Clarification on Material Efficiency research

Ok so I’m just getting into some industry and have researched a few rig blueprints to 8% ME. Then I wanted to do some calculations before committing to the long haul of getting all the way to 10%.

So say I have a blueprint that requires 14 armor plates, 9 Contaminated nanite compounds, and 17 fried interface circuits.

My understanding is that when you calculate material efficiency, you can’t be left with part of a component. It will always round down.

For example, at 8% ME, the 14 armor plates would be reduced by 1.12. But the .12 doesn’t count, so at 8% efficiency, it would still require 13 armor plates per run.

The big deal is when, for this particular item, it requires 9 contaminated nanite compounds (which is by far the most expensive of the materials required), and starting with only 9, even if I take it to 10% ME, it would only take away .9 units which isn’t enough to get to the next whole number so it would STILL require 9 CNCs for each run, which means the time spent to get to 10% ME research is worthless as regards this material component and offers negligible savings for this product on the whole.

I guess the reason I’m posting is just to make sure my understanding of how all this works is correct. Because this would mean that the profitability of ME research is significantly dependent on the numbers of materials originally required. So of course something that requires 109 of a X material reduces by 10 and the loss of the .9 isn’t that bad. But something that only requires 9 or 19 to start with, well then the extra research is a lot less effective.

Any insight you can offer is greatly appreciated! Thanks!

You are correct: It rounds up to the next whole item.
If you need 30 widgets to make one item then with ME=8 that’s 27.6, you will need to provide 28 widgets to make one item. A saving of 2 widgets.

But. And it’s an important but…
The ME is applied after the total needed for the whole job is calculated.

So, in the above. To make ten (10) items you would need 300 widgets on an unimproved blueprint, but with ME=8 that becomes 276 widgets (a whole number) and thus a saving of 24 widgets.

3 Likes

Ok if true that makes a huge difference so that the more runs you do at a time, the better the savings. Is there an official document/explanation that spells that out? I could have sworn I read that it was on a per run basis, but I freely admit I may be (and hopefully am) misremembering.

Probably: https://eve-industry.org/export/IndustryFormulas.pdf
Not official but worked out from in game measurements and tests as well as CCP’s published information.

The one wrinkle is you never go below one item per run.
So if you need two Widgets and one Thingie per Item on an single run of an unimproved blueprint:
Improving it to ME=8 and doing 100 runs requires 184 Widgets but still 100 Thingies.
So if you need one widget per item and do 100 runs it’s still 100 widgets, not 92.

3 Likes

You’ve been incredibly helpful, thank you so much!

1 Like

As you advance in your industrial career blueprint quality will become increasingly important because you will be doing more runs per job - for advanced components I typically make one or two thousand at a time. At that scale, the blueprint ME/TE and structure rig bonuses make a big difference.

Starting out, a lot of the T1 products are marginally profitable even with a fully researched blueprint once you factor in the various taxes and fees you’ll encounter on your way to market.

There are exceptions - some small rigs and many scripts are very profitable and can be researched to ME10 in a few days. The downside is that these tend to be inexpensive items so you need to sell a lot of them! Use a tool like https://www.fuzzwork.co.uk/blueprint/ to play with the numbers.

1 Like

First, all subcap items MUST be 10/20 . The time to research is very small (actually BC and BS can be long but still)

Second, remember that ME does not affect item requirements that are 1. eg if a nighthawk bp needs 1 drake to manufacture, even with 99% ME over 10 000 runs it will still need 1 drake per run.
So the material usage of a job, for a BP that requires M material per run, over R runs, with effective material mult MM (with MM = 1-ME/100), is :
M==1 ? R : ceil( MM×R×M)

This is based purely on curiosity, but what do you classify as “very small” in regard to research times? Obviously that’s a subjective question, but I’m just wondering what you consider it to be.

edit : the values given in those screenshots bear only meaning for me. For you different values would make sense. Still the ratio of times should remain approx the same.

You need to work not by one bpo but with several . eg ATM I am building items with like 20 BPO in parallel (carbonide something) . So here the time to research is ×20 (because I need 20 lines at the same time). Same for T2 of small modules, you want like 20 or so of them (for copy, craft T1 fast).

The rigs on the other hand … they are so fast to build, you may want at most 3 for M, 5 for S, 1 for L, and they take like … dunno
Checks the actual research time ingame
L is 37D ; M is 24 ; S is 12D (that’s the time to go from 0/0 to 10/20, just checked on random rigs)
=> 37D for L, 72D for M, 60D for S

Just checked CDFE (I think it’s the most sold rig), and it’s 7/12/3/2 BPO needed

So in terms of research, days, it is
S : 6.1×2×7 = 85D
M : 12.2×2×12 = 293D
L : 18.5×2×3 = 111D

for I believe what is the highest volume of rigs in the game.

FYI here are the requirement for me to provide enough auxiliary nano pump, both in T1 and T2 production :

it’s 3/3/2/1

for the same activity,

frigate BPOs :

destroyer BPOs I need :

for the cruiser/battlecruiser :

(talos, orcale is 1 and thorax is 4 )

FYI 97D to make an augoror BPO 10/20, times 3 =291D : for one of the lowest volume cruiser, it requires as much time to research as the highest rig.

So yeah the time to research the rigs is very low, compared to the time of other stuff. If you want to be competitive you should research them to 10/20.

(/20 is not really necessary. It reduces the temporal cost of making stuff, but if you can’t build very interesting stuff that temporal cost is very low)

That’s a LOT of information, but I appreciate it! I see your rationale, but I think you’re operating at a far higher level than I plan to what it comes to concurrent jobs and such. It it’s good to know what some folks are doing and what the upper levels of industry are capable of.

Again, those number made sense the moment I posted them, but are completely wrong just as of now.
Still my point is about the time to research, for an amount of BP to use to provision a market, and the ratio should remain in the same amounts.

But don’t consider the numbers I showed you as being anything relevant in themselves. Because besides the number of BPO I have, those screens are wrong.
I specifically made modifications to show you the data I wanted to show you without showing you things I would not want to show or that you could guess from the screen. e.g. you can see that several BP I have much more than needed, while other I have 0, which should be a hint that this is not my real work.

So no that’s not a lot of information. The sole information is : rigs require much less time to research, and are very cheap, so you rather research them all to 10/x than use them with less than 10 ME.

Spurred by your information, it seems like you can supply a LOT of whatever you’re making, which obviously means you’re choosing items with high turnover. So a question related to that:

When you look at the market history window, the bar graph at the bottom indicating volume, is that volume of fulfilled buy/sell orders for that particular station, for the region, etc? (Really this question applies to all the information displayed in this window.)

I love figuring things out for myself and completely understand why you’d hide some info in your screen shots. I DO want to figure out my own “secrets.” But I need to have a firm understanding of al the tools available.

You’re making too much assumptions here.
I do what I do, and the least people can do what I do, the more value I make so obviously I won’t help you do what I do. But I can help you understand what YOU can do.

History is per region.
It’s not the volume. It’s the tampered volume. The only way to have correct volume is by saving it before CCP tampers with it, so eg towards 23:55 every day.
It’s obvious that CCP tampers with it since the data for one day is inconsistent from one day to another. No example provided because that’s redacted data ; let’s just say that the “high” value for today (2020-08-04) can become LOWER if observed tomorrow, which is not possible unless there is tampering by CCP.

I actually did not “hide” things. I instead showed you exactly what you needed to see, so what I would do if I were in your case.

Maybe I’m going places too dark and scary to consider (haha), but in what ways and for what purposes do you think CCP is tampering with the information in the market history window?

Also, I didn’t mean anything negative when I said “hide,” I just meant that if you chose not to share some info, I totally understood why.

I asked and a GM told me that it is indeed tampered with, when I showed him proofs that the data presented in the game was incorrect. He also told me they had a reason to do that, and gave me a reason which was actually absurd (resilience to market manipulation). I can’t share discussion with a GM (it’s against EULA and it could give information I’m willing to hide), but basically all I can day is, that data is not trustworthy.

IMO it’s just like that and they don’t understand/consider why it is bad to tamper with data you propose to people (it defeats the role of that data, especially the way they tamper with it). They had what seemed a good idea, and applied it, without realizing it was a bad idea. It happens a lot ^^ .

I can’t talk about things I don’t know, so since I don’t have access to the algorithm used, all I can say is it IS tempered, and I have my idea of what they do (mostly because I have the same result when I apply my own correction algorithm), but talking about that is just speculations.

Yes.

Sorry that this isn’t a specific question, but do you have any other advice or tips based on things that people often get wrong due to misinformation or fallacious thinking?

People sometimes believe that their understanding of real-life markets can be applied to the in-game market. Then they get annoyed when the two are not the same.

1 Like

I suppose that’s where ignorance might come to my rescue, then! I won’t have that problem, certainly :-). Though I do find it an intriguing (and believable) point.

Both a lot - and actually none.
You need experiment to get your beliefs rights. Me telling you won’t be as good as you experimenting errors, and learning.

so… just try. Remember that whatever you belief, may be actually completely wrong, and the money you invest in industry is wasted until it magically comes back multiplied - but just because it happened once does not mean that you were right.
At first when I started industry, I used PVE to feed money, and spent a lot in it. A lot of errors, a lot of experiments. Worth it in the end ^^

This is just a general case of what @Xeux wrote. Just because you can make a model that matches what you observe, does not mean that what you observe actually follows that model. So try to make model, try to understand, try to optimize, but always remain conscious that you know nothing, John Snow. You are playing a game, but at the same time you are part of that game, so people are playing with you.

I think what you’ve just described is why I’m having fun this go around in Eve. In prior attempts I wanted to do everything right the first time around and It really ruined my day if I messed up. Now I am older and perhaps a bit wiser and just want to learn as I go. Obviously I’m still asking questions because it’s be silly not to, at least at a certain level. But I’m much more willing to try things and, in the case of industry, if I spend all the time and energy to research a BPO and make a bunch of that product and it turns out it wasn’t as good a decision as I thought, well, that’s not time or isk wasted, it was time and isk spent learning, Which I can now realize has tremendous value all its own.

So thank you for the perspectives and insight, and I’d gladly accept more from anyone who wants to share :-). I look forward to a progressing (and somewhat dabbling) career in industry!

1 Like