How long can you delay founding City 1

[couerdelion]

I think many of us have moved the settler from their original starting position in order to get a better one. One simple objectuve that is often considered is moving to a nearby plains hill to acquire a value extra hammer in the city square. I thought it would be interesting to consider how long we might spend wandering around before the delay in founding the city costs more than the benefit gained.

The comments here are based on epic speed so times for other speeds should be adjusted accordingly. Since this is largely a theoretical discussion, I would also add that the considerations here are more designed around a civ with Hunting as a starting tech which allows for the scout to give valuable support in this early search.

My approach has been to consider the value we have at start up which comprises surplus food, hammers and commerce. I will also work on the principle that we are trying to optimise growth in our civ and that, at this stage in the game, a 1.5% growth rate might be considered satisfactory.

Food, hammers and commerce will all be converted into a “gold” value at the following rates

1f=2h=3c=3g.

Let’s also ignore the question of what other resources are there in the city site. This factor is just too variable to make any meaningful analysis but it goes without saying that you would have to consider this. Switching from a city fat-cross with corn/gold/ivory/rivers/floodplains/forest to one with desert/incense is plainly nonsensical but I will leave it to the reader to judge those issues themselves.

Let’s suppose we start with a basic city tile (2/1/1) and assume that the other tile we can work is 2/1/0. From the start the city is generating +2 food, 2 production, and 9 commerce for a total value of 18 gpt (2f+2h+9c=6g+3g+9g=18g)

The “alternative” site plains hill and also provides a 2/1/0 tile to start with. The value of this starting position is 19.5 representing an 8.3% increase on our first option. If we can found this city within 5 turns of the original start we will have created an average return over our optional start of 1.5% p.a. If it takes longer than 5 turns we are not getting enough to justify the move.

If the “improved site” is more dramatically enhanced 2/2/9 to 3/2/14 (settle plains hill with wine by river/work oasis tile/financial civ), then you can spend 24 turns to get here!!! - But I am not recommending this sort of delay since your settler will certainly become food for the wild beasts if you spend that long settling. But it does indicate that delaying for longer periods can be justified.

[absimiliard]

I'm not convinced 1.5% adequately represents a typical return I would find acceptable.

I realize some game-long calculations have come out near that. Stuff from the discussions of discount rates and inflation seem to support that contention.

I can't help feeling that the very early game is different though. Consider as a possible measure of acceptable growth rates the time/food needed to grow from size 1 to size 2 in your capitol. If I work a typical 2f/1h tile I'll grow in 10 turns, yes? So 1.5% seems a bit low to me.

In support of my assertion that 1.5% is too low, though not particularly supporting my growth-based model, I'll point out that most people would think anyone insane for waiting 24 turns to found their first city, even if they could found on such a wondrous square as you've described. The length of time is, frankly, outrageous.

In fact the disconnect between what the math said and what my gut said about 24 turns was one of the things that made me question a 1.5% growth rate.

Other than that, looks solid.

-abs

[Arrian]

I've never delayed more than 1 turn.

-Arrian

[Quillan]

Honestly, I don't like the mathematical values assigned to different things used in comparisons like this. It just seems too arbitrary to me. However, I have to agree with Vel on this one; the initial turns of the game are the most important. The period of time from the start of the game until the barbarians appear is the make or break time for the game. If you delay founding the city at all, even by one turn, you've cut into your research. Losing research puts you behind the curve in whatever you're going after. Trying for a religion? That lost turn or two could mean the difference between founding it and someone else founding it. Going for early scouts to get more huts? Lost turns means fewer huts for you to grab. It means lost turns of growth on your city, it means lost turns of production for your military, and your culture.

This is just a gut call, but I only delay founding the first city for a VERY important reason. If my initial scout shows a much better site right nearby, I'll move. I'll take a plains hill if there's one there, but 3 turns is my absolute limit.

[couerdelion]

Originally posted by absimiliard
In support of my assertion that 1.5% is too low, though not particularly supporting my growth-based model, I'll point out that most people would think anyone insane for waiting 24 turns to found their first city, even if they could found on such a wondrous square as you've described. The length of time is, frankly, outrageous.

I agree that 24 turns is insane but wanted to demonstrate that, in this particular case, you were still able to produce a 1.5% increase in output over the period.

Be careful though because I am talking epic speed here. That 24 turns represents just 16 at Standard speed and the 1.5% rate would be 2.25% on standard. I tested the 1.5% simply by taking a standard city tile two more 3/0/1 tiles and seeing how much the output grew during the 11 turns when it gets to size 2. The answer was 1.64%.

There are some tactics that require a quick start (Early Religion is perhaps one) but even then, either production of commerce might be quickly recouped if the timescale of the initial play is a little longer than 5-10 turns.

[absimiliard]

*edits away almost everything he wrote*

Your numbers work. But my gut still says the cost of lost turn-advantage is just much higher.

I wonder if my gut, or the equation, is wrong?

-abs

[Kuciwalker]

The cost of lost-turn advantage in the very early game is FAR higher than you're estimating.

[Anath]

I'm very new to this but to me it appears that moving your settler for 2-3 turns is ok. In my current game I had a crummy position to beging with so I hopped away 2 tiles and still Im ahead of the AI atm. Admittedly Im playing at a very low difficulty setting so maybe I should be quiet.

I support not taking the first best place that comes across even though waiting more than 5 turns prolly mean a barb or a beast just made sausage out of you.

[rah]

In an SP game 3 turns max.
In an MP game 1st or 2nd turn max

[couerdelion]

I've realised one mistake I have made in the calculations. The undelying asset that we have on founding is a combination of food, production and commerce (or science and beakers). I've placed a value on these by looking at the output - albeit using a average rate of growth - and for simplicity left the number as a turn-based measure rather than as a measure of the value of the asset itself.

What I missed was the asset that would be generated in period in which the city was actually founded - the accumulated production, population growth, technology. For short period, I can't see this having a great impact but it will certainly change the 24 turn figure.

With regards to "opportunity cost" the term, I believe is used far too freely and I don't think fully understood. Turn 1-10 opportunity cost really only has a bearing if you are trying for an early religion. In this situation, even one turn may cost you. But beyond that you have to be a little clearer about examples of what you are in such a rush to do that will be damaged by a 10 turn delay in founding a much stronger city site.

[trev]

I almost always move to get a plains/hill or similar benefit, I think 2 moves on quick is acceptable ranging up to 10 moves on marathon (This amount of moves can be made up by increased speed of first build that a good site gives - usually worker or settler)

[Blake]

At normal speed i would think that 5 turns would be often fine.

For example doing settler first, with a standard start the city produces 4"hpt" and the settler takes 25 turns. If you moved to a Plains Hill the city would be 5hpt and the settler would take 20 turns. So you can move for up to 5 turns and still get the 2nd settler out at the same time. True, you lose 5 turns of research, but for every turn after turn 25 you also have +1hpt in your capital.

I think going walkies can definitely be worthwhile if your capital is sub-par and moving to found on a plains hill is definitely a good idea (all other things being mostly equal).

[couerdelion]

I redid the calculations allowing for the “invested assets” or food, production and beakers of the early game and was surprised to find that it had almost no significant effect. The reason is that the compounding effect of the calculations assume that this is being done already so you’re getting the same effect.

The one factor that can change things a little is the discount rate. If the rate is changed to 2% per turn (ie 3% on standard) then the more extreme example above cannot be justified by spending more than 18 turns moving. That’s still very high!!

So we can adjust this one factor, all you need to do is figure out how you are going to build to generate more than 2% per turn in this early game.

Using a 2% rate moving first cities are justified if

1) Basic city tile with FP to Plains Hill with FP (3.5 turns)
2) Basic city tile with 2/1/0 to Basic city tile with FP (6.5 turns)

It’s roughly 2.8 turns (1.9 on standard) per additional point of output (ie commerce=gold). You can use this to do the sums with your own measure of relative hammer/food values. But if you start with at least an FP tile, you would only give up 2.4 turns per additional unit of commerce.

Having generated the “theory”, I intend to run a few test cases to see how this plays out – this will mean putting on hold a small renaissance/industrial age war that I am currently waging against the Vikings but no matter, Nidaros is not going anywhere fast.

I will try this out in the unforgiving environment of an Immortal level game since by 0 AD you will have a fairly good idea if the plan has worked or not. With any luck I’ll get the chance to go wandering for 10+ turns, build a great city next near to one of my neighbours and save time with an early rush. Possibly Genghis or Cyrus might be a good on to play with but I will definitely pick a civ with Hunting despite the fact that settlers can also pop huts

It might be interesting is some others tested this theory closer to the “edge”

[absimiliard]

Okay, now I'm more excited.

I'll work on the discount rate assumption of 3%, on Standard. My gut tells me it still seems low, but it's when your gut and your theory conflicts that interesting discoveries are made.

So by your numbers we have about 2 turns per cpt we can add to our starting location. That's very interesting and strongly implies that my instinct to found quickly may be in error if I'm not chasing an early religion. (and I since I run all random games I'm usually not playing someone who can chase an early religion.)

I'll give it some testing tonight if I can. See how it shakes out.

You may have just fundamentally altered my early game Lionheart, depending on how the testing goes. Nice job.

-abs

[rah]

But it's just like civ II yeah, if you delayed you could get a city to make up for it quite quickly, but there is no guarentee that you'll find better land.

If it's just a few turns, I'll accept that you have another unit out and you know your target, but 10-18 turns, how do you know where you're moving and is the grass really greener where you're moving.