Part-building units

[couerdelion]

The question was raised by Vel about whether or not there might be something to gain by part-building a unit but not finishing it and leaving it in the build queue. It was suggested that this might be a mechanism which could be used to save on military upkeep.

To begin to look at this we need to figure out first how the costs work and I will be working here with standard speed, monarch level and assuming no inconvenient civics to spoil my numbers (eg Vassalage, Pacifism)

The unit costs can be calculated as

Total units - Free units - Handicap costs

Free Units is a population dependent number. You have a basic level of 8 free units:

plus 1 if total population is 1-4
plus 2 if total population is 5-8
plus 3 if total pop is 9-12
etc

total pop is simply the total of the sizes of each city so you'll end up with the same number of free units when you have 10 cities size 1 or one city size 10.

For the purposes of this discussion, I will assume that population is fairly fixed. In any case, you will not be adjusting your population to save on military costs since each population will almost certainly add more to city maintenance costs. Also, on average, each extra population point is only going to save you 0.175 gpt (on average) - trust me on this for now. The numbers used to derive this will come soon.

After deducting free units we also get a handicap cost so that we don't pay for every extra unit above our free support level.

The formula for handicap cost (expressed mathematically) is:

INT[{Total Units- Free Units} * 3 / 10] + 1

So from your excess units the free ones are the 1st, 4th, 7th, 11th, 14th etc.

Moving now to the original question, we can look at this at a macro level or a micro-level. By this I mean that the micro-level analysis will say that either the extra unit is free or it costs 1 gpt - and in any random situation, there is a 70% chance that it will cost you. The macro-approach will simply say that each extra unit costs you 0.7gpt.

Now to the wider issue of whether or not it is worth investing hammers in a part-build unit but not to finish this to save on costs. Here I will take a macro-level approach to valuing assets and I will assume that hammers and gold are broadly valued at a rate of 2h = 3g

Lets now assume that we have x hammers "invested" in this part-build unit. After 10 turns, hammers will be lost and since we are only saving 1gpt (max) from not having the unit, we are losing 1 hpt but gaining 1gpt. Our above relations tells me that this is making us a loss so we can only assume that the unit will not be allowed to degrade through our failure to finish it.

This then puts a maximum saving of our part build of 10 gold - hardly ground-breaking stuff anyway but the cost of that saving is an investment of x hammers. Looking at our hammer/gold equation we will then only do this if

10 gold > x hammers, => x < 6.7 hammers

So we are definitely not talking here of a unit that is one turn from completion but rather one that is perhaps a simple whip away.

There are therefore three reasons why we might start a unit build and leave it for finishing later.

1) We have nothing else to build now but will do shortly. It gives the city "something to do"
2) We are near to growth and want to use this extra population for some other build (eg worker/settler).
3) We need extra "insurance" so want to put a unit as part-built to allow us to quickly get one if needed.

The first two examples are those things that would be considered anyway without reference to the specific desire to hold an unbuilt unit in reserve. So we are very much left with the insurance question.

Now here I can certainly see some value in the part-build if we consider that the x hammers (small number) is put in the unit to start with to allow it as a reserve whip (chops almost certainly won't work). But if we are past our free unit level then I don't think we are normally going to be that exposed - putting aside the situation when all our units are away at war. But it will work in the very early game where you want to get some high value building early and do not want to be saddled with too many early military builds. This is essentially a gamble against barbs (or aggressive neighbours) and if the gamble does not pay off and you are attacked you have that insurance unit to whip out instantly.

If you're someone like me and like to play this sort of early game looking for turn advantage at the expense of some immediate security then I think this option can be quite valuable. I even tried it in a recent game and placed two Archers in temporary build in both of my cities. For the total cost of 3h, I had two units ready just in case and I eventually whipped the archer in my capital after losing 1h from the degrading. The other part-build unit "lapsed" without value but the real value was that extra security during the 10 turns that I had the unit there. In the meantime I had a chariot built anyway so the value of further "insurance" was much lower.

[Cort Haus]

Do you know what the hammer decay rate is, couerdelion?

[Velociryx]

Excellent analysis to get the ball rolling, Lionheart!

I realize that the question of civics and their interactions with upkeep went beyond the scope of your initial article, but based on your experiences thus far, I think that one could safely conclude that such a tactic could be quite helpful to those players running an early pacifism sling play, as the maintenance costs would be higher for them in any case.

I'm just guessing here, but as I understand it, there's no hammer lost for the first ten turns that a unit is in the queue, but not "first in line" and afterwards, the build degrades at the rate of one hammer per turn.

That being the case, one would not need to allow the unit (archer, or whatever) to totally fade from the queue, but periodically, you could "bring that unit to the fore" to give him a fresh infusion of hammers--and presumably reset the ten turn counter, although I have not tested this.

Further, if "occassionally" working on a unit DOES "reset the clock" so to speak, and forestalls hammer degredation on the build, then it is possible to keep the unit in a mostly built state indefinitely (since you know what your hammer outputs are, you could allow the build to degrade to the point where one turn of working on the unit later would still leave it incomplete, and then switch away from it to something else....

-=Vel=-

[Arrian]

This strikes me as something that could be realistically done in a democracy game or other PBEM game... where you have tons of time between turns and are willing to use it on stuff like this.

Never, not in a million years, would I go this far in SP. That's just my own personal line in the sand, though, YMMV.

-Arrian

[Velociryx]

Oh, no doubt about it...this is definitely not for everyone. It smacks of EXTREME levels of MicroManagement, and likely would only appeal to those players who were looking to squeeze out every drop of commerce they could to speed through those techs.

Truthfully, the only real use I could see for it would be in garrisons, as you would want to maintain an imposing standing force for either attack/aggressive defense purposes, but this methodology, especially when coupled with keeping your # of workers to the minimum you need to get by would likely go a long ways toward either freeing up "slots" for a bigger offensive army or for giving you just a pinch more commerce turn to turn.

-=Vel=-

[Dominae]

Re: Part-building units

Originally posted by couerdelion
There are therefore three reasons why we might start a unit build and leave it for finishing later.

1) We have nothing else to build now but will do shortly. It gives the city "something to do"
2) We are near to growth and want to use this extra population for some other build (eg worker/settler).
3) We need extra "insurance" so want to put a unit as part-built to allow us to quickly get one if needed.

I agree that these are the good reasons to micromanage partial builds; savings on upkeep is not one of them.

Another point is that in SP the AI bases its military decisions on how powerful you are. So not having units around means a greater chance of being attacked. Usualy this is a bad thing, but I suppose you could lure the AI into a bad offensive using prebuilds. However, in my games I find that my Power is less than most AIs anyway. If I'm at the threshold where prebuilds actually matter, my military is more than strong enough to deal with the AI, in fact I should be going offensive instead. The "baiting" tactic thus works quite well without prebuilds: you just need to have enough units to defend your lands, which is far less than the Power graph would lead you to believe.

[Velociryx]

I'm not sure I agree, but I admittedly haven't run enough tests to see if there's perhaps an opportunity for marginally faster teching.

It seems to me, however, that in the early game, when your total commerce is quite small, even a nominal savings on total costs (1-2gpt) can be compelling, especially if you've already popped, chopped, and otherwise rushed through every bit of infrastructure you need, and thus, have scant need for hammers for a given window of time, why not swap them out for a couple more commerce and get the next tech a turn quicker?

The folks who would stand to see the most gain from it would be the early game pacifist sling players (who will also likely be running a Philo Civ)...in such an environment, each garrison would actually be quite expensive, and while the peaceful land grab is in full swing, there's a window of opportunity to exploit some cost savings.

In larger, more mature Empires, and/or depending on the civics you're running, the cap (and maintenance itself, for that matter) is rarely even a consideration, but very early on, it might provide a few crucial extra points of commerce...I'm not yet sure, but it's...interesting.

-=Vel=-

[Jaybe]

As I play marathon speed, I presume the 10-turns-before-hammer-decay is game speed dependent. Also that map size doesn't change any of the results.

[ben04]

The problem I see is that you can only get one unit out a turn. Even if you spot a SOD early you will at most have 5-6 turns. This might not be enough to get those units finished so the enemy actually takes a city full of nearly finished units.

On the other hand if you can maintain a force large enough to take that SOD out it wont make a big difference anyway.

In the early game I rather have a realy warrior to go pop huts or a real archer to drive the power graph high so the AI leaves me in peace.

[Quillan]

I don't think I would keep a stack of "almost finished" units in the queue, as that would be WAY too much micromanagement to keep up. However, one wouldn't be too bad to manage. If you kept one of those units in each city, you could pop out a number of units in an emergency without having to pay maintenance on them. I might give that idea a try my next game.

The only problem I see is obsolescence. Suppose I have some archers hanging in limbo, and get to where I can build longbows. Do they update? Or do I have to scrap them and almost build a new crop of longbows, then repeat with musketmen, riflemen and infantry later?

[Common Sensei]

Originally posted by Quillan
I don't think I would keep a stack of "almost finished" units in the queue, as that would be WAY too much micromanagement to keep up. However, one wouldn't be too bad to manage. If you kept one of those units in each city, you could pop out a number of units in an emergency without having to pay maintenance on them. I might give that idea a try my next game.

The only problem I see is obsolescence. Suppose I have some archers hanging in limbo, and get to where I can build longbows. Do they update? Or do I have to scrap them and almost build a new crop of longbows, then repeat with musketmen, riflemen and infantry later?

I'm pretty certain a unit being built is upgraded before being finished if the proper tech comes along. When building a Musketman while researching Rifling, if Rifling is discovered before the unit is complete the build order shifts to Rifleman, and the build progress is not lost.

[zeace]

Originally posted by Common Sensei


I'm pretty certain a unit being built is upgraded before being finished if the proper tech comes along. When building a Musketman while researching Rifling, if Rifling is discovered before the unit is complete the build order shifts to Rifleman, and the build progress is not lost.

If that's the case then there's definately some good times for this... Your workers are 15 turns away from connecting your city to your copper mine, but you might need a second (or third) warrior garrison to hold off the raging barbarians. Build ~10% of him, and if the barbs come, pop-rush finish him, otherwise wait until the city has copper and all that production moves to an axe. And the whole time you don't pay maintenance on your "extra" garrison either.

Still, this level of MM is insane...

[Quillan]

I have had "in production" units upgrade before completion in the past, I am just not certain of the circumstances that trigger it. I have had other units that didn't. I think (not certain here) that first the game finishes all units that would complete in the turn, and then if not changes the build to the upgraded unit. If I'm right, then an archer that is one turn from completion wouldn't change into a longbow, it would finish as an archer, while an archer 2 turns from completion would become a longbow with potentially 2+ more turns to go (due to increased hammers needed). I have no idea if it affects things not currently at the top of the queue.

[Swiss Pauli]

I think the queue realted 'upgrade' occurs when all possible upgrade options become tech enabled. Thus, a longbow will be switched when rifling is researched (and iron is available), but an archer only when you have machinery (and iron) and feudalism.

[Velociryx]

That is correct, and Warriors upgrade to Spears if you're in the process of building and hook up copper.

-=Vel=-