edit, might it be that current devs don't understand the basis of the original skillpoint numbers anymore? What?
Aineko Macx
#28 - 2014-07-12 09:56:47 UTC
Steve Ronuken wrote:
SDE conversions

Thx Steve, again you bring light and order into this mess.

Before I begin any development on it, are you somewhat sure you'll keep the proposed scheme for the industry tables?

Some notes:
- Column 'typeid' in industryActivitySkills should be 'typeID' for consistency
- If you have composite indices over (typeID, activityTypeID), another simple index over typeID is redundant. The only use I could see is for implementing foreign keys, which is pointless for this static data.

iveeCore 3.0: The PHP engine for industrial activities and CREST library
Steve Ronuken
Fuzzwork Enterprises
Vote Steve Ronuken for CSM
#29 - 2014-07-12 16:49:21 UTC
Aineko Macx wrote:
Steve Ronuken wrote:
SDE conversions

Thx Steve, again you bring light and order into this mess.

Before I begin any development on it, are you somewhat sure you'll keep the proposed scheme for the industry tables?

Some notes:
- Column 'typeid' in industryActivitySkills should be 'typeID' for consistency
- If you have composite indices over (typeID, activityTypeID), another simple index over typeID is redundant. The only use I could see is for implementing foreign keys, which is pointless for this static data.



I've had people talking to me about this.

There will almost certainly be tweaks, but only to column names. The other structure should be static.

activityTypeID will become activityID
typeID will become blueprintTypeID (to make it obvious.

All names will become properly camel case (and ID will remain ID, rather than Id)

The single column index isn't entirely redundant. But you make a reasonable point.

Woo! CSM XI!

Fuzzwork Enterprises

Twitter: @fuzzysteve on Twitter
Zifrian
Federal Defense Union
Gallente Federation
#30 - 2014-07-13 23:02:38 UTC  |  Edited by: Zifrian
After getting the new blueprint data loaded, I noticed there is no longer a techLevel field for blueprints (I can get all other fields of the old invBlueprintTypes table). This value usually signaled 1,2, or 3. I've been looking at the dgmtypeAttributes table (attribute 633) for the meta value but this really isn't the same information since it's the meta level of the item but not the blueprint. It's also not easy to determine tech 3 blueprints unless I hack it to look for the 4 ship types.

Any way to get this data in the current SDE or is this an oversight?

Thanks

Maximze your Industry Potential! - Download EVE Isk per Hour!

Import CCP's SDE - EVE SDE Database Builder
Apelacja
Sad Najwyzszy
#31 - 2014-07-14 14:20:58 UTC
so what is a cost to resesach something to me 9?

Rly 20 times base cost :?
CCP Nullarbor
C C P
C C P Alliance
#32 - 2014-07-14 14:29:41 UTC
Apelacja wrote:
so what is a cost to resesach something to me 9?

Rly 20 times base cost :?


Research in a system costs the same amount per second of time, so the longer it takes to research that level the more it costs. However keep in mind the base level is the value of the input materials to build the thing x 2% x the system cost index which will be somewhere between 0 - 14% we believe.

CCP Nullarbor // Senior Engineer // Team Game of Drones
CCP Nullarbor
C C P
C C P Alliance
#33 - 2014-07-14 14:33:48 UTC
Zifrian wrote:
After getting the new blueprint data loaded, I noticed there is no longer a techLevel field for blueprints (I can get all other fields of the old invBlueprintTypes table). This value usually signaled 1,2, or 3. I've been looking at the dgmtypeAttributes table (attribute 633) for the meta value but this really isn't the same information since it's the meta level of the item but not the blueprint. It's also not easy to determine tech 3 blueprints unless I hack it to look for the 4 ship types.

Any way to get this data in the current SDE or is this an oversight?

Thanks


I doubt we will put it in the blueprints yaml, I will talk to greyscale about whether we should set that data on dogma for the blueprints.

We don't actually use techLevel on blueprints for anything though, which is why it was removed.

CCP Nullarbor // Senior Engineer // Team Game of Drones
Querns
Science and Trade Institute
Caldari State
#34 - 2014-07-14 15:13:55 UTC  |  Edited by: Querns
e: it's back online
Looks like sisi crest is down. Sad

{
message: "Not accepting connections",
key: "NotAcceptingConnection",
exceptionType: "ServiceUnavailableError"
}

Is there any chance of this coming back up before Crius hits?

This post was crafted by the wormhole expert of the Goonswarm Economic Warfare Cabal, the foremost authority on Eve: Online economics and gameplay.
Zifrian
Federal Defense Union
Gallente Federation
#35 - 2014-07-14 16:52:08 UTC
CCP Nullarbor wrote:
Zifrian wrote:
After getting the new blueprint data loaded, I noticed there is no longer a techLevel field for blueprints (I can get all other fields of the old invBlueprintTypes table). This value usually signaled 1,2, or 3. I've been looking at the dgmtypeAttributes table (attribute 633) for the meta value but this really isn't the same information since it's the meta level of the item but not the blueprint. It's also not easy to determine tech 3 blueprints unless I hack it to look for the 4 ship types.

Any way to get this data in the current SDE or is this an oversight?

Thanks


I doubt we will put it in the blueprints yaml, I will talk to greyscale about whether we should set that data on dogma for the blueprints.

We don't actually use techLevel on blueprints for anything though, which is why it was removed.

I figured as much. I'll find a work around for my purposes.

Thanks

Maximze your Industry Potential! - Download EVE Isk per Hour!

Import CCP's SDE - EVE SDE Database Builder
TheSmokingHertog
Julia's Interstellar Trade Emperium
#36 - 2014-07-15 03:22:54 UTC
Zifrian wrote:
CCP Nullarbor wrote:
Zifrian wrote:
After getting the new blueprint data loaded, I noticed there is no longer a techLevel field for blueprints (I can get all other fields of the old invBlueprintTypes table). This value usually signaled 1,2, or 3. I've been looking at the dgmtypeAttributes table (attribute 633) for the meta value but this really isn't the same information since it's the meta level of the item but not the blueprint. It's also not easy to determine tech 3 blueprints unless I hack it to look for the 4 ship types.

Any way to get this data in the current SDE or is this an oversight?

Thanks


I doubt we will put it in the blueprints yaml, I will talk to greyscale about whether we should set that data on dogma for the blueprints.

We don't actually use techLevel on blueprints for anything though, which is why it was removed.

I figured as much. I'll find a work around for my purposes.

Thanks


Cant you do a merge from a old labeled dump with the new one, just compare the ItemID of the bleuprints in question between the 2 sources. Then cache this list in a table and use it from then on...

"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 =-
Zifrian
Federal Defense Union
Gallente Federation
#37 - 2014-07-16 12:06:14 UTC  |  Edited by: Zifrian
TheSmokingHertog wrote:
Zifrian wrote:
CCP Nullarbor wrote:
Zifrian wrote:
After getting the new blueprint data loaded, I noticed there is no longer a techLevel field for blueprints (I can get all other fields of the old invBlueprintTypes table). This value usually signaled 1,2, or 3. I've been looking at the dgmtypeAttributes table (attribute 633) for the meta value but this really isn't the same information since it's the meta level of the item but not the blueprint. It's also not easy to determine tech 3 blueprints unless I hack it to look for the 4 ship types.

Any way to get this data in the current SDE or is this an oversight?

Thanks


I doubt we will put it in the blueprints yaml, I will talk to greyscale about whether we should set that data on dogma for the blueprints.

We don't actually use techLevel on blueprints for anything though, which is why it was removed.

I figured as much. I'll find a work around for my purposes.

Thanks


Cant you do a merge from a old labeled dump with the new one, just compare the ItemID of the bleuprints in question between the 2 sources. Then cache this list in a table and use it from then on...

But when new bps are added it's out of date. I'd also rather have something that is repeatable with logic each sde instead of special cases. The meta data works fine for what I need now. The only issue is that T3 item meta data isn't consistent and marked as meta 1 or 4, which neither make any sense to me. So I'll just add in a special case for these for now.

Maximze your Industry Potential! - Download EVE Isk per Hour!

Import CCP's SDE - EVE SDE Database Builder
Cryten Jones
Advantage Inc
#38 - 2014-07-16 12:38:10 UTC
CCP Nullarbor wrote:
Zifrian wrote:
After getting the new blueprint data loaded, I noticed there is no longer a techLevel field for blueprints (I can get all other fields of the old invBlueprintTypes table). This value usually signaled 1,2, or 3. I've been looking at the dgmtypeAttributes table (attribute 633) for the meta value but this really isn't the same information since it's the meta level of the item but not the blueprint. It's also not easy to determine tech 3 blueprints unless I hack it to look for the 4 ship types.

Any way to get this data in the current SDE or is this an oversight?

Thanks


I doubt we will put it in the blueprints yaml, I will talk to greyscale about whether we should set that data on dogma for the blueprints.

We don't actually use techLevel on blueprints for anything though, which is why it was removed.




I use this as part of a system to work out what can be invented from something. Eg add a new T1 BPO t the system and have it ask if you want to add any inventable types at the same time.....

Any chance of having some form if link from a T2 type back to its T1 version?

My preference would be to have a T2 print have a value for it's parent print...

Thanks
Hirogenale
The Scope
Gallente Federation
#39 - 2014-07-16 13:45:39 UTC
Cryten Jones wrote:
CCP Nullarbor wrote:
Zifrian wrote:
After getting the new blueprint data loaded, I noticed there is no longer a techLevel field for blueprints (I can get all other fields of the old invBlueprintTypes table). This value usually signaled 1,2, or 3. I've been looking at the dgmtypeAttributes table (attribute 633) for the meta value but this really isn't the same information since it's the meta level of the item but not the blueprint. It's also not easy to determine tech 3 blueprints unless I hack it to look for the 4 ship types.

Any way to get this data in the current SDE or is this an oversight?

Thanks


I doubt we will put it in the blueprints yaml, I will talk to greyscale about whether we should set that data on dogma for the blueprints.

We don't actually use techLevel on blueprints for anything though, which is why it was removed.




I use this as part of a system to work out what can be invented from something. Eg add a new T1 BPO t the system and have it ask if you want to add any inventable types at the same time.....

Any chance of having some form if link from a T2 type back to its T1 version?

My preference would be to have a T2 print have a value for it's parent print...

Thanks


You could simply do a join from invBlueprintTypes and invTypes on productTypeID=typeID to get the groupID and then a join from the result of that with invGroups.
The CategoryID in there then gives you the Information if its T1, T2, T3, Faction or whatever.

You could find out the parent BPO with the invMetaTypes table, search for the parentTypeID of the item your BPO produces and then look up the BPO that produces this parent Item

isn't exactly the most intuitive way to look for stuff like that but its possible... xD
CCP Nullarbor
C C P
C C P Alliance
#40 - 2014-07-16 22:52:52 UTC
Cryten Jones wrote:
CCP Nullarbor wrote:
Zifrian wrote:
After getting the new blueprint data loaded, I noticed there is no longer a techLevel field for blueprints (I can get all other fields of the old invBlueprintTypes table). This value usually signaled 1,2, or 3. I've been looking at the dgmtypeAttributes table (attribute 633) for the meta value but this really isn't the same information since it's the meta level of the item but not the blueprint. It's also not easy to determine tech 3 blueprints unless I hack it to look for the 4 ship types.

Any way to get this data in the current SDE or is this an oversight?

Thanks


I doubt we will put it in the blueprints yaml, I will talk to greyscale about whether we should set that data on dogma for the blueprints.

We don't actually use techLevel on blueprints for anything though, which is why it was removed.




I use this as part of a system to work out what can be invented from something. Eg add a new T1 BPO t the system and have it ask if you want to add any inventable types at the same time.....

Any chance of having some form if link from a T2 type back to its T1 version?

My preference would be to have a T2 print have a value for it's parent print...

Thanks


This information is available in the blueprints.yaml by looking through the "products:" list for each of the invention activities.

A side effect of this is it sets us up for a possibly conceivable future where you could invent something from multiple different blueprint types.

CCP Nullarbor // Senior Engineer // Team Game of Drones