Science & Industry

Some of us have already acted upon this information.



I must say, the module invention and production times are being shaken up more than I thought they would be. Interesting.

I will most certainly be acting upon some of this information as soon as I can log on.

A few more thoughts on the impact of increased time to market.

This will have a huge impact on increasing price spikes in the early days of alliance wars. During week 1 of the Fountain War I recognised that Goonswarm demand meant that Oneiros prices had soared, and I started vomiting them out as quickly as my available capital allowed (I was capital limited, not production line or science line limited).

At the time Oneiros build price was about 120m and the pre-war price had been about 135m, but Mynnna's minions bought enough to make the price soar.

I clearly wasn't the only person that made this shift, and after ~4 days (which was my time-to-market) the Oneiros price stopped rising and stabilized at an extremely profitable 190-195m. Over a few weeks, this slowly declined as more people got into the market, and the production cost went up as people started charging more for Photon Microprocessors.

Had time-to-market been 8 days instead of 4, the equilibrium reached may have been different, and definitely less stable.



Edit: Scrap the conclusion. Looks like I had some numbers wrong and the Oneiros may be quicker to get to market, not slower.

Oh man, 90 day base invent time for a JF.

Definitely watching this space.

Greyscale - when can we expect to get this on sisi? A lot of us are simply tinkering with spreadsheets right now and holding back opinions until we can see this in action.

I considered including decryptor effects, but decided against it, as we still have no idea what they'll be changing to.

If I do include decryptor effects, I also have the option of easily optimizing for profit/hr and having the data use the optimal decryptor for each item (while Symmetry is very popular, it's still not the optimal decryptor to use for most items). That, of course, would create another problem: decryptor prices aren't going to stay the same post-Crius, and which decryptor is optimal does heavily depend on their price.

If I have enough time, I actually have the proper code scaffolding to go in, get volume/day data for decryptors, T2 items, everything, assume that decryptor effects are staying the same but just scaling up from -4/-4 base to 0/0 base, and come up with a proper estimate of how decryptor demand would change. Probably not going to happen though. It'd easily be several paragraphs of code, whereas everything else I've done in this thread has been just copy-pasted pre-existing code with trivial changes.

That said, I'm not doing anything more today, as I'm already half-asleep. Tomorrow I'll actually look at that data I had my scripts spit out, and hopefully be able to make some sense of it. I might even decide to involve decryptors somewhere; suggestions are appreciated.

Is it me or Fuel Block manufacturing time is now 3x longer ? from 5 min to 15 min.

alright, i can't sleep and decided to take another look at this: Time for more assumptions that may, or may not be wrong. read everything with a big grain of salt.

I have tweaked some values so that i end up with the actual invention time that will be relevant ingame (that means taking all inventiontimes from greyscales .csv and halving them. nobody will do serious invention in a station if you can double your output in a POS.

i end up with some values that i actually really like and a lot more that i don't like. let's just go through a bunch of them.

i tooked at the value purely based on one thing. play sessions. players tend to be online at the same time every day. that is usually in the evening between 18.00 and 00.00 local time for up to 4 hours. it is much more likely that someone skips a day than it is that he or she can log in at a different time.

i therefore assume that a multiple of 24 minus X, or anything under 4 hours is actually a good number for invention jobs. everything else most likely means some kind of wasted slot time. my OCD hates wasted slot time.

so far, i have completely ignored the whole manufacturing - invention balancing thing. this post is purely about invention times and optimisation of slot usage.

lets dive into the numbers:

Weapons

6 hrs for small weapons. whelp.
13hrs 30 minutes for medium weapons and most medslot modules. that's just bad. it feels wrong that modules that are so important are made under such bad conditions.
22 hours invention time for large weapons. that is actually a great number.
12 days for capital modules. is that for a 10 run or a 1 run copy ?

Ammo
19 hours for small ammunition. fine. BUT: a MASSIVE difference to the status quo. good to know ;)
2 hrs for mining crystals. why the difference to small ammo ? forgot them ?
1 days, 13 hours for medium ammo. seriously ?
2 days, 8 hours for large ammo. whelp.

Rigs
12 hrs for small rigs. just as bad as medium modules.
24 hrs for medium rigs. oh dear
37 hrs for large rigs. okay-ish. you loose a bit of sciency slot time, but i really like long timers ;)
50 hrs for capital rigs. oh dear.

Ships
4 hrs for frigages. get that to ~20 hours.
9 hrs for destroyers. bad. get that to 40 hours
9 hrs for cruisers and industrials. whelp. should be longer than destroyers, get it to 60 hours or so.
12 hrs for Battlecruisers. seriously ? 5 days would be appropriate
13 hrs for battleships. seriously ? should be much (!!) longer. i would suggest getting that up to a week, maybe 10 days. i would prefer a week, mostly because ~building a bunch of marauders/blops sounds like a funny project for a weekend

Drones
2 hrs for small drones. okay
4 hrs for medium drones. not that cool, but okay-ish
6 hrs for large drones. whelp

I very much (!!) appreciate your move towards significantly increased job length. everybody hates the clickfest and this should really help with some people carpal tunnel syndrome. good job

however, the actual numbers you guys ended up with are often quite meh-ish or just horribly bad.. it's really hard to optimize slot utilisation with those values.

That by itself is not a bad thing. it just doesn't feel right. I even see that there may be some game-design reasons for non-optimal timers. (longer play sessions, multiple daily logins, ...)

however, profit in manufacturing should come from making the correct decisions to the following questions

1. what should i build ?
2. where should i build it ?
3. how should i build it ?

But not based on logging in at work or staying up longer because of a video game.

Therefore i have two suggestions:

1. move back to the drawing board with rank selection. choose a finer scale and work from there. if i understood the concept of ranks correctly it's something that we won't actually see ingame, it will just work in the background, therefore a weird rank won't confuse anyone who looks at blueprint data ingame.

how about multiplying the rank scale and all current ranks by the factor 10 ?

From there, you can move things along the rank scale a bit to end up with ~nicer invention times~ without disrupting the general idea that everything is based on one single value.

If i understood your formulas correct (and sadly, i'm very much not a math genius) this won't break things as long as you just divide everything by 10 at the end.

2. split invention into things that are done within one play session (that means invention times of ~3 hours) and things that are done in multiple session (that means invention times of ~20, ~44, ~68, ... hours)

From there on, work with decryptors and meta items to make sure that enough ~stuff is supplied. if you feel like some special rules are necessary in order to keep the market alive until you can do some real improvements, go ahead (it may be necessary to increase the runs on module or ammo T2 BPCs for some time)

I will (hopefully) post something similar about manufacturing tomorrow and some more things about decryptors and meta items in the invention process later this week.

T2 BPO are going the way of the Titan - slowly nerfing them and buffing everything around them, until they are pretty much useless.

Does that include rigs?
At the moment invention produces 1-run BPCs.

Just to make sure I'm not crazy with how I price things, and determine if something is worth making:

(I'm bad, because I normally ignore the time to copy the blueprint, and actually /do/ the invention. I don't ignore the cost of invention.)

I take all the material costs (including an average invention material cost) and subtract that from the sale price. I then divide that by the number of hours needed to make the thing. If the total build time is below 24 hours, I normalize the time to 24 hours (As I'm generally logging in once per day.)

If that doesn't beat (in general) 100,000 isk per hour, it's not worth making.


So, for an Expanded Cargohold II (it's a go-to example for me.) current material cost is around 148,000 at jita sell prices (the only real way to price things, except possibly components.) The invention cost is around 75,000 per unit.

Current sell price is around 510,000.

This leaves a profit of around 287,000 per unit. As I can only make 10 per day, per slot, I multiply the profit by 10, then divide it by 24 to get the isk/hr which is around 120,000 isk/hr.

If the build time went up by a factor of four, the price would have to go up by around 200,000 for it to remain a viable product for me to make.

For the odd days I can make two batches, it becomes less viable.


Summary:
Increasing build times will directly impact on module pricing, as serious manufacturers don't look at profit on a percentage basis, but instead on an isk/hr level, with a baseline of what isk/hr they'll accept as viable.

Revising earlier post, since I was using a wrong number: https://docs.google.com/spreadsheets/d/1ZcEnJQDAYqMARdp8bpMdIz4Ai6c4xeBmk8NDQMI5IqY/edit#gid=0 Assumptions remain the same, listed on the sheet. Only mistakes at this point, I hope, should be places where I applied the wrong decryptor by changing formulas en masse.

Goes without saying that different decryptor choices will significantly change the outcome...


This isn't a bad idea. It's not quite as easy as going, "okay that build time is for a 10 run job (for modules)" because then you just wind up lowering module build times by 50-70% given the current numbers, but it wouldn't take much more fiddling than that to get a favorable change.