Constructor armies/stacking

[beans]

Is there a benefit to putting engineers in an army and having them work together? They don't seem to work any faster or any more efficiently, and it's faster to have each engineer work independently on a few buildings rather than having them all work on the same buildings.

I used to play Planetary Annihilation, where construction speed is linear to the amount of constructors and buildings can be built hysterically fast with enough constructors. Maybe there's a minor speed boost, but I'm so used to linear stacking, I'm not noticing it. That's my best guess, because I can't see any real benefit to stacking workers in Ashes.

[Hesa_Heart]

its not linear , im not sure of the right word except each addtional enginner effect reduces each time one is added to the group so putting 10 enginners on a single build is not good, i found 5 work best. im sure some of the others more versed in the math can ex[lain it better

[TSP]

Originally posted by bnb_home1:

its not linear , im not sure of the right word

Diminishing returns.

I personally strongly suggest that you do not have an Engineer assisting a building unless you absolutely need the unit that is being produced right this instant (For reasons of dying or something similar). You are much better off building another factory with that Engineer instead. Takes some investment regarding metals and radioactives for sure, but it is generally considered the correct play.

[Griezell]

the way i see it that you never want to use more then 2 on building stuff and never more then 1 assisting a facotry . only more then a few for assisting an dreadnought factory but that wil melt your metal inome

[Kirdaiht]

If you need that specific thing ASAP, you put as much engineers on it as you can spare. The returns diminish, but every engineer still shaves of a few precious fractions of a second.

If you want as much of a thing as possible, work in parallel with as little engineer assitance as possible. Every construction project ideally has one engineer, and every engineer has one project.



Factory assisting is usually not worth it. There's usually other better things the engineer can do, like build another factory. Some exceptions exist. Use your own judgement to find out if your situation is one. (Example, if you don't have any tasks for that engineer, Then he might as well assist a factory and be somewhat productive. Same if you rather have one dreadnought now over two dreadnoughts in 2 minutes)

[centaurianmudpig]

My current play style is to have an Engineer support each Factory, it really does help speed up production so long as you have the Metal.

When it comes to constructing buildings, again if you have the resources to sustain engineers paired, I will do so.

Production speed is increased and will give you an advantage over a single Engineer.

Each time you add an Engineer to assist in construction the speed increase is decreased each time, hence the deminishing returns. What the sweet spot is would require some investigation and perhaps be based on playstyle and the map being played on.

[Kirdaiht]

I've done some testing with factory assisting and think the conclusions are worth sharing.


A single engineer adds half of a factory worth of buildpower. So a factory with an engineer assisting will finish 3 units in the time a factory without assitance will finish 2.

An extra engineer costs 400 metal, and a factory worth of engineers costs 800 metal (Note, this is for the first engineer assisting a factory only)

A PHC frigate factory costs 150 metal. Instead of building two engineers to get one factory worth of buildpower, you could also build 5 extra factories, for 5 factories worth of buildpower, and still have 50 metal to spare. Cost for cost, building more factories is 5,3 times as effective as building engineers to support your existing factories.

On the PHC air factory, the cost of a new factory is 175 metal. Now we can only build 4 factories for the price of two engineers, (with 100 metal to spare). Cost for cost the factory is 4,6 times as effective.


For the armory it's slightly more complicated. An armory costs 250 metal and 90 rads, and converting those 90 rads in their equivalent value of metal is not always obvious and depends a lot on the map and your playstyle. I'm going to estimate 1 rad to be worth 2 metal, which makes the armory cost 440 metal. This puts the armory on being 1,8 times as effective as an engineer. If you have an air-heavy playstyle or have a huge rad deficiency you might find that a single radioactive is worth a lot more than 2 metal. Using engineers instead of building extra factories becomes worth it if you give 1 rad a value of 6,1 metal. These situations are probably rare.

Now on to the dread launch. Your dread launch costs 300 metal and 320 rads. With our previous conversion rate of 1 to 2, this is 940 metal equivalent. In this situation you do want to assist your building instead of building a new one. Building a new dread launch is 0,85 times as effective as adding one engineer to an unsupported dread launch. So you do want to support your dread launches with at least one engineer. (In some situations where you have relatively high rads, you might find another dread launch a better option)

But what about a 2nd engineer?
This is actually quite simple. The 2nd engineer is half as effective as the first engineer. If all your factories already have one engineer supporting them, adding a 2nd one adds 1/4th of a factory worth of buildpower. Now our dreadlaunch+engineer combo is equivalent to 1,5 factories. so we need 6 extra engineers to match this. That's 2400 metal compared to 1340 metal equivalent. The 2nd engineer is unlikely to be worth it for the dread launch.

Last point. what if we feel like 2 dread launches just take up too much space and we rather just support our first dread launch untill it is twice as fast.
Each assisting engineer is half as efficient as the one before it. so to get to 2 factories, we first add 1/2, then 1/4th, then 1/8th, then 1/16th (you can see where this is going). We will never actually make our factory twice as effective as the first.

I'll probably post this as a guide somewhere in the very near future.

[Glock Glock]

Thank you for sharing this valuable info. I usually only assist my HQ in rushing out a 2nd engineer early, and the math you've done here supports that assisting is generally not worth it.

Originally posted by Kirdaiht:

I've done some testing with factory assisting and think the conclusions are worth sharing.


A single engineer adds half of a factory worth of buildpower. So a factory with an engineer assisting will finish 3 units in the time a factory without assitance will finish 2.

An extra engineer costs 400 metal, and a factory worth of engineers costs 800 metal (Note, this is for the first engineer assisting a factory only)

A PHC frigate factory costs 150 metal. Instead of building two engineers to get one factory worth of buildpower, you could also build 5 extra factories, for 5 factories worth of buildpower, and still have 50 metal to spare. Cost for cost, building more factories is 5,3 times as effective as building engineers to support your existing factories.

On the PHC air factory, the cost of a new factory is 175 metal. Now we can only build 4 factories for the price of two engineers, (with 100 metal to spare). Cost for cost the factory is 4,6 times as effective.


For the armory it's slightly more complicated. An armory costs 250 metal and 90 rads, and converting those 90 rads in their equivalent value of metal is not always obvious and depends a lot on the map and your playstyle. I'm going to estimate 1 rad to be worth 2 metal, which makes the armory cost 440 metal. This puts the armory on being 1,8 times as effective as an engineer. If you have an air-heavy playstyle or have a huge rad deficiency you might find that a single radioactive is worth a lot more than 2 metal. Using engineers instead of building extra factories becomes worth it if you give 1 rad a value of 6,1 metal. These situations are probably rare.

Now on to the dread launch. Your dread launch costs 300 metal and 320 rads. With our previous conversion rate of 1 to 2, this is 940 metal equivalent. In this situation you do want to assist your building instead of building a new one. Building a new dread launch is 0,85 times as effective as adding one engineer to an unsupported dread launch. So you do want to support your dread launches with at least one engineer. (In some situations where you have relatively high rads, you might find another dread launch a better option)

But what about a 2nd engineer?
This is actually quite simple. The 2nd engineer is half as effective as the first engineer. If all your factories already have one engineer supporting them, adding a 2nd one adds 1/4th of a factory worth of buildpower. Now our dreadlaunch+engineer combo is equivalent to 1,5 factories. so we need 6 extra engineers to match this. That's 2400 metal compared to 1340 metal equivalent. The 2nd engineer is unlikely to be worth it for the dread launch.

Last point. what if we feel like 2 dread launches just take up too much space and we rather just support our first dread launch untill it is twice as fast.
Each assisting engineer is half as efficient as the one before it. so to get to 2 factories, we first add 1/2, then 1/4th, then 1/8th, then 1/16th (you can see where this is going). We will never actually make our factory twice as effective as the first.

I'll probably post this as a guide somewhere in the very near future.

[Ticktoc]

When an engineer assists a factory what -metal is it creating per second? Does it tell you nowadays if you select it?

[realdead_man]

Ok, that is production. What about build speed?

I been wanting to test it my self....never did.

Building a dread launch:

How much faster is it built with 2 engineers vs 1? or 4 vs 1? etc..

[Burnin-Teddy]

Originally posted by Ticktoc:

When an engineer assists a factory what -metal is it creating per second? Does it tell you nowadays if you select it?

There was a post somewhere that had tables denoting the impact of each additional engineer on a build project. Might be somewhere under "Guides" but I'm not 100% sure anymore. It essentially shows that each additional engineer is half as effective as the previous one. Capped at five engineers iirc.

Anyways, to answer your question: When an engineer assists a project it primarily increases the resource drain. Since the costs of the project don't change this results in a faster build time.

So for example, let's say you're constructing a PHC engineer.
Base values are 400 metal in total costs, 16 metal drain per second resulting in 25 seconds build time (400 / 16 = 25).
Add one engineer and the metal drain increases by 50% to 24 metal per second. This results in a build time of roughly 16.67 seconds (400 / 24).
Add a second engineer and the metal drain increases by 50% of the last increase, to a total of 28 metal per second. Accordingly the build time is now reduced to about 14.29 seconds (400 / 28).
And so on...

[Kirdaiht]

Originally posted by realdead_man:

Ok, that is production. What about build speed?

I been wanting to test it my self....never did.

Building a dread launch:

How much faster is it built with 2 engineers vs 1? or 4 vs 1? etc..

Every new engineer adds half the build-power of the last one.

2 engineers is 1,5 times as fast as 1 engineer. 3 is 1,75, 4 is 1,875 and so on.

Building two things at once is faster than having two engineers build them sequential.

[centaurianmudpig]

There is a guide for Build Time & Costs. Just goes to show that the one engineer is the best for producing units.

http://steamcommunity.com/sharedfiles/filedetails/?id=549485829

[Ticktoc]

Originally posted by Burnin-Teddy:

Originally posted by Ticktoc:

When an engineer assists a factory what -metal is it creating per second? Does it tell you nowadays if you select it?

There was a post somewhere that had tables denoting the impact of each additional engineer on a build project. Might be somewhere under "Guides" but I'm not 100% sure anymore. It essentially shows that each additional engineer is half as effective as the previous one. Capped at five engineers iirc.

Anyways, to answer your question: When an engineer assists a project it primarily increases the resource drain. Since the costs of the project don't change this results in a faster build time.

So for example, let's say you're constructing a PHC engineer.
Base values are 400 metal in total costs, 16 metal drain per second resulting in 25 seconds build time (400 / 16 = 25).
Add one engineer and the metal drain increases by 50% to 24 metal per second. This results in a build time of roughly 16.67 seconds (400 / 24).
Add a second engineer and the metal drain increases by 50% of the last increase, to a total of 28 metal per second. Accordingly the build time is now reduced to about 14.29 seconds (400 / 28).
And so on...

Thanks for that, it was not what I was expecting. I thought the engineer doing the assisting would also incur a resource cost just for the act of assisting itself and this would be where some of the inefficiency would come from. Doesn't seem to be the case and the inefficiencies seem to be scaled in but it looks like it scales down the improvement whilst also limiting how much faster you burn through resources at the same time. I suspect this was done to make it harder to completely tank your economy quite so easily like you can in FAF. Very useful to know, thanks again.

Thanks for that link also Centaur, some of those numbers have changed but the principles seems to hold true.