How optimum is optimum?

[atreas]

For those that haven't already read it, there is a detailed article (http://forums.civfanatics.com/showthread.php?t=154872) about the method for optimum early growth, where it is stated that the sequence Worker-Worker-Settler is the optimum one in conjuction with BW and chopping. I have some objections about the GENERAL validity of this statement in many cases, especially in epic or slower speeds and higher difficulty levels, which I would like to post as a source for discussion. Of course, a need for reading the above article is both necessary and extremely useful to ANY civ player - note that this article is stating that is focusing on Prince level; so in most cases I am not objecting it, but just discuss other levels and speeds.

1. Only 100% science please

There is a many times "hidden" precondition that you run science on 100% during all the time to BW. At the article it is stated that

"Epic BW=21, W=19 turns" (in epic you discover BW in 21 turns and produce a worker in 19 turns). This is false for the majority of civs in higher levels - the truth is that without an organised civ you need 28 turns (due to 80% science) or, with a binary tech search, 27 turns (if I didn't miscount - I get bored with this up-down thing).

So, the effects described can be achieved in lower levels and organised trait, but not necessarily in higher levels and other traits. When you have to lower your science slider the result is that you end up with a worker that has nothing to do except wait or mine. If at this moment you start a second worker, you can't use chopping (no BW yet) so the whole article loses its validity. If you switch in the meantime to anything else (like a warrior), then it would be probably better to have started first with the warrior and maybe (depending on the "lag") grow your city to size 2 first (usually the grow needs 11 turns in epic - well, 11 is not far from 9 and the increased city size will speed up the worker creation). So

#1. If you have to lower your science slider, then probably the "worker-worker-settler" sequence isn't optimum, because of the "delays" in techs.

2. Not hills/plains please

The article is based on an assumption about the city tile, that:

"A size one city has 3 free production"

if you "translate it to civ", that excludes a city which is built on a hills/plains AND has into the radius any 3 food tile (like corn, floods, etc.) or even a forest tile, because in this case the city has a different base production, according to the article terminology. This case is not at all rare (note that many players tend to build always on a hills plains tile) - and creates again similar problems with the case #1 - very fast production of a worker that will have nothing to do. To "ease" a bit the numbers, I will now allow even 100% science, and we have:

Epic: BW 22 turns, W 15 turns (7 turns lag).

Again, what is clear is that the proposed strategy cannot be applied "fully" and the numbers given in the article cannot be trusted for this case. So

#2. A hills/plains city tile may change much the strategy that you will follow, especially if you have also to lower the science slider.

3. Starting techs

Of course, the proposed strategy has obviously no meaning for the (not few) civs that don't start with Mining (longer way to BW). But we have already seen that there are many cases where you end up with a worker that hasn't the right to chop - so if he is to do something constructive, he must MINE a hill (if there is a hill available, that also doesn't have a tree on it). Still, this is the "somewhat worse" case D of the article (unless you happen to mine the "correct tile" that will happen to have bronze - if there is such a tile in the first city). There is only ONE civ in this game that starts with the two "best" techs for the proposed strategy (Agriculture+mining), so as to be able also to improve a +3 tile: the Chinese. So:

#3. To get to the "equally good" case C of the article (improve a +3 resource), you must either be Chinese, or be quite lucky with the huts.

4. Commerce

This article states specifically that it doesn't deal with commerce: the proposed strategy is obviously giving worse commerce results than any "growing city" strategy, since the additional tile would give additional commerce (and thus beakers). But how much? Of course, this depends on the tiles you have in the city radius, but with a hills plains city and TWO "3 food" tiles (floods/corn/wheat, etc) (not the most unusual in the world for the capital) you can easily verify that the sequence "warrior+grow city+worker+worker+settler" (and to make it easier in 100% science!) is ONE turn later than the proposed strategy in production but can be at least TWO turns ahead in commerce (if the tiles have also commerce), plus with an increased city size at the end (more profit to come in the next turns). If we are forced to lower the science slider, this strategy could be the winner by far. So

#4. If you have a city with potential commerce+food tiles, then even with 100% science there may be alternative tracks that loses a bit of production in favor of science.

5. Overkill of trees

This is, IMO, a strategical disadvantage of the proposed method: it counts only the "current" production achieved but doesn't count what you have lost. Any tree can be chopped at any time: but timing is crucial to get the maximum effect. If we have in the one side of the equation the "assets" (trees) and in the other the "achievements" (produced units), we see that we just transposed our assets - well, that's not a production boost, since it can be done at any time in the future. As correctly is stated in the article:

"You can already see from the above that starting a second city adds much more production than growing the first city, and that the best improvements are almost as valuable in the short run as founding a second city. Normal improvements increase total production modestly, but only on certain tiles."

If we take this thought one step further, using an old chess thinking method about "necessary moves in a plan" we see that the second worker just speeds up a little the conversion of an asset (tree) to a settler - well, how necessary is this? Of course the first worker is crucial, but the ideal would be to have a city that helps - through the improved tiles - and not wait everything from chops. I have the feeling that the second worker is just an overkill of trees - most probably because I never had a problem with waiting a bit before his production, in order to minimize the "chops without a contributing city". To see the difference that a contributing city may make, see #6.

Plus, the missing trees could be used in many ways, and since the first city is usually a "rich" one I generally prefer to chop the second worker from the SECOND city - that way it will leave me some trees for a possible needed wonder (or even an extra axeman). So

#5 How necessary is really to chop IMMEDIATELY a second worker? Can't it wait a bit until city grows in size and/or have some nice improved tiles, so as to preserve some trees (assets) for other uses?

6. The "special cases" - welcome to the fishermen and goldminers club

This is where my quest really started - I happened to have a city in hills/plains with corn into its radius and also a fish resource - and the best of all, I was playing a civ that started with Fishing! Out of curiosity I started to count (I gave myself also mining, to be "a fair battle"). It wasn't even close: starting with a work boat and then worker+worker+settles was equally good in production (I had a "contributing city") but was FAR better in science - PLUS at the end of the count I had a city with size 2 instead of a city with size 1 (more commerce and production to be won in the future). Of course this is a special case, so I counted again with a clam resource: now the proposed method was a bit faster (2 turns) in production but FAR worse in commerse, science, and future prospects. Another advantage of the work boat: it eliminated all "lags".

One more "exception" i have found is the "gold" tiles: of course, they provide SO MUCH gold that even with other things equal it's useful to deviate a bit and mine them first - even if you didn't have to, due to the level "lags". So

#6 There are also some special cases where the proposed strategy is clearly inferior. One such case is the sea resources for the fishing civs, and another is the gold tiles, due to the commerce you lose by not mining them ASAP.

7. Marathon

In most of this I focused on Epic, but Marathon shows even bigger problems: as stated in the article about the turns needed for a worker/BW

"Marathon BW=49, W=30 turns. At Marathon/Deity, it takes 67 turns to get bronzeworking"

You can imagine yourself that 37 turns without something to do is clearly NOT optimal. Again, only Chinese would be able to farm anything - so the problem is just aggravated. I think that in Marathon and higher levels this strategy doesn't have a chance at all (after all, the numbers can't even be applied due to the "best case" 19 turns lag for BW). Note that, for a city with +3 food, it takes 22 turns to grow in size - so now a good alternative is to grow first the city (build a warrior) and then build a worker (due to the increased city size, it will need less than 30 turns, so it will come approximately at the time of BW). So

#7. In Marathon, the proposed strategy can't be applied.

 

[ownedbyakorat]

I've been finding some success in using the first worker to chop a settler rather than a second worker... seems to get me started fast, especially when copper does not show up in my home city radius.

 

[Dusty Monkey]

As for #5 - every turn that you dont have that second city but COULD HAVE had it is a turn where you LOSE production that could have been gotten from a second city. Its a rare day when an additional population point on the 1st city will cover that loss. There is no make-up test at the end of the game where you magically get those turns back.

 

[atreas]

As for #5 - every turn that you dont have that second city but COULD HAVE had it is a turn where you LOSE production that could have been gotten from a second city. Its a rare day when an additional population point on the 1st city will cover that loss. There is no make-up test at the end of the game where you magically get those turns back.

Exactly: that's why it's suboptimal to build a second worker - building first a settler BOOSTS production, when by building a second worker first you are just transforming assets. It is stated in the original article also - the earlier second city is obtained by building a settler after first worker.

EDIT. I think I must rephrase the original post because it is not clear as I wrote it. Second worker speeds up the CHOPPING of the settler, not the settler itself.

 

[DaveMcW]

What good is it to have a 2nd city with no improvements? The point of a 2nd worker is to immediately improve (chop) around the 2nd city.

The reason you go Worker,Settler rather than Settler,Worker is the worker can help chop the settler.

 

[atreas]

What good is it to have a 2nd city with no improvements? The point of a 2nd worker is to immediately improve (chop) around the 2nd city.

The reason you go Worker,Settler rather than Settler,Worker is the worker can help chop the settler.

The reason is that you save trees for later choppings if you can combine chopping worker with a city working on improved tiles. More trees = more future chops.

 

[DaveMcW]

Why should I care about future chops?

Most valuable things to chop:
1. Settlers
2. Workers (to chop settlers faster)
3. Axemen (to defend the settlers)
4. Wonders (when I can't afford any more cities)

 

[atreas]

Was that an argument? Because the answer could be that all three things you mentioned ARE future chops - unless you intend to play with only 2 cities and 2 workers. Plus, we have different ideas on what is good to chop and what isn't - but that's not the point of it. The point is that IT IS SUBOPTIMUM - it's exactly case C in the original article.

EDIT: Since now the 3 items have become 4, I have not a big objection on the list of "chopping items". Still, what I am trying to say is that, if something can be done with a chop less, even with a cost in turns, the result of the calculation isn't "sure" - the only thing that boosts production is a new city, all others are just transformations of one kind of assets to another, with the timing being quite subjective. But it's very difficult to calculate it until the age of wonders to see exact numbers - so I just give my estimation.

 

[maltz]

In higher difficulties and larger maps, indeed it will take much longer times to get BW. However, the earlier one is able to chop, the better chance to survive in the game.

So I would quene a warrior first, and try to make the pop. grow to 2, then switch to worker. By the time the worker is done, BW should be ready.

When one gets extremely lucky (or extremely patient), a goody hut will pop bronze working!

 

[Helmling]

This is what is great about Civ4, gentlemen. There is no single way to proceed. Everything in this game is contingent. You have to adapt and roll with the circumstances and with your overall goals for the game.

In some games, the worker-worker-settler route may be most effective--perhaps even most games. But if you cling slavishly to one strategy, you will be vulnerable.

 

[opensilo]

atreas, I think you make some very valid observations. After reading that article, I started a few games with worker-worker-settler.

I quickly realized that while good in theory, there were often times (on emperor and with a non-organized civ that didn't start with mining), that it was much better to start with a warrior. The only a starting worker could have done was to mine a bare hill, like you say.

Also, one of the problems I kept running into was early barbarians interfering with my workers and my settlers. Having an extra warrior is very helpful to keep the fog of war at bay.

I realize that the second pop point tends to add only a single "value" because that pop needs two food to sustain itself, whereas a new city gives you 2 food, 1 production (or two production if on a plains/hill) and a commerce coin. Still, by queue switching, you can grow your city while working on a warrior (and chopping worker or settler) to size 2 to get (usually) and extra value point.

atreas mentions starting with a workboat for seafood resources or working a gold mine if that's available. I'd also like to add an anecdote. I was very satisfied with researching animal husbandry quickly (started with one prereq as I recall) when my capital had both cows and pigs in the fat cross. If you don't have many trees, or want to use them to chop a wonder (perhaps itself a suboptimal play, admittedly), then workers who can only work on mines and chopping aren't all that useful.

 

[atreas]

Good points. Note that the real value of the extra population isn't so much the extra production, but the possible extra commerce. It clearly depends on the tiles you have around your first city - I used a simple pattern of a city with a river, but maybe if it was a city with a lake then the results would be different (the idea is to have both citizens to work on tiles with commerce and 3 production).

As for the queue swaping, I didn't mind checking it because I think that this is a technique you can use in BOTH cases (so I have no reason to believe that it would alter anything).

I believe that the best conclusion was said by maltz - the goal seems to be to search for BW and have your first worker coincide with BW (if you can, in the meantime grow also the city). This looks to me like a very good general rule, with the application depending on the specific case.

 

[Selous]

I believe that the best conclusion was said by maltz - the goal seems to be to search for BW and have your first worker coincide with BW (if you can, in the meantime grow also the city). This looks to me like a very good general rule, with the application depending on the specific case.

nice artical .... i agree with it mostly, most games i build warrior, worker, settler while making a bee line for BW but sometimes the game dictates different. if you have a goldmine mine that puppy and get mining (and then BW) but if you start with pigs and cows (or wheat ectect) then i gofor agriculture first

anyways, to the quote, if you beeline to BW and have at least 1 worker when you get BW that is optimal

 

[East St Trader]

Circumstances alter cases.

I doubt there is a clear cut sequence that would always be right. And I doubt that the resources available within the city radius are the only relevant variables.

What is going on with neighbours may well be very significant. As, for example, if there is a close neighbour and some plainly critical nearby resource - copper for the sake of argument. That may make building a settler quickly - or perhaps a warrior or two to cramp the neighbours style - critical.

Hut outcomes I suspect would matter.

Which is not to say that having a sequence or two that optimises productive capacity for a set of variables based on terrain is not a helpful tool. So I am glad to have read the thread.

I just don't really expect to follow a set pattern of early builds too often in actual play.

 

[alatari]

We are talking about ohioastronomy's thread on optimum starting strategy and his preference is worker/worker/settler through forest chopping.

I only play marathon, huge maps at diety level so it is of no use to my games and especially this last game playing the Inca's in South America on KREarth's Huge map. Not a forest to be seen within 20 tiles of the starting location. Jungle as far as you can see and about 700 science units till bronze. No plains or grasslands with river access either. Just desert and hills.

The only strategy that seems to work for the first 40 turns is to build a settler immediately with no defenders. Animals won't tread into your cultural influence and the last couple of tiles is a risky run to the settlement site which is on stone quarry.

Bronzeworking is incredibly important but so is founding a religion and getting exploring boats out to meet other civs so that future trade of techs is possible. While you maybe playing at diety levels the AI's are playing at noble level...

Is there a good thread to diety player starting strategies?

 

[drkodos]

1. The original piece was composed before the 1.61 patch that nerfed chopping.

2. It is still a very good startegy (Pun intended) on a majority of starts, Monarch and lower. It does not work in every instance, nor did the author claim it does.

3. There is no panacea. There is no spoon, either.

 

[Stolen Rutters]

I can easily see how chop rushing was superior before v1.61.

I go worker-settler and steal additional workers. I try to limit chops to only speed the first settler, since the forest spreads when I leave it alone. Sometimes, I even get way more than my original forest back. That way, there are many more forests created for wonder builds in the future. Clearcutting destroys regrowth potential, and gives me no "forest trails" for my early woodsman II "Anti-Barbarian Defense Force" and I can usually whip more effectively than chopping before mathematics, anyway.

 

[alatari]

Got to love that woodsman zone defence!

 

[Eqqman]

Is there a good thread to diety player starting strategies?

For the experts, yes. For the rest of us, I believe it's 'cross your fingers and pray!'

 

[cleverhandle]

(Cross-posted to the OT in Strategy Articles)

OK, I'm probably kicking a dead horse here, but I've been away for 6 months and this is the first I've seen of this particular analysis. Seems interesting...

I spent the last few hours analyzing a fairly run-of-the-mill Peter start on Emperor (pic, game). In general, I think ohioastronomy lays out a useful way to analyze a starting sequence, though I disagree with him on some points. Mostly, it's the stereotypical "every game is unique" thing. This is especially true in his valuation of worker turns - valuing them at (chop value)/(chop time) is naive, if not disingenious. There's really no way to quantify the value of a worker turn without seeing the start in front of you, hence my analysis. Starting techs and traits also play a big role.

That being said, I did notice some interesting things that should apply to many "typical" games (no religion race, fish, gold, or bizarre terrain). The Excel file for these numbers is available here - corrections warmly invited...

1) Pop growth is bad. Nah... just getting your attention. What is bad is waiting to grow to 2 before starting on a worker if there are decent (+2 or +3 power) improvements available. The naive Grow-Worker-Settler plan (I know I started out doing it) gave a pitiful normalized production of 162 using 2 trees and had terrible worker position for improvement of the second site. Unusual commerce situations aside, this path really screws everything up.

2) Pop growth is good, if done intelligently. Getting a worker out first and carefully timing a "growth intermission" so that the second pop could begin to work a mine immediately after it was born created what were arguably the best scenarios. Worker-Worker-Settler produced a normalized 246 using 2 trees and Worker-Settler-Worker produced a normalized 269 using 2 trees but with a worker disadvantage. Furthermore, the worker position in Worker-Worker-Settler was perfectly arranged for one more chop without wasting time - that would put it at an impressive 266 production at the cost of 3 trees.

3) Stay away from clear-cutting unless your resources really suck. The extreme clear-cutting case produced a normalized 232 with excellent worker position but at the cost of 4 trees. Mixing in a good special and intelligent growth gave better results with less chopping.

4) A settler before military is not suicide. A settler sent out at turn 34 was quite safe from barbs (checking World Builder) on a standard Emperor map. He will need animal coverage, though. How to handle that will obviously depend on the map and scout situation. Working a single warrior into the queue is only a 1-2 turn delay for any of the situations if they are planned correctly. Several of the Worker-Settler-Worker scenarios (including the 269 production one above) get a Settler out by turn 32 or earlier.

All in all, there's no substitute for getting in there and crunching the numbers for a particular scenario. Unfortunately, that's pretty time-consuming, and general rules only take you so far...