These forums have been archived and are now read-only.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: Researching, the Future

First post First post First post
Author
Previous Topic Next Topic
LHA Tarawa
Pator Tech School
Minmatar Republic
#761 - 2014-05-09 16:40:37 UTC  |  Edited by: LHA Tarawa
So, if waste = 10% * 1 / (1+ME) was too complex... could they have resolved the complexity by
1) Renaming it to waste_divider,
2) Starting it at 1 instead of 0.

That would have simplified it to waste = 10% / Waste_Divider.
Weaselior
GoonWaffe
Goonswarm Federation
#762 - 2014-05-09 16:55:02 UTC  |  Edited by: Weaselior
LHA Tarawa wrote:

I think it is bad that mega corps will get a major advantage over little corps because they can crank out capital ships in batches, while I think that is a bad thing.... so I'm a liar. Got it.


that's intended. what is a lie is your professed newbie-concern and your attempting to claim you're opposing these changes for any reason other than they might hurt you, personally, in a slight amount: you (a capital producer) are upset because you may get outcompeted. your arguments besides pure-self interest are trumped up.


LHA Tarawa wrote:

Here again, we're clearly not speaking the same language. I consider taking away a competitive advantage that people had been using to increase profits is screwing them.

On the other hand, you continue to dismiss the massive time increase to sub-1% waste as "just a scaling thing", but CCP still has not offered any indication they intend to alter the scaling. I do not see how they could alter the scaling to make 0% waste reasonable, without making 1% or less too quick.

What you dismiss as a "scaling thing" is really the result of removing partial % research.


i have no patience for people whining about a competitive advantage that was not real. the scaling thing is not related to atomization: it is related to the choice to make each level take more time than the next, and the time increase between levels, and that this was not scaled to current research times. it probably should have been.


LHA Tarawa wrote:

I could not disagree more. Research only needed to be done once per BPO, whereas manufacturing is done repeatedly, and BPO research it was the same for everyone, regardless the size of corp/alliance they were in. Now the rounding at the job level will favor those in the largest corporations that can run capital ships in large batches rather than one at a time.

I respect your opinion, but disagree that this is "good" complexity.


interpreting ME was a thing that needed to be done when buying bpos. all complexity came from the fact that the figure you were given didn't mean what any normal person would think it means. now, complexity comes into play in such things as "am I better off putting in a long build job, tying up capital for a long time for higher profits per unit, or am I better off doing short runs and churning capital repeatedly". you have demonstrated your inability to compete by not realizing this tradeoff: you think this is the same as the old method where you just find the right answer and do that. it's not. as a rich and powerful industrialist i may choose to crush you through long jobs - or, I may choose to crush you through cycling capital through short jobs and saturating the market and making a smaller profit but making it more often so it compounds

LHA Tarawa wrote:

Because of the scaling,
In the new system the difference between not being able to do research between 10% and 5% compared to not being able to do fraction of a % between 1% and 0% is a ration of 1.5 hours to 4 2/3rds days (that is for items with multiplier of 1. expanding that ratio to capital it is the difference between a month, and 7 years).

For the old system, it was similar ration... for something like a battleship, the jump for 10% to 5% was 4 days while the jump from 1% to 0% was years.

I think it is far more important to be able to do partial research of steps that take days (or years) than hours (or days).

Then again, I have to remember were speaking different languages, so perhaps in your languages 2 orders of magnitude difference in "trumped up nonsense". Maybe it is just as important to be able to do partial research for things that take days, as it is to be able to do partial research that takes decades.


in the old system the difference between ME1 and ME0 was huge for a titan bpo but for a time it didn't make sense because at the time bpcs were so valuble it would take a long time to pay off keeping it idle

now, i could research it quickly to intermediate values between 0 and 5%. this is not a real concern and is merely something you're trumping up. the problem - that it scales too hard on the high end - is fixed by fixing that. it never should have been a "years" jump from 1% to 0% for t1 ships, that was stupid and it's great they're going away. i have no perticular opposition to a research point system where intermediate research can be done in chunks so it's not lost if you want to put your bpo in research for a month then use it before the next level is done, just like skills.

LHA Tarawa wrote:

Here, we must be speaking different languages again. In your language, "dumb" must mean the same as "cogent" means in my language.


this is basically a dunning-kruger problem: you don't have the knowledge to know how little you know. or put another way: when you've been posting bad opinions nonstop, why would we trust your opinion on if they're dumb? if you knew they were dumb you wouldn't have posted them and if they were good they'd stand on their own merit instead of you needing to repeatedly assure everyone no actually they are smart

Head of the Goonswarm Economic Warfare Cabal Pubbie Management and Exploitation Division.
Weaselior
GoonWaffe
Goonswarm Federation
#763 - 2014-05-09 17:05:20 UTC
LHA Tarawa wrote:
So, if waste = 10% * 1 / (1+ME) was too complex... could they have resolved the complexity by
1) Renaming it to waste_divider,
2) Starting it at 1 instead of 0.

That would have simplified it to waste = 10% / Waste_Divider.

there is literally no merit to this suggestion whatsoever, i don't know how many times I can say that you simply do not understand the issue

the issue is that the UI should give you meaningful information. giving you ME and you needing to divide 10% by ME, then compare that waste to the waste of the alternative ME avalible, is horrible UI. it's crap, it has no redeeming features whatsoever

tweaking the exact formula you need to use to make numbers meaningful is not helpful in the slightest. you really do not understand the problem at all, and that's why your solutions are such garbage. you have glommed onto that this may hurt you in a tiny way, so because your isk/hour depends on you not understanding the problem you refuse to understand it

Head of the Goonswarm Economic Warfare Cabal Pubbie Management and Exploitation Division.
LHA Tarawa
Pator Tech School
Minmatar Republic
#764 - 2014-05-09 17:05:34 UTC
Weaselior wrote:
LHA Tarawa wrote:

I think it is bad that mega corps will get a major advantage over little corps because they can crank out capital ships in batches, while I think that is a bad thing.... so I'm a liar. Got it.


that's intended. what is a lie is your professed newbie-concern and your attempting to claim you're opposing these changes for any reason other than they might hurt you, personally, in a slight amount: you (a capital producer) are upset because you may get outcompeted. your arguments besides pure-self interest are trumped up.


You continue to be dismissive, and speaking a different language.

My concerns cover all players.
Existing players that have put years into research only to have other players "catch up".
Newer players that will have much longer research times.
Small corporations that do not run carriers in batches.
All manufacturers that have the complexity moved from 1-time research to every time they start a manufacturing job.


My main concern is the general EPIC FAIL of their attempt to remove complexity because they failed to understand the real source of the complexity. They attacked something that wasn't a problem, failed to address the real problem, then put in a hack that creates new problems.

As for harming me, the last capital I built was my rorqual... Oh, take it back, my freighter, because last time we moved I could jump the rorq, but just sold the freighter and built a new one.

I'm certainly not a player in the capital ship market. I don't even own any capital ship BPOs. I do have a full set of ammo, rig, crystal, T2 component and such. I play the low end of the market.

And yes, I usually batch up rig and ammo runs, so I won't personally be hit by the changes.

But that doesn't mean I can't see the major flaws in the changes they are hacking together to attack something that wasn't a problem in the first place, and certainly have better options, even if some people can't do very simple elementary school math.
LHA Tarawa
Pator Tech School
Minmatar Republic
#765 - 2014-05-09 17:27:38 UTC  |  Edited by: LHA Tarawa
Weaselior wrote:
LHA Tarawa wrote:
So, if waste = 10% * 1 / (1+ME) was too complex... could they have resolved the complexity by
1) Renaming it to waste_divider,
2) Starting it at 1 instead of 0.

That would have simplified it to waste = 10% / Waste_Divider.

there is literally no merit to this suggestion whatsoever, i don't know how many times I can say that you simply do not understand the issue

the issue is that the UI should give you meaningful information. giving you ME and you needing to divide 10% by ME, then compare that waste to the waste of the alternative ME avalible, is horrible UI. it's crap, it has no redeeming features whatsoever

tweaking the exact formula you need to use to make numbers meaningful is not helpful in the slightest. you really do not understand the problem at all, and that's why your solutions are such garbage. you have glommed onto that this may hurt you in a tiny way, so because your isk/hour depends on you not understanding the problem you refuse to understand it


And you continue to be dismissive of all other options that could have had the UI present meaningful information.

The UI could EASILY have shown "Waste Base, Waste Divider, and Waste Effective = with Waste Effective simply Waste Base / Waste Divider.

What I see as having no redeeming features is the proposed redesign.


What we have reduced this to is a simple difference of subjective opinion:

My opinion:

The complexity of the old system involved applying the % to the actual input needs to determine rounding point.

If there was confusion over waste = .1 * 1/(1+ME), it was minor and there were many ways to remove that confusion without radical overhaul.

The redesign is bad because it takes away the advantage of people who worked hard to research BPOs to a high level.

I think the redesign is bad because it takies away the ability to research fractions of waste between 1% and 0%. That is too long of a jump to not be able to do in pieces, and receive benefit from those pieces.

Moving complexity from research that is done once per BPO to manufacturing is bad, because I don't see this as "good, meaningful" complexity.

Requiring billion ISK ships be produced in batches to take advantage of material reductions is too big of an advantage for mega corps over little corps.



Your opinion seems to be:

All the complexity of the old system was from waste = 10% * 1/(1+ME) and any attempt to simplify that, such as waste = 10% / waste_divider would be unacceptable.

Screwing people that spent time, ISK and effort researching BPOs above ME 10 is good.

CCP will do something to fix the long research times in the new plan.

The complexity that was bad in research, is good in manufacturing, and was intended change, even though that change was not part of the original change and was hacked in after people like me pointed out that they were not dealing with the actual complexity in research, which was combining the waste % with items needed and then identifying inflection in the resulting rounding.

Capital ship construction should only be done by mega corps that can crank out billion ISK ships in batches, not by small industrial corps that would do manufacturing of capital ships one at a time.
Weaselior
GoonWaffe
Goonswarm Federation
#766 - 2014-05-09 17:40:22 UTC
LHA Tarawa wrote:

You continue to be dismissive, and speaking a different language.

My concerns cover all players.
Existing players that have put years into research only to have other players "catch up".
Newer players that will have much longer research times.
Small corporations that do not run carriers in batches.
All manufacturers that have the complexity moved from 1-time research to every time they start a manufacturing job.

i continue to be dismissive because you merely spam the same argument over and over again rather than understand and respond to criticism

your concerns do not cover all players. they cover only yourself. i have repeatedly torpedoed every attempt to generalize your concerns outside your narrow self-interested attempt to protect your personal industry. the one legitimate concern you raise i have spoken at length about and it does not at its core relate to the change in system but perticular decisions in its implementation that can be altered

your dumb argument about carriers in batches i addressed, you did not understand the response so you've just repeated your point. a dismissive noise is all this deserves.

and lastly, you continue to be unable to understand good complexity vs bad complexity

Head of the Goonswarm Economic Warfare Cabal Pubbie Management and Exploitation Division.
Weaselior
GoonWaffe
Goonswarm Federation
#767 - 2014-05-09 17:44:09 UTC
LHA Tarawa wrote:

The UI could EASILY have shown "Waste Base, Waste Divider, and Waste Effective = with Waste Effective simply Waste Base / Waste Divider.


still terrible ui with no redeeming features


LHA Tarawa wrote:

What we have reduced this to is a simple difference of subjective opinion:

no. our differences are rooted in that your points are objectively and demonstrably wrong. that you're resorting to saying it's your opinion and just repeating it endlessly proves that. your inability to parse and respond to criticism proves that. all of the things you say are "your opinion" are wrong, I have explained why they are wrong, and you have not even attempted to rebut the reasons why they are wrong.

needless to say, given your inability to even attempt to rebut my arguments, your attempts to explain what my "opinion" (in actuality, my description of objective reality) is are nonsense not worth even reading (and I literally didn't read them). my points stand on their own and are unassalibly correct. it is, of course, ironic you complained about strawmanning and your sole response now is trying to "rephrase" my posts

Head of the Goonswarm Economic Warfare Cabal Pubbie Management and Exploitation Division.
Quintessen
The Scope
Gallente Federation
#768 - 2014-05-09 17:54:12 UTC
@LHA Tarawa

I, and many, get that you feel that rounding is the source of a lot of complexity, but rounding is something that myself, and a lot of others are perfectly fine with -- something generally learned in primary and secondary school.

Since this is not the only factor in rounding, rounding is generally going to be required. It's already baked into the system and has been for some time. It's not really reasonable for CCP to replace all instances of rounding because a small sub-set of people don't like the idea that they will have to round things off in their head.

What you're proposing with player-specified lengths of time is far more complex, I think, to the average person than simple rounding (as opposed to complex rounding like bankers rounding). The idea is that every user will know exactly what they're going to get out of it and what the cost is when they put the blueprint in for research. Because the percentage values will be there and will be cumulative the blueprint ME% is simply a factor in the final values. You might say that multiple factors are too complex, but that type of complexity is where actual decision making is taking place and therefore good complexity.

Though they've been addressed before, I will address your four points again.

1) Some people will gain more advantage than others. This happens and CCP has repeatedly state that they are free to do this and they do not compensate people for these kinds of changes. CCP must be allowed to rebalance and adjust their game for it to be long-term viable.

2) Longer research times are a problem for newer players and should be looked at.

3) Researching partial levels is something they might want to look at, but they may also want people to have to risk things for long periods of time as a sort of gameplay experience. Perhaps this is to get people to be invested in defending territory so that they don't lose progress in research.

4) With respect to the complexity of the current ME system versus the new ME system, we're simply going to have to agree to disagree. But much of the player base, myself included, along with CCP seem to think that it's not a problem. Continually stating that rounding is the issue isn't going to get you very far since you've already said it many times.
Quintessen
The Scope
Gallente Federation
#769 - 2014-05-09 18:27:49 UTC
LHA Tarawa wrote:

Moving complexity from research that is done once per BPO to manufacturing is bad, because I don't see this as "good, meaningful" complexity.


Except, CCP wants the complexity at manufacturing. They want you to consider location, teams and facilities in your calculations for both research and manufacturing so that your manufacturing reductions are coming from multiple sources. CCP has stated that this is good complexity because it forces you to make decisions and weigh pros and cons. I think any argument down this path isn't going to work.

LHA Tarawa wrote:

Requiring billion ISK ships be produced in batches to take advantage of material reductions is too big of an advantage for mega corps over little corps.


Perhaps CCP could increase quantities so that instead of 4 of an item, it takes 40. Multiply all instances faucets and sinks for a given item by 10. That's a simpler solution to the problem even though it looses some flavor value since even a carrier doesn't need 40 drone bays, but that would diminish the rounding issue you've described.

Finally, any response that asks them to throw out all the work they've done up to this point is folly. They're simply not going to take their big industry expansion and remove all the industry elements from it. And they wouldn't just need another month to add what you're asking for. It would basically be starting from scratch. I suggest trying to work within their system with smaller changes. The time for big changes will come again, but not right this second.
LHA Tarawa
Pator Tech School
Minmatar Republic
#770 - 2014-05-09 18:44:04 UTC
Quintessen wrote:
@LHA Tarawa
I, and many, get that you feel that rounding is the source of a lot of complexity, but rounding is something that myself, and a lot of others are perfectly fine with -- something generally learned in primary and secondary school.


I never suggested they get rid of rounding. Their stated goal was to remove complexity, and since the complexity is because of the rounding, they failed to achieve the goal. As for learned i secondary school, if they can multiply % and determine round, then waste = 10% / (1+ME) should have been no problem at all!

Quintessen wrote:

What you're proposing with player-specified lengths of time is far more complex, I think, to the average person than simple rounding (as opposed to complex rounding like bankers rounding).


That is only necessary because CCP has removed the ability to research fractions of a %, and the jump from 1% to 0% is insanely long. It is a hack to fix a problem that CCP created by their failed attempt to remove complexity.

Quintessen wrote:

Though they've been addressed before, I will address your four points again.

1) Some people will gain more advantage than others. This happens and CCP has repeatedly state that they are free to do this and they do not compensate people for these kinds of changes. CCP must be allowed to rebalance and adjust their game for it to be long-term viable.


If there were a valuable reason to do this, I would agree. In this case there is not.


Quintessen wrote:

2) Longer research times are a problem for newer players and should be looked at.

3) Researching partial levels is something they might want to look at, but they may also want people to have to risk things for long periods of time as a sort of gameplay experience. Perhaps this is to get people to be invested in defending territory so that they don't lose progress in research.


Requiring they go right back to something that we have now... or perhaps what we would have had if ME had been converted to "research points" and be like ME * 100.

Quintessen wrote:

4) With respect to the complexity of the current ME system versus the new ME system, we're simply going to have to agree to disagree. But much of the player base, myself included, along with CCP seem to think that it's not a problem. Continually stating that rounding is the issue isn't going to get you very far since you've already said it many times.


Again, rounding is not an issue. It is simply the source of the complexity. And CCPs stated goal was to remove complexity, and since they utterly misunderstood the source of the complexity, their goals have not been met.

I'm not saying they should have removed rounding. I'm saying they shouldn't have tried to remove the complexity that rounding causes, by attacking the wrong thing. If ME was really such a problem, when they should have looked for minimal impact ways of clearing up the confusion, like making it start at 1, or making it a "research points" with units small enough to allow research to whole (or for the long research at the upper end, fractions of whole %.
LHA Tarawa
Pator Tech School
Minmatar Republic
#771 - 2014-05-09 19:10:39 UTC  |  Edited by: LHA Tarawa
Quintessen wrote:

Perhaps CCP could increase quantities so that instead of 4 of an item, it takes 40. Multiply all instances faucets and sinks for a given item by 10. That's a simpler solution to the problem even though it looses some flavor value since even a carrier doesn't need 40 drone bays, but that would diminish the rounding issue you've described.

Finally, any response that asks them to throw out all the work they've done up to this point is folly. They're simply not going to take their big industry expansion and remove all the industry elements from it. And they wouldn't just need another month to add what you're asking for. It would basically be starting from scratch. I suggest trying to work within their system with smaller changes. The time for big changes will come again, but not right this second.


I thought of that as well, and for at the high end, like capital components where there is plenty of room to shrink the size as they increase teh number, this will work.

However, where I see it breaking is at the low end. No item has a size below .01. I believe that internally, the m3 is actually stored as an int 100 per m3, and only converted to decimal by the UI. This means the big jump of 10X for rigs would effectively increase rig volume an order of magnitude.

I guess that is not an issue, because running sub-capital rigs in batches of 10 for best round is not nearly as big of a concern as capitals.

I agree it is folly for this release to tell them to throw it out and start over. I've been there. Done an iteration review and had users point out my design was utter fail becuase I didn't understand the problem. No time or resources to do it right, no flexibility in the ship date. So, the team ends up thrashing hacks trying to fix it.

We have this high performance system. Hundrends of millions or billions of records. Tens of millions of transactions an hour. We get a story "Send out notifications of transactions to down stream systems so they can track activity level". Now, to hit our performance metrics, we do a TON of asynchronous processing. We don't want to lock one part of the code while a bottleneck part of the code is cranking, so everything gets queued up all over the place. So, we implemented the new notifications in the same asynch way. Queue up events and send out as we get around to processing them, rather than blocking other processing.

So, in iteration review, a month before we ship, I AGAIN stress the asynch inherent in the design. A down stream consumer, who had hears "asynch processing" like 5 times in prior iteration, including full diagrams of the design finally see the output and aks "why did the add come out after the update?" Again, I casually say, multi-threaded asynch, so output order will be totally random on CPU and other resource time sharing in the server farm.

Well, that will never work.. we need to keep our downstream systems in lock step with the main database.

I pull up the story and show them it only says activity level.

A discussion about how the story was poorly written, is incorrect, the requirements are totally different....

Neadless to say, 2 years later, I'm still not allowed to toss the initial design, start over. No time, no resources, PM that worte the initial use case refusal to admit he gave us flawed requirements.... So, instead we hack, hack, hack, hack, trying to impose synchronicity onto a fundamentally asynch design.



There is a slight difference here.

Had we not shipped the initial flawed design, there would have been no functionality. We were creating new, not reworking existing. In this case, the existing functionality is there.

We sell upgrades for millions of dollars and ship twice a year. Had we delayed, there was only a painful (basicaly calling it a defect) way to ship which causes people to lose bonuses for defect counts that escape to the field. CCP is going to monthly drops, and they are free, so less pain in delaying the functionality.

They can still keep parts of the June release. For example, refine changes. Damage to T2 inputs. Move extra items to base. No remote research. The new UI. Infinite slots, solar system wide usage cost scaling. Compression arrays. New ships, changes to ships (including barges/exhumers). Changes to copy times.

The only things I've asked for are
1) Change to cost scaling to base it on the concurrent jobs in a facility rather than solar system level usage over the last 28 days. The can go forward with the current horrid design on cost scaling, and fix it post June.

2) Revisit the research changes, and do this NOW, because this one can not simply be rolled back after. Once they do the big round up, that can't be undone.


So, in summary.... DO NOT ship the research changes. Ship everything else. Post June, revisit the manufacturing cost scaling to be facility based on how many jobs are running right now instead of how many have run in the solar system over the last 4 weeks and make MUCH smaller changes (if any) to research.
LHA Tarawa
Pator Tech School
Minmatar Republic
#772 - 2014-05-09 19:25:28 UTC
Quintessen wrote:
LHA Tarawa wrote:

Moving complexity from research that is done once per BPO to manufacturing is bad, because I don't see this as "good, meaningful" complexity.


Except, CCP wants the complexity at manufacturing. They want you to consider location, teams and facilities in your calculations for both research and manufacturing so that your manufacturing reductions are coming from multiple sources. CCP has stated that this is good complexity because it forces you to make decisions and weigh pros and cons. I think any argument down this path isn't going to work.


Then that could still be done, without the other breaking changes to research.

Undo everything with research except the change to when the round happens.

The complexity moves to manufacturing and they don't have to change to only round up.



Let's think for a moment why they changed to only round up.

If you needed 4 before, then at upgrade 4 * 1.11111 rounds to 4. So no change.

Then with ME 10% and Outpost 5%, the 4 * 0,85 = 3,4 which would round down to 3. So, they hacked in a ceil so you can't get belwo the old base even on 4.

But that hack for 4 not going to 3, makes it impossible for the old 5 to get back down to 5 without multiple carriers in a batch. 5 * 1.11111 rounds to 6. 6 * .85 = 5.1.. which ceil() back to 6.



This is the point in the release cycle where hacks to cover up one fail begin to cascade into new fails.
Weaselior
GoonWaffe
Goonswarm Federation
#773 - 2014-05-09 20:07:08 UTC
LHA Tarawa wrote:

Again, rounding is not an issue. It is simply the source of the complexity. And CCPs stated goal was to remove complexity, and since they utterly misunderstood the source of the complexity, their goals have not been met.

this, right here, is one of the reasons that it's obvious you don't understand or respond to when your arguments are demolished, and instead just mindlessly repeat them

the goal is to reduce bad complexity, and increase good complexity

now, this is seperate from your complete inability to understand what complexity is but CCP has said repeatedly what I just said above, I have pointed it out to you repeatedly, and your response is to keep copying the same screeds

Head of the Goonswarm Economic Warfare Cabal Pubbie Management and Exploitation Division.
Weaselior
GoonWaffe
Goonswarm Federation
#774 - 2014-05-09 20:10:32 UTC
LHA Tarawa wrote:
Quintessen wrote:
@LHA Tarawa
I, and many, get that you feel that rounding is the source of a lot of complexity, but rounding is something that myself, and a lot of others are perfectly fine with -- something generally learned in primary and secondary school.


I never suggested they get rid of rounding. Their stated goal was to remove complexity, and since the complexity is because of the rounding, they failed to achieve the goal. As for learned i secondary school, if they can multiply % and determine round, then waste = 10% / (1+ME) should have been no problem at all!

what quintessen told you was that whatever your personal difficulties with rounding are, they're not problems for anyone else

your response was to mindlessly assert again that the complexity is because of the rounding, something so obviously not true that you will never convince anyone. like this is the "fire is wet" part of your argument: it is untrue. it is obviously untrue. it is so untrue that by asserting it (repeatedly) you make clear the issue isn't just you're wrong, its that you have no connection to reality

Head of the Goonswarm Economic Warfare Cabal Pubbie Management and Exploitation Division.
Seith Kali
Caldari Provisions
Caldari State
#775 - 2014-05-09 20:27:36 UTC
If I was an undergrad programmer dipping my toes into my first ever fizzbuzz like mr programmer over there I'd probably scratching my head over a bit of simple rounding too. Except it isn't complicated so I guess 140 isn't nearly enough IQs.

Please stop derailing this thread with your issues with basic integer math. I can recommend some excellent literature on the subject if you need it.

Apprentice Goonswarm Economic Warfare Consultant - Drowning in entitlement and privilege. 
Darin Vanar
Federal Navy Academy
Gallente Federation
#776 - 2014-05-09 20:31:55 UTC
LHA Tarawa wrote:
So, up to this point, we've been focused on how the changes will effect newer players, with 4-5x as long to get a BPO to sub 1% waste.

Well..... I hope it didn't slip past anyone, but if I understand correctly, the hacks they "tossed in" over the last couple days in an attempt to make this less complex, just increased the time it takes to get a carrier BPO to "pretty good" from 3 months to 5 years.


OLD:
The round used to happen at the run level, and round closest.

Chimera, base needs are (4, 5, 10, 5, 10, 37, 10, 4, 4, 4, 10, 10, 10) . With no research, the 5s become 6 and the 10s become 11, while the 37 becomes 41 so (4, 6, 11, 6, 11, 41, 11, 4, 4, 4, 11, 11, 11). Total extra was 13. % extra was 13/ 123 = 10.57%.

However, under the old, spend 3 months researching to ME 2 and those dropped to (4,5,10,5,10,38,10,4,4,4,10,10, 10). That is one wasted cap drone bay. So waste is 1/123 or 0.8%.

To get rid of that last 1 wasted drone bay would require ME=7 which was another 10.5 months. Few would bother unless you planned on making copies to sell. Then, the extra research didn't make the BPO sell for much higher price, but it did make it sell faster, meaning you could sell more, meaning more total profit.


NEW:
Round happens after full job calculation, and ALWAYS ROUNDS UP!

At deploy time, the base become the old waste numbers: (4, 6, 11, 6, 11, 41, 11, 4, 4, 4, 11, 11, 11).

ME 2 was 7.5% reduction in waste, which is no longer possible under the new "whole %" so let's go with the low end... 7% reduction.... which will take 19000 seconds * 854 = 188 day base (6 months), but let's say you can still cut 30% off that for skills and POS mod and get it down to 4.5 months.

Now your base numbers are multiplied by .93, then rounded up to get: (4, 6, 11, 6, 11, 39, 11, 4, 4, 4, 11, 11, 11). A whopping reduction of 2 drone bays. 2 / 136 = A change fro 1.5% reduction from the base 136 items needed from a 7% researched BPO.

(AS I said, the complexity has always been in applying the %s to the items needed and determining the rounding inflection points. And their last minute hack to apply at the job level and then round up, simply moves the complexity to manufacturing job start instead of research).

NOW, if I happen to be a mega corp that can crank out carriers in batches, and happen to run 3 chimera in a job, I can get 3 for (12, 17, 31,12,31, 115, 31, 12, 12, 12, 31, 31, 31) which is 378 items from a base 408 = 7.3% reduction.

See, ALL the complexity (always was, is and STILL WILL BE) of applying the % reductions to the items needed and finding the inflection point in the rounding is still there. All this change, breaking STUFF to remove complexity, didn't actually remove any of the complexity. All it did was move the complexity from BPO ME research to manufacturing job start.


Now, back to my main point: To get my chimera "pretty good", if I'm a small corp that runs billion ISK ships one at a time, I need to get those 11s back to the old base of 10 and the 6s back to 5. I'm half in luck. 11 * .91 is 10.01, which is still going to round to 11, but IF I take the BPO to perfect 10% reduction 11 * .9 = 9.9 which will round to 10. I'm only "half in luck" because 6*.9 = 5,4, which is still going to round to 6.

So, instead of 3 months to get to 0,8% waste, I have to go to -10%, which will be 256K second * 854 which is 7 years BASE time, and then if you can reduce that 30% for skills and POS mods is "only" 4 years 10 months. And for that almost 5 years, you're getting (4, 6, 10, 6,10,37,10,4,4,4,10,10,10). (136-125) / 136 = 8% reduction.


The complexity was NEVER in waste = .1/(1+ME). The complexity is, was, and STILL WILL BE from applying the % to the items needed and finding the inflection points in the rounding!



So, again we:
1) Screw over players that spent a lot of time on research because people that spent way less get to catch up in the big conversion.
2) Screw over newer players with WAY longer research times.
3) Remove the ability to research to between whole % waste.
4) Do not remove any of the actual complexity, which has always come from applying the % to the items needed and identifying inflection points in the rounding, we simply move the complexity to manufacturing job start optimal runs.

They didn't understand the problem (complexity was in applying % to items needed and finding inflection point in the round, not in the old formula waste = .1 / (1+ME)).

So, they attacked the wrong thing (waste formula instead of round). In doing so, they broke a lot of stuff (can't research between whole %) which had big implications ( The big round up; the jump from 9% to 10% has always been huge, but in the past it was unnecessar becuase I could do partial %s, but since I won't be able to do partial %s, it iwill be more important to get to perfect).

When it was pointed out that the complexity is in the round not the formula so they dodn't actualy remove any complexity, they slapped in a fix to apply the round at the job level and round up, which simply moved the complexity to manufacturing, AND broke a bunch more stuff, like having to run billion ISK ships in batches.

FAIL!

FAIL on so many levels!


Stop hacking. Stop breaking stuff!!!!!!

Start over.


+1

This is getting worrisome.

CCP, what is going on with the new ME? It seems you are scrambling to fix something but you are incorporating more issues. You are clearly running out of time and this isn't ready for release. If you run DB scripts in the current state, you are going to create a disaster.

Here's a crazy idea. How about you delay this ME system from the Industry changes? Keep the old system, but incorporate your removal of slots and increasing install costs based on usage, and deploy this (whatever this is) when it's *ready* for release.

It might not even be necessary.
Darin Vanar
Federal Navy Academy
Gallente Federation
#777 - 2014-05-09 20:38:54 UTC
And to keep the research POS relevant, keep the system usage costs apart from the POS entirely.

The old system is far more elegant in design. I don't know why you are changing so much just to do away with slots. You don't need to overhaul the Jeep into a Minivan just to change the tires.
Weaselior
GoonWaffe
Goonswarm Federation
#778 - 2014-05-09 21:02:06 UTC
another blatantly obvious issue missed by LHA: waste no longer exists

its not a thing that's there, there's the cost then you apply things that reduce cost

it would be garbage to have teams reducing costs by even, nice amounts, pos and ouposts reducing costs by even, nice amounts

then bpos reducing amounts by 10% - 10%*(1/1+ME)

like man, would that be a moronic system

Head of the Goonswarm Economic Warfare Cabal Pubbie Management and Exploitation Division.
Weaselior
GoonWaffe
Goonswarm Federation
#779 - 2014-05-09 21:05:28 UTC
Darin Vanar wrote:

CCP, what is going on with the new ME? It seems you are scrambling to fix something but you are incorporating more issues. You are clearly running out of time and this isn't ready for release. If you run DB scripts in the current state, you are going to create a disaster.

Here's a crazy idea. How about you delay this ME system from the Industry changes? Keep the old system, but incorporate your removal of slots and increasing install costs based on usage, and deploy this (whatever this is) when it's *ready* for release.

It might not even be necessary.

the scripts are simple as all hell, incredibly simple, even LHA could code the logic, because it's an incredibly simple conversion formula that's put in the initial post that has exactly 7 states: if it's ME0, ME1, ME2, ME3, ME4, ME5-9, ME10+

you can just do seven passes of a stupidly simple script

seriously people, put some effort into your baseless attacks

Head of the Goonswarm Economic Warfare Cabal Pubbie Management and Exploitation Division.
ElectronHerd Askulf
Aridia Logistical Misdirection
#780 - 2014-05-09 21:08:36 UTC  |  Edited by: ElectronHerd Askulf
CCP Greyscale wrote:

Yup, there is a little bit of additional rounding waste in the negative MEs. I had this in the blog for a while, but I took it out because it was fiddly to explain and I figured someone would figure it out pretty quickly in the comments anyway. So, you get a small prize of nonspecific nature for being the first (that I've noticed!) to flag this up.


Now that I've written up some code for the new world, I've run into this as well. This rounding waste in the case of a Jaguar adds up to somewhere around 700,000 isk for a build from components. Given the profit from that build would be about 2 million, that's significant.

Additionally, turning the crank on a naive implementation of this stuff leads me to the improbable situation of requiring 2 rifters to build 1 jaguar. This is based on Nullarbor's comments regarding having a 'ceil' prior to multiplication by number of runs in a job (to support the 2% material reduction in POS modules effecting low-input blueprints). Will there be still be inputs that will be treated differently, or am I missing additional rounding somewhere? I'll throw in a 'floor' for now, but I'd really like to know I'm putting it in the right place before I start making decisions about this.