[GS] Rethinking Lumber Mills

[knighterrant81]

I can be unimpressed with poor presentation and overly narrow conclusions without being frustrated.

If the intention is to provide an analysis and conclusion that only applies to a very narrow playstyle (speedy victory on Deity), then by all means make that clear in the initial post. That way I'll know not to bother discussing the issue because I am rather indifferent to that style of play.

Either way, I appreciate your effort to clarify.

Its fair to say that these threads don't do a good job of advertising themselves. We probably need a separate theorycrafting forum for Civ. Usually when Lily posts, you know this is the kind of thread you're getting into. Again, he's been making posts like this for years so we usually know that's what's going on but it wouldn't hurt to have a trigger warning on these threads

That said, I'd encourage you to read them with the attitude of "what can I learn from this" rather than indifference - just because you aren't trying to win in 150 turns doesn't mean you can't learn that Chop is almost always better than a Lumber Mill.

 

[Victoria]

So you shall chop now.

I have been pondering this some more.
A chop appreciates in value at the same speed as inflation as inflation and chop are based on the highest of science/culture.
So chopping now vs chopping then is no different in value ignoring what is produced and when.
The real values are the production loss, cost of a charge at the time and what you are chopping when.
If you are working a 2/2 tile and have to still work it after a chop there is a disadvantage it chopping now rather than later (without considering anything else)
Considering that the inflation of gold/faith is higher then a builder charge cost is likely less but certainly less at around Liang/Pyramids/Feudalism (any or all)
So apart from what you are chopping there is a fine point around feudalism in particular but certainly now now if now is before these times and certainly not before Magnus.

So as lily hinted, chopping to defend yourself is all important and will always be a chop now unless it was not necessary/very useful.
If we take the example of a library
Chopping a library may be a disadvantage now is science starts outrunning cheap district placement before you can get settlers out (if not working on district discounts)
Chopping a library now gives +2 science but will unlikely finish the library with one chop while chopping then will and possibly also donate to your university but more importantly if you have 1-2 CS up by then your library will be worth +4 +6 or even +7 with Hypatia. So the small loss of +2 per turn if not needed for Dom at the time may appreciate more than inflation considering the builder charge and the loss of worked tile production.
What about Settlers? Fantastic value if you have a place to settle them, the earlier the better but they inflate in price at an alarming rate meaning too many too fast (before early empire and then before Ancestral hall) they just do not become worthy of a chop or do they? This is much harder but chopping in some is worth it but how much wood do you have?, how many builders? and more importantly it is not worth it to chop in builders early. So chopping in settlers is limited by how many, builder availability and chop resource availability. Would you use every last chop in a city for a settler? Why not?

If we also consider that inflation is not flat at all. It is a curve starting slowly where too much science culture is limited and large amounts are hard to get and increases rapidly as we get to the end of the game. Also we tend to not finish techs/civics as already mentioned to keep inflation down for district placement. We get back to the choppity chop thread @Sostratus provided which pinpoints the mid game as the place to chop. Ideally falsely inflating by finishing small techs or civics to provide and inflation boost to chops.

Chop n improve is another feature. If you are not going to have a very large empire (your choice) is chopping and improving better than just chopping? Changing that 2/2 forested hill into a 3/2 or a 4/1? How much does that 1 production appreciate over the game? It seems like it appreciates by thousands. Worth a charge if you can afford one?

So... you shall chop now is not so simple to me but perhaps I am just confused. It seems this thread discards chop appreciation and I certainly appreciate chops.

 

[DanQuayle]

If you have a piece of forest, chop it or place lumber mill?

Before I tend to think this as no problem, you shall always chop. However after calculation I start to challenge this conclusion on online speed.

Build Lumbermill: +2 production, +1 from trees, so actually +3 production per turn altogether. In late game you can still remove and chop. +1 from Steel and +1 from cybernetics. +1 appeal and becomes +2 after Conservation.
Chop: +production right now. Also clear a tile for your district/wonder/other improvements.

Suppose you get X production from chopping now, and Y net prod(consider the builder charge cost ) from chopping in late game. There's T turns between now and late game. Suppose value inflates at t% per turn.


Simpliest Model:

Your profit from chopping now: X (suppose you don't build other things thereafter)
Profit from late-game chopping: 3*(1-t%)/t%+Y*(1-t%)^T (ignore bonus from appeal or steel)

On online speed suppose X=40, Y=60 and t=7%, T=50, then you get 41.45 from building lumber mill but only 40 from chopping now. So you shall build lumber mill if your citizens do not have a better place to work on.

On standard speed suppose X=80, Y=120 and t=5%, T=75, then you get 80 from chopping now but only 59.56 from lumber mill. So you shall chop now.


Conclusion:
On online speed, if your citizens do not have a better place to work on, you shall build lumber mill instead of chopping right now. Surprisingly this yields a little more than chopping. If they do have better tiles to work then you shall chop. Keep (rain)forests for district adj or Earth Goddess is not always inefficient.

On standard speed you shall always chop.

The formula you are using is derived from the financial formula of the present value of an annuity which assumes that all "cash flows" can be reinvested at the same constant interest rate. This is however not the case here.

All the periodic "cash flows" (the +3h from lumber mills each turn) do not yield any interest. Getting +3h every turn for 10 turns is the same as getting +30h after 10 turns. The only benefit you get is when you complete the production of a building/district/unit. As these are linear increases there is no compounding effect (which is the case when reinvesting money in the context of the original formula).

The use of this formula is thus flawed in this context and does nothing but translate the general idea that "getting stuff earlier is better than getting stuff later" by discounting later values.

Furthermore, you need to take the differential production yield into consideration.

Spoiler :

While +3 might be a good approximation for forests on flat lands for a large portion of the game (ie post-construction and pre-Steel), it is not the case for forests on hills. (That is when you do not have access to a UI that provides production.)

A plains hill with a mine post-industrialization is 1f5h which is the same as a lumber mill post-construction, but pre-Steel. The difference between placing a lumbermill on a forest on a plains hill at construction (1f5h) versus chopping the forest and placing a mine over the same plains hill (1f3h) is +2h and the cost of a builder charge. As soon as you reach Apprenticeship, the yield differential becomes only +1h and when you reach Industrialization, it becomes zero until you discover Steel.

For a proper evaluation, you need to divide the game into segments and evaluate how long each will be: Start-Mining, Mining-Construction, Construction-Apprenticeship, Apprenticeship-Industrialization, etc.
You also need to take into account that you can skip Construction altogether and go straight to Industrialization.

To properly model what is going on in the game you need to do properly-modeled simulations with staircase/jump functions.

 

[Fluphen Azine]

I have 1378 hours played on Civ VI. I may not be in the 2k club, but second game?

Many players play 1400 hours in 140 days or less!
Just for some perspective.