EVE Information Portal

Invention skills are so useless it's generally considered a waste to train them to IV. That's how little they matter with the current formula.

In the case of the capital ship pre-reqs being changed it was done in a way that no one really had to train any more than they already had to keep the same level. In the case of drones skills were bought up to match other skills that you may have trained.

The existing skills do lose value though, in the case of Sin used as an example in the blog, someone who really only builds ships would have mechanical engineering which is now useless to them and wouldn't have high energy physics which before they wouldn't of trained.

So the same treatment of skills towards ships will not be seen towards the industry? c/d

Pointless? On the contrary, the price will of course be supported by the market. If there is demand the price will fluctuate accordingly. The point is these rare decryptors and data cores would offer significant bonuses that will of course need to be balanced with cost.

Come on man! Think of the possibilities to both offer explorers and builders/inventors. Win/win/win.

If they are going to add a feature that enables different levels of success, why not have an extremely rare chance to "Invent" a T2 BPO.

Then there would be no need to address the existing T2 BPO's in the game. It would be possible, although extremely rare for anyone to get one. The more invention you do the higher your chances of getting an ellusive BPO.

/Que the Flames

Edit;
I know how bad it would be to introduce more T2 BPO's. With time the need for invention would completely dissapear as T2 BPO's would eventually cover all the production needs.

However, As someone else earlyer in the thread mentioned. it really does not make sense for items to need to be "invented" thousands of times. Eventually those rare invented items should become common place enough to not need to be invented any longer. i.e. enough T2 BPO's around that they no longer need invention.

Such a system would allow for newly introduced T2 items to gradually become common items, with T2 BPO's for production. Eventually invention would only be useful to prevent the prices of the T2 BPO's from getting too high, as there would always be an alternative to using them, and a way to aquire additional T2 BPO's if market demand goes up.

Good riddance meta items in invention.

Saving Datacores seems fine, but I don't feel +ME/+TE is an elegant design.

+Me: Sure I will save some components, but I build in batches and will always plan using the basic ME2/TE4 bpc.
+Te: Having a few TE5 bpcs finishing a few minutes earlier in a batch of TE4 will change nothing, I will just wait for my whole 10 lines to finish before setting up another batch.

Also science skills need some love, I understand you don't want them to be the new material efficiency skill for T2 manufacturing, but still...maxing them to V for such a minimal gain defies logic.

Want to know how to spreadsheet these numbers? Work with averages.

https://docs.google.com/spreadsheets/d/1ikRteYe8iSCau-H2Y0J5vO0Dl1RTrfZfd_tFQcrCXD8/edit?usp=sharing

It has been said repeatedly in this thread that the science skills are not worth enough. It has also been said that we don't want these skills to be "level 5 or go home" like the old Production Efficiency skill was.

So here is some math to show how much effect these skills really have:

Imagine a module for which the T1 version costs 50k isk and T2 version costs 1M isk. It takes 6 datacores to invent (3 of each type) and returns a 10 run blueprint on success. I am doing this in a system with copying and invention cost indices of 3%.

The installation cost to make a T1 BPC = 50000*0.02*0.03 = 30 isk.
The installation cost to attempt each invention = 1000000*0.02*0.03 = 600 isk.

Datacores have cost between 80k isk and 150k isk each for quite a while now. Assume that after the market shakeup prices settle to about the same place. (FW runners choose different datacores to supply and things even out.) Let's use 120k isk for our datacore price.

The new system will return 38.5% of datacores per failure. Those returned datacores can be sold off or used for the next batch, so we don't count them as part of the cost of invention. The isk cost of datacores for one invention attempt = 120000*6*(1-Percentage of Returned Datacores)

The cost of each invention attempt is the installation cost of the copy + the installation cost of the invention + the datacore cost.

The Cost per Successful invention = Cost Per Attempt / Success Chance

The Cost Per Unit = Cost per Success / Runs per Success

With the current success formula, a character with level 3 in Science 1, Science 2 and Racial Encryption Methods has a 46% chance to invent a module. Someone with level 4 in those skills has a 48% chance. Someone with level 5 in those skills has 50% chance (Not 50.4% as was mentioned in the dev blog.) Calculate cost per module out for 46%, 48%, and 50% Success Chances and you will get:
................Success Chance...................46%...................48%..................50%
...........Returned Datacores..............20.79%..............20.02%.............19.25%
Datacore Cost per Attempt........570,312.00.........575,856.00.......581,400.00
.......Total Cost per Attempt........570,942.00.........576,486.00.......582,030.00
...............Cost Per Success.....1,241,178.26.....1,201,012.50.....1,164,060.00
......................Cost Per Unit........124,117.83........120,101.25........116,406.00

So, a character can train three rank 5 skills (two science skills and one encryption skill) from level 0 to 3 in about three days. Their T2 BPC will cost them 124k isk of the 1M isk finished T2 module.
That character can then train those three skills to level 4, and it will take them about 12 days. At that point, their T2 BPC will cost them 120k isk. That 4016 isk savings is 0.4% more profit over the character with rank 3 skills.
That character can then train those three skills to level 5. It will take about 60 days. They save 3695 isk per unit over the character with level 4 skills. 0.37%

This item was chosen to illustrate the maximum benefit one can get from the skills. For most other items the manufacturing cost is higher and the blueprint cost is a much smaller portion of the finished T2 item cost.

Yes, you heard that right! For the low, low cost of two months training you too can earn less than half of a percent more profit on your inventions!

About 0.2% of all materials.

But really a lot less than that because with limited runs and rounding we will miss out on most of the savings.

For me, a couple exceptional successes will just mean that I need to buy fewer materials for next week's run.

I'm not going to lie. A lot of this blew over my head. The real question I have, does CCP do this type of analysis on things that they change/ want to change and if not why don't they?

Here is my proposal for a new invention success formula:

Success Rate = Base Rate * (1+ (Science Skill 1 + Science Skill 2 + Encryption Skill)/15)
Where Base Rate for Modules, Ammo, Drones = 26%
Base Rate for Frigates = 23%
Base Rate for Cruisers = 20%
Base Rate for Battleships 17%
Base Rate for Freighters = 14%

............Current Formula.....................................................Proposed Formula
Base.....Skills 3.....Skills 4.....Skills 5.........................Base.....Skills 3.....Skills 4.....Skills 5
40%..........46%.........48%..........50%........................26%........41.6%......46.8%.........52%
35%.....40.25%.........42%.....43.75%........................23%........36.8%......41.4%.........46%
30%.......34.5%.........36%.......27.5%........................20%...........32%.........36%.........40%
25%.....28.75%.........30%.....31.25%........................17%........27.2%......30.6%.........34%
20%..........23%.........24%..........25%........................14%........22.4%......25.2%.........28%

My new numbers were chosen around characters with level 4 science and encryption skills. Those characters would have approximately the same success rates in both formulas, but those with lesser skills have fewer successes and those with greater skills have more successes.

Plugging the proposed success rates into the same hypothetical module in my previous post:
................Success Chance...................41.6%...................46.8%......................52%
...........Returned Datacores...............22.484%...............20.482%.................18.48%
Datacore Cost per Attempt............558,115.20...........572,529.60...........586,944.00
.......Total Cost per Attempt...........558,745.20...........573,159.60............587,574.00
...............Cost Per Success........1,343,137.50........1,224,700.00........1,129,950.00
......................Cost Per Unit...........134,313.75...........122,470.00...........112,995.00

So the 12 days of training to get skills from 3 to 4 will yeild 1.18% more profit on this T2 item. The 60 days training to get skills from 4 to 5 yields 0.947% more profit. 1% is comparable to moving from a 5% system to a 4% manufacturing system, and less than is saved by using most teams. Much better return than previously, but certainly not required.