Granary vs Settler vs. Worker

[Raliuven]

Granary vs. Settler vs. Worker Test

This question seems to come up rather often. Do you want to push out a settler right away or invest in a granary? How about adding a worker to your Civ right away to make sure the tiles are improved and Citizens are happy and productive? Is that a cow I see in the fog? Should I round it up before it wanders away? I hear that farming wheat is what the cool kids are doing! Maybe we should head that way!

The main goal of this post is to explore the strengths and weaknesses of some of the various ‘starts’ to a PTW came. Sorry, I don’t have conquest so that is not included here. So no agricultural trait!

The information in this post is not to suggest that other starts are not valid or do not have a purpose. Your planned VC, map settings, barbarians, etc. will all dictate a different approach to the game.

I have designed a series of tests to gather some raw data on the subject. The first step was to create the map. The base template is a small island on a huge map for the human player and an even smaller island for our rival Civ (1 opponent).

1. I selected Persia as the rival just because we hate each other so much and I find satisfaction in the fact that his immortals can’t swim. The opponent Civ really makes no difference because we will never meet.
2. Barbarians are turned off. The island is ‘safe’
3. The template map has no luxuries and no strategic resources. The focus of this test is to get raw data on REX
4. The template map has no additional resources (food bonuses) to begin. Food bonuses will be added to the map in the following order.

a, No bonus. This is the baseline. [NFB]
b. Food bonus out of reach of 1st city placement, within 2nd city placement. [NFB1stradFB2ndrad]
c. Food bonus out of reach of 1st city placement, within 2nd and 3rd city placements. [NFB1stradFB2nd3rd]
d. Food bonus in reach of 1st city, but nowhere else. [FB1stradOWNFB]
e. Food bonus in reach of 1st and 2nd city placement. [FB1st2ndrad]
f. Food bonus in reach of 1st, 2nd and 3rd city placement (party time!) [FB1st2nd3rdrad]
​

5. Each map will be played by two Civs for the first 60 turns, ending in 1525 after all actions are taken but before moving to turn 61. The Player civs will be the (PTW) Iroquois (religious/expansionist), and America (industrious/expansionist). I selected these Civs because I wished to see how industrious civs compare to non-industrious [NI] civs. I chose expansionist so that we start with the ability to make a Granary. I might also expand the test to compare commercial civs for the reduced corruption effect.
6. Each of the Civs (Iroquois and America) will play each map 3 times (total of 6 tests per map). Though in no particular order, the tests will be:

a. Granary first. [GF] Every city must produce a granary before producing a settler. I did not go so far as to say it must produce a granary first. Some cities, particularly those founded late in the test may produce a warrior instead.
b. Settler first. [SF] Every city must produce a settler before producing a granary. Same as noted above.
c. Worker first. [WF] This is test the theory that an early worker first will help growth. The first city must produce a worker before producing a granary or Settler. Thereafter, the goal will be to produce enough workers to ensure that all citizens will be working on improved tiles.
​

There are a few more ‘rules’ I have imposed.
1. The first city must be settled in place. No moving, even if our start location allows us to see a bonus food source. It just makes the mechanics easier. Besides, we are already on a hill and a river. Rome started the same way; it should be good enough for me!
2. The initial scout must be moved out of the city and disbanded. No Scout for you!
3. All future units will be allowed to scout the island. An “island completely scouted” will be noted in the log. This is to ‘display’ the strength of a start that uses an early warrior or scout for these duties.
4. Me trade fire for your rock. Tech research is set to minimum. Research may NOT be turned completely off. Research must advance at a bare minimum each round. Start with Alphabet and then move to writing for all tests.
5. The luxury slider and military police will be used to combat all unhappiness. No clowns were harmed in the making of this test.
6. Peasants will not be repressed. I will not use pop-rushing, at least in the initial tests. Variants may be tested after initial test data is complied and depending on how irritating the peasants have become. I reserve the right to shoot peasants on a case-by-case basis (though not for the purpose of converting their carcasses into shields).

I will firm up these details as they strike me or someone asks for clarification. I will upload the various maps and attach some screen shots. I also have some preliminary data, but I realized I will need to rerun some of the worker first tests. I wasn’t completely true to the test parameters. So I have about 12 of 36 tests already done.

I will be tracking nearly every part of every round with the following exceptions. Notes will be a little sloppy on recording worker moves while giving an idea of what they are doing and where. I will not track military moves – they will scout or move between cities as needed to combat unhappiness and push back the fog.

Some of the main stats I will be looking at are: Average per turn [gold, expenses, income, food left in the box, shield left in the box, shield waste, gold waste, food and shields]; # of citizens, cities, settlers, workers, scouts, warriors, final fpt, spt, gpt, treasury, the founding dates for each city, the average time to found a city, and the average time between founding a city. I may generate more data than this, but it is a starting point. If anyone can think of something else to watch, let me know and I will see what I can do.

I will try to put this information together in some presentable and meaningful way to spark discussion. I will also make the data available. I will also make all of the game starts available (eventually, some are yet to be created). All games are PTW. I can't upload the spreadsheet, though. If anyone is interested, I will try to find a way to make it available.

**Warnings** Me. Yep, I’m a warning. The test is only as good as player that is doing the test. I have PTW. I am comfortable playing monarch; I have beat emperor; I have not yet tried deity and I occasionally play a 16 civ Pangaea on warlord and go to war with all 15 civs at the same time just for fun. It may be obvious that I prefer Monarchy as my system of government (though that doesn’t factor into this test, we’ll be in despotism the entire time). The reason I am doing these tests is because I have NOT mastered the art of an optimal start, but I am versed enough to make a strong base and win at mid-high levels.

Each test takes 3-4 hours depending on how loud my 2 month old yells or how demanding my 11 & 10 year old sons become. (Yes, there is a 10 year difference between them. Sigh.) I will continue to fill in the test data as it becomes available. I can usually knock out one each night and a few on the weekend.

Comments and questions welcome. I’ve received endless amounts of generous and wise advice from the great players that haunt these electronic halls. Hopefully this can be of use to new players to answer (at least for themselves) the age old question – Granary, Settler or Worker first?

**EDIT** House cleaning - I posted the starting position pictures to post #2 and I will keep them there. I did this so I can post all of the data in post #1 to keep them together. For now I will keep the start position files here (Post #1) but I might move them later if needed.

 

[Raliuven]

I have updated the raw data sheet with an explanation of each category (just in case it is not clear). I have also uploaded a Report Card to the first post. I will keep refreshing those posts to keep it all in one area.

The report card is an analysis of my findings. The Report Card will use the gathered information from the raw data to give each test a ‘grade’ in specific areas. The purpose of the Report Card will be to suggest the use of a particular starting approach given the facts ‘on the ground’. The report card will compare the performance against starts within the similar category. Although the industrious civ and non-industrious civ are together, I will not rate them against each other. They are together for the purpose of comparing the civ trait.

I have completed the report card for the NFB test. I should have the NFB1stradFB2ndrad done soon, which should make an interesting comparison. Just how does 2 bonus wheat resources change a game?

The report card measures 6 main areas: Available Gold, Population, Food Production, Shield Production, Military and REX.

As a summary, GF won the NFB test overall. Its most glaring defect is its lack of scouting and military. Otherwise, it basically won all other categories. SF made a good showing. WF suffered the worst overall rating.

1 Available Gold/ 2 Population/ 3 Food Production/ 4 Shield Production/ 5 Military/ 6 REX
GF(NI)NFB 1 A/ 2 A/ 3 A/ 4 A/ 5 F/ 6 A-
SF(NI)NFB 1 C/ 2 B-/ 3 C/ 4 B/ 5 A/ 6 B+
WF(NI)NFB 1 F/ 2 B/ 3 D/ 4 B/ 5 B/ 6 D

GF(I)NFB 1 B/ 2 A/ 3 A/ 4 A/ 5 F/ 6 A-
SF(I)NFB 1 B-/ 2 B-/ 3 C/ 4 C/ 5 A+/ 6 B+
WF(I)NFB 1 B+/ 2 B/ 3 C+/ 4 B (B+)/ 5 C/ 6 C+

The full analysis is in the report card and I am open to other interpretations.

 

[templar_x]

i cannot open the saves now as i dont have the game at hand, nor CA2. i may be able to do that next week.

i had a look on the word files though. quite a bit of work invested there!

if you want feedback on whether someone could find more optimal outcomes for certain test runs, we´d in addition need the spreadsheets for any of those starts. i hope you´ve still got them around.

if not, i could play a few of your games and deliver what i think i could get outta them under the given situations. but i would not even need to fire up civ if i had the spreadsheet and a pic of the alternative starting situations *with* ressources.

reading that you provided 1-2 wheats in the ressources start i am sure that you are aware that most players would choose a cow over a wheat. it gives you an additional shield, which can be very helpful in the set-up of a factory.

templar_x

 

[AutomatedTeller]

having a cow makes getting a worker first even less important. Really, with a irrigated cow on grass start, all you need is one forest and 1 BG.

 

[Raliuven]

Well, as it turns out, I planned to test both. I placed the bonus wheat in a way that was rather complimentary to the development of the first city, following the trail of BG. In the 3rd set of tests, where the 2nd and 3rd city can both capitalize off of food bonuses, I placed a cow in that set. However, chasing the cow does means you will have to work away from your 1st city's BG, at least at the start.

And although it doesn't show in the pictures, I actually placed one of the wheat in the NFB1stradFB2ndrad test on a BG - which IIRC, never happens in a real game. So to be true, the Wheat+BG is really the equivalent of a cow on grass (I think that is accurate, correct me if that is wrong).

Screenshots and another opening save to come in short order. I also haven't done the final test & report card on the 2nd set of tests. RL got in the way, but I am aiming for Sunday.

RE: the spreadsheet. I tried to upload it but got an error. Not sure why. I use excel 2007, but I saved it as 97-2003 and it still wouldn't upload. It is pretty big, but less than 2MB. It will proabably bloon from there - that is only 12 of 36 tests. I would welcome any suggstions on how to get around this?

I will post a few screen shots of the final boards (not all of them, but maybe the winners?) - at least the 'city' path I took should be clear, since each city is labeled in the order I settled them.

 

[Raliuven]

Here are the screen shots for the GF test games. One thing to note is that you will see a fish resource in the test game. I was playing with the editor and left them in by mistake. I never used the fish resources. They are not in the save games I uploaded because I eventually removed them so they wouldn't cause any inadvertant problems later.

 

[templar_x]

having a cow makes getting a worker first even less important. Really, with a irrigated cow on grass start, all you need is one forest and 1 BG.

You'll Need another Food Bonus without the Agri Trait. And there is None in ptw.

Templar_x

 

[darski]

Ok... I just can't see any cows or wheat anywhere... maybe I am just used to different graphics but could you point them out for me?

 

[AutomatedTeller]

You'll Need another Food Bonus without the Agri Trait. And there is None in ptw.

Templar_x

That is true - with a non-agri civ, all you need is an average of 5 shields working, so you don't even need the forest (since it takes 6 turns to replace the 2 citizens you use for the settler. Even less reason to need the worker.

 

[templar_x]

If i Count correctly, a Granary at +4f is just As efficient As at +5f, each Time saving 4 Turns of Food growth per settler. In a 6 Turner, u really hardly ever Lack shields, so u won't Need Many mines and thus no other worker, just As AT Said.
The One Situation Where worker First Might work is with Lots of Food and absolutely no shields around.

Templar_x

 

[Raliuven]

Okay, this is a little late, but I have updated the report card and raw data for the NFB1stradFB2ndrad test. This examines the benefits of 2 wheat bonus food sources.

As a summary, I have to say for total point count, that SF takes the lead, though that may be deceptive. So for those that are only reading the summary, SF is a more balanced approach, but GF still produces a stronger result in 4 of 6 categories, particularly those that focus on expansion/growth. SF leads in military and gold.

1 Available Gold/ 2 Population/ 3 Food Production/ 4 Shield Production/ 5 Military/ 6 REX
GF(NI)NFB 1 C-/ 2 A+/ 3 A/ 4 A/ 5 D/ 6 A-
SF(NI)NFB 1 A/ 2 C/ 3 B+(A-)/ 4 B/ 5 A/ 6 B-
WF(NI)NFB 1 C-/ 2 C/ 3 B-/ 4 C/ 5 F/ 6 D-

GF(I)NFB 1 B+/ 2 A/ 3 A/ 4 A/ 5 F/ 6 A+
SF(I)NFB 1 A/ 2 B/ 3 A-/ 4 B/ 5 A/ 6 (A-/B+)
WF(I)NFB 1 C+/ 2 C/ 3 C-(D+)/ 4 C/ 5 C/ 6 D

The full analysis is in the report card.

One thing I found interesting was how far behind the WF approach fell. Examination of the data shows (IMO) that this start just does not effectively recover from the resources drained by this approach. Originally I had thought that the WF approach would allow settlers to reach their destinations quicker and immediately work on improved tiles right away, thus making up for the delay in producing the first 2-3 cities. This does not appear true.

Regarding gold production, GF had stronger production gpt possibilities, but expenses for granaries and luxuries (due to lack of MPs) eliminated this lead. In the case of non-industrious civs, this reduction is considerable. Such a dramatic cut in available gpt may/will have an impact on research/trade capabilities.

 

[Raliuven]

Ok... I just can't see any cows or wheat anywhere... maybe I am just used to different graphics but could you point them out for me?

I do not use any mods whatsoever. My PTW is as installed, with the exception of the font folder I had to change to get the game to launch.

In the test games, one has no food bonuses. On the other two, you will find the bonuses to the east of the main city (which is pretty much at the center of the island). One set is the southeast (2 wheat), the other to the northeast (1 cow, 1 wheat). You can see them in the pictures in Post #1.

I haven't created the other three test environments, but I will keep those resources consistent as they are added. I will be adding some food bonuses to City #1 eventually, haven't decided where or what yet.

 

[Raliuven]

I have updated the Raw Data and Report card for NFB1stradFB2nd3rd - this tests the effect of adding a cow and a wheat to a 3rd possible city site. This means there are now 4 bonus food sources near the 1st city, but outside of its 21 tile radius.

Here is a summary of the results:

1 Available Gold/ 2 Population/ 3 Food Production/ 4 Shield Production/ 5 Military/ 6 REX
GF(NI)NFB 1 B+/ 2 A+/ 3 A/ 4 A/ 5 D-/ 6 A-
SF(NI)NFB 1 A+/ 2 C+/ 3 A-/ 4 A-(B+)/ 5 A/ 6 A-
WF(NI)NFB 1 A-/ 2 D/ 3 B+/ 4 B/ 5 C+(B-)/ 6 A

GF(I)NFB 1 A-/ 2 A/ 3 A/ 4 A/ 5 F/ 6 A-
SF(I)NFB 1 A/ 2 B/ 3 A-/ 4 B/ 5 A/ 6 A
WF(I)NFB 1 B-(C+)/ 2 D/ 3 B-(C+)/ 4 C-(D+)/ 5 A/ 6 C-

The full analysis is in the report card. Photobucket wasn't working last night, so screen shots will follow, hopefully tonight.

Some observations:

WF- I am actually surprised me by how little difference there is between the NI and I tests. The improved worker performances and even the increased number of workers had little overall effect and made little difference in the REX. The only difference is that WF can complete much better from a military standpoint, but that's not much help. I even tried a few forest chops with the Industrial workers - something I would NOT try with non-industrial workers - to added a boost to production and it did little good. IMO, early forest chops are a bust - they might be applicable in a very shield poor area.

One point on WF is that in the non-industrial test, it actually managed a faster REX than either SF or GF. Once it began to produce settlers/cities, it did so with increasing speed over SF or GF. It was relatively competitive in other areas, but didn't win in any other areas - so take it for what it is worth.

SF - Settler First closed the gap on REX with GF. It even won in the industrial test. It could be viewed that if there is an abundance of bonus food around, settler first can be competitive on the REX stage and it maintains its military advantage. I do worry about what happens after turn 60, though. GF has a solid food production and population base, SF remains second tier for the most part. However, I would say that SF is a good balanced approach - good REX, good military, high gpt income.

GF - Militarily it remains weak. REX is always strong and it is gaining in the treasury department. If there are good huts, this start can/may/will suffer. Can be competitive on the tech research stage in regards to gold expenditures, except that it will not gain early contacts very quick. Strong population counts also lead to higher shield production - only those shields are sunk into infrastructure investments.

That concludes the No Food Bonus in the first city radius tests. I will arrange another attachment that allows comparison between test types so it is easier to see how each approach changed as food was added. That might generate more conversation.

The next test will be removing all food bonuses and then adding a few in the 1st city radius. I will have updated screen shots available soon.

 

[Raliuven]

I have completed the first set of tests were there is food within the 1st city radius, but nowhere else. This is the FB1stradOWNB set.

Here is a summary of the results:

1 Available Gold/ 2 Population/ 3 Food Production/ 4 Shield Production/ 5 Military/ 6 REX
GF(NI)NFB 1 B-/ 2 A/ 3 A-/ 4 A-/ 5 B-/ 6 B-
SF(NI)NFB 1 A/ 2 A/ 3 A/ 4 A-/ 5 A/ 6 A-
WF(NI)NFB 1 A-/ 2 A-(B+)/ 3 A-/ 4 A/ 5 A/ 6 A

GF(I)NFB 1 A/ 2 A/ 3 A/ 4 A/ 5 F/ 6 A+
SF(I)NFB 1 A/ 2 A-(B+)/ 3 B/ 4 B+/ 5 A+/ 6 A
WF(I)NFB 1 B-/ 2 A/ 3 B/ 4 A-/ 5 D/ 6 A-

The full analysis is in the report card.

A couple of observations - In these tests the WF approach actually appears competitive. Particularly in the non-industrious test, the WF approach pulls ahead or is at least equal in some areas. SF falls behind in military might in the (NI) test and WF makes up the difference. General feeling is that the WF fell flat in the NFB starts because it could never make good the initial food 'down payment' needed for the worker. As that 'down payment’ becomes more insignificant and the early worker improvements are needed earlier, the WF begins to pay off. WF is a somewhat plausible approach in the FB1stradOWNB set.

In the (NI) sets, GF and SF seems to be losing their specialization/advantages. The sets have relatively similar ending stats, trading a few warriors here for a granary or an additional worker.

In the (I) set, one of the biggest points of interest was the large treasury that GF was able to generate - previously a strength of the SF start. However, Militarily, SF dominates this set, but not unlike the ratio seen in the (I)NFB tests, just on a larger scale. In terms of producing scouts for early exploration, this has the potential for a lot of exploration power. Not to mention an early conquest. One last note - in the WF test in this set - I put a lot of weight in the fact that the WF had a very sound infrastructure at the end. GF also did an excellent job. SF was struggling.

I will update the report card and raw data sheets in Post #1. Tests are going a little slower because of the quicker expansion - more cities = more documentation. I also have to start GOTM106 at some point if I want to submit this month. But I am dedicated to finishing this test. I hope to have the next serious update this weekend.

 

[Raliuven]

I have completed the FB1st2ndrad tests. This is a world where there are food bonuses with reach of the first city site. There is a cow and wheat, both within reach of irrigation. There is also food bonuses located near a second possible city site – these bonuses are the same type and location as the NFB1stradFB2ndrad test.

Here is a summary of the results:

1 Available Gold/ 2 Population/ 3 Food Production/ 4 Shield Production/ 5 Military/ 6 REX
GF(NI)NFB 1 C-/ 2 A/ 3 A/ 4 B+ 5 D-/ 6 B
SF(NI)NFB 1 A/ 2 A/ 3 A-/ 4 A-(A)/ 5 C+/ 6 A
WF(NI)NFB 1 A/ 2 A/ 3 A/ 4 A/ 5 A/ 6 C

GF(I)NFB 1 A/ 2 A/ 3 A/ 4 A/ 5 F/ 6 A
SF(I)NFB 1 A/ 2 C/ 3 A/ 4 B/ 5 A/ 6 A
WF(I)NFB 1 A/ 2 B+/ 3 A-/ 4 A-/ 5 D/ 6 A-

The full analysis is in the report card. I will update the raw data sheets and report card.

(NI) test summary – for the first time, I call WF the winner by point count. It is a very close call. SF has a strong showing, only .5 behind. The facts on the ground will determine which one is the better approach. GF was the big loser here. Lack of Gold and Military could spell doom for this start. All starts suffered from happiness problems – particularly GF. This may account for its poor showing in treasury department. GF would gain the most from an early luxury or two. It is also significant, in my opinion, that WF is so competitive. For the first time an emphasis on worker improvements overcomes strict expansion. Given the vast difference in worker counts and infrastructure, I believe that WF could easily have staged a later victory. At some point, SF would have to slow down to produce workers and fill in the infrastructure. Until them, SF is like a house of cards. Lot of cities but no way to move troops. Power is dispersed, making individual unit production lag. GF had a moderate core and decent growth. I actually think that GF is the path of moderation for this set of tests – a role usually filled by SF.

(I) Test summary – SF is the winner by the numbers. It is important to note that the military ranking pushes SF over the top. GF and WF have many important strengths that should be taken into account. SF spread fast but has a very weak infrastructure. It may be a good start if there are neighbors nearby and you need to grab land and fight an early war. GF is a good blend – good REX and a good supporting infrastructure. WF is competitive here. It has good REX and excellent infrastructure. This approach may be good to use if there is less immediate pressure by neighbors. I think that WF sets up a strong base for Turn 61 and beyond. SF, in my opinion, is overheated. It is going to have to stall at some point to build infrastructure – at that point, GF and WF would likely overtake it. Of course, SF’s superior military could also be decisive. That is all outside the scope of this test, but important to consider.

 

[Lanzelot]

Interesting. I think that these last two sets of tests are now getting close to "real life" scenarios, in particular to the GOTMs on higher difficulty levels. (Maps with no food bonus in the first or second ring are quite unlikely.) So we see that all three approaches can be good in certain situations. (Hmm, that doesn't make it easier, does it... No clear-cut recipe.) It just depends on a lot of other factors. Is there a luxury around to support the higher population of the GF scenario? Are there close-by neighbors who might box you in (or declare war). Etc.
But at least these numerical results give you a foundation for your decision.

A result that I did not expect, was that the GF approach in the non-industrial case was trailing behind on gold income in both cases, FB1stradOWNB and FB1st2ndrad! My intuition had told me the exact opposite. (In fact, in games on Emperor/Demigod, where the Republic slingshot was on danger, I always went "warrior-warrior-granary-settlers" in order to have a better research output. I figured the higher base population built up during the time the scouting warriors and the granary were completed, would yield more gold during the early game.)

And part of me still thinks that GF should yield a better income than the other two in a real life game. Perhaps the "problems" with your test are indeed, that you have no luxury, and also that you build "not enough" workers. E.g the first screenshot shows 10 towns, but only 6 workers. With more workers (after the granary) the picture would perhaps look differently. (This is also supported by the fact that the industrious case looks quite different in your tests.)
So I think that with 1-2 luxuries and more early workers, the granary first approach might still be better than the other two in terms of gold?!

So far we have only been discussing PTW. In C3C I think that the agricultural trait would make an even bigger difference than the industrious trait in your tests. I expect that with an agricultural civ the GF strategy would be the strongest approach. (Perhaps I have time for some tests...)

 

[Raliuven]

One thing I would like to reiterate is that the test parameters are fairly rigid which is why the test data is ‘suggestive’ but not conclusive.

As you point out, there are some situations where it would have been better in the GF test to push out a few early workers, especially in the less productive outer city ring rather than to pursue a granary. But the test parameters restricted me from doing so. The WF approach gave the best flexibility after the initial worker build, but I would hazard to surmise at this point that WF is shooting yourself in the foot if there is no immediate food bonus nearby to recoup the down payment quickly. When there is a food bonus nearby (the producing city), it seems pretty competitive.

Taking the most recent set of (I)FB1st2ndrad test, the GF approach produced a Granary, worker and settler by turn 24. WF accomplished this in turn 27. However, WF had an extra worker between turn 10 and turn 27. GF’s worker didn’t arrive on the scene until turn 24 (It was Granary/Settler/Worker). You could flip this to Granary/Worker/Settler, but then you would have worker available on turn 20 but City #2 would be delayed until Turn 25. Either way, the WF approach gives you 10-17 extra worker turns which can be decisive - thus the more competative results.

However, in say test (I)NFB1stradFB2nd rad, were there was food in the 2nd radius (comparable above, just removing the food bonus from the 1st radius), GF accomplishes all three by #36, though the build is (Granary, Settler, Warrior, Settler, Worker), 2nd and 3rd city foundings in t25 and t35. WF accomplishes the same goal by turn #41, with 2nd and 3rd foundings in t32 and t42.

Your point about GF start having a stronger gold position – I don’t disagree with you on principle, GF should have strongest gold position. I actually decided I should go back to my test data and pull out the average luxury rate for each test – I kept track, so it shouldn’t be hard to produce the data. So a future posting of the raw data will include that column. I very much agree that an early luxury would have the greatest gpt impact on the GF start (generally speaking). With only one exception, GF had the highest population – and thus the highest unhappiness- and the fewest MPs available to make people smile. The benefit of even 1 luxury could tip the scales here. GF consistently has the highest population and thus a stronger (if not strongest) gpt production/potential. Costs almost always kills GF. From the reverse, it always has the highest expenses, except the one time it lost the population count to WF, and then WF had the highest expenses.

Although I only have 6 tests left, I am thinking I might (MIGHT) run a few more to test more things that have caught my interest – set up a food rich shield poor start and see if the WF can win that. And a start with a single food bonus at the 1st radius only. As you point out, that would be more typical of a GOTM start. I think these are harder to optimize. With a +4fpt bonus and a granary, you need to have a city – or perhaps 2 cities – close enough to the +4fpt bonus to pick up the bonus when the main city does not need it. Without a granary there is no problem, but does that waste the potential?

One other thing I think the NFB tests are good for is looking at the NFB1stradFB2ndrad and the FB1stradOWNB test. This would be a good comparison of what moving a settler on the 1st or even 2nd turn could mean to your Civ just to get that food bonus. I know there has been a lot of discussion on that point. I would need to shave a few rounds of production off the FB1stradOWNB test to represent the wasted opening turns, but just looking at the raw numbers it seems unlikely that the loss of 1-2 turns would seriously degrade the potential of getting that food in the 1st city radius. Of course there are other factors, such as immediately appearing ‘weak’ to other civs and inviting attacks, etc. to factor into a real game. I would hazard that within the initial 60 turns, you would have overcome that weakness. So moving the worker to capture a good spot may be a feasible and even better than settling in place with the intent to pick up the food bonus at the second city.

For comparison, in the GF(I)NFB1stradFB2ndrad test (the one with the best REX score by my calculation), the cities were founding in t0, t25, t35, t47, t48, t50, t55, t57, t60. For GF(I)FB1stradOWNB test, modified for a 2 turn move, it would be t2, t25, t31, t35, t39, t43, t47, t52, t56, t56, t57, t60. That is 9 cities vs. 12 cities and even with a 2 turn penalty, the FB1strad test matches or overtakes every founding date. In my mind, that means that moving the settler even 2 turns to capture a decent food bonus is worth it. Caution: this test makes the FB sites capable of becoming a settler or worker factory (+5fpt). Results might not work the same with less than a +5fpt bonus. I would hazard to say at this point that moving 1 turn to capture a +4fpt bonus would probably be worth it.

 

[Raliuven]

I have added and Appendix A to the Report card regarding the following luxury discussion. The Data I am pulling from can be found in the report card and at the very end of the Raw Data Sheet. I did not re-create it for the Alternate view. However, I have updated both data sheets with the luxury columns for comparison. I equated 1 luxury=10% entertainment. I would be interested if everyone 'generally' agrees with this estimate. The following analysis is dependent on that assumption, so be forwarned.

I'm working on the last set of 6 tests, but GOTM106 is taking much longer than I expected. I will get to them soon.

Analysis of Entertainment and Luxuries
The question is how would adding luxuries into the test effect the overall strength of a starting position. This section is hypothetical, but upon looking at the data, I think that we can extrapolate who the winner would be in a given situation, or if it would make little difference.

Just for the purposes of sanity, I will equate 1 luxury as providing enough to lower the Entertainment bar by 10%. Thus, 1 luxury will reduce entertain by 10%, 2 by 20%, etc.

One caution I would give is not to automatically turn the best treasure/gpt producer into the winner in a science research race. Scouting, goody huts and early contacts can also be decisive. Strength in one category is not conclusive.

WARNING – I AM EXTRAPOLATING FROM THE DATA; THIS DOES NOT ACTUALLY REPRESENT A TEST WITH LUXURIES PRESENT ON THE MAP.

(NI)NFB tests
If anything, GF would pull further ahead. I think this test is the clearest indication that if luxuries/entertainment is taken out of the equation that GF will be the strongest gpt producer.

(I)NFB tests
Much the same as above – WF won this category by a small margin, but all are close. A luxury probably wouldn’t make a huge difference in the outcome, but all ending treasures are generally the same.

(NI)NFB1stradFB2ndrad Tests
Here the interesting thing is that GF and WF ended up with about the same treasury. However the introduction of luxuries would potentially make up for a large amount of missing gpt for GF. Given the entertainment counts, I believe it would take 3-4 luxuries to enable GF to compete with SF. Even with 4 luxuries I don’t know that GF could overtake SF in this set of tests. At the more reasonable level of 1-2 luxuries, I would still give SF a healthy win here, but GF does become more competitive. WF is stagnant. Happy perhaps, but still gpt poor.

(I)NFB1stradFB2ndrad Tests
Again SF is the winner, though only luxury would even the odds for everyone. WF would still trail but is at least competitive. GF and SF would likely be a draw. A 2nd or 3rd luxury would be of some use to SF.

(NI)NFB1stradFB2nd3rd Tests
With the addition of the 1st luxury, GF and WF becomes more competitive. GF has the most ground to gain from a 2nd luxury. At the 3rd Luxury everyone evens out, with SF maintaining a small lead. All ends remains close.

(I)NFB1stradFB2nd3rd Tests
GF could pull even with SF with even the 1st luxury. It would be close. Further luxuries continue to favor GF but on a small scale. WF remains the weakest here with the least to gain all around with the smallest gains. This dooms WF to the weakest gpt production regardless.

(NI)FB1stradOWNB
I think it is clear the GF woud be the biggest overall winner with the 1st luxury. WF becomes a little more competitive with SF but SF keeps the lead. GF is also the biggest winner with the 2nd luxury, but is probably still in 3rd place though is now in the ‘B/B+’ range. WF wins the most from a 2nd luxury, closing the gap with SF. Who has the lead at this point is unknown. It takes a 3rd and especially a 4th luxury for GF to stay in this race.

(I)FB1stradOWNB
GF and SF are neck-in-neck to begin. GF takes the lead with the 1st luxury by a nose and continues to pull away with each additional luxury. SF stops benefiting after the 2nd luxury and its ‘potential’ is fully developed – and still behind GF. The margin is not likely to be huge, but I still give GF the win. WF Remains at the back of the pack the entire time, gaining as much as GF with the 1st luxury and even more from the remaining luxuries, though not enough to close any gaps.

(NI)FB1st2ndrad
The numbers here are incredibly lopsided. GF is dying from entertainment costs and it would take a 3rd and 4th luxury for it to recover from its gpt grave. SF and WF remain competitive from the 1st luxury to the end – SF actually gains slightly more than WF to begin with 1 or 2 luxuries. WF recoups at the 3rd and 4th luxuries. Realistically, I believe that SF will probably keep the lead. Depending on how the rates drop from 40% - 30% on the 1st luxury, WF might stage a comeback. Doubtful in my mind. Even so, SF and WF remain very, very close.

(I)FB1st2ndrad
Not only do we have GF at the head of the pack to begin with, GF also stands to gain the most from the 1st and 2nd luxury. WF becomes even more competitive, but SF keeps pace. I believe GF would take the lead and keep it; WF would pull up equal to SF and both WF and SF would remain on GF’s heels and keep a ‘A’ rating. GF moves to ‘A+’ in my opinion.

(N)FB1st2nd3rdrad
By the numbers, GF has the greatest gpt potential, which is generally true across all tests. In this case, SF has a sizeable advantage in both treasury and average luxury rate. As previously noted, SF’s average is 60% that of GF. Adding the first luxury, GF stands to gain the most, which would be expected. SF gains the least. With the addition of the second luxury, SF has nearly eliminated its luxury needs. Realistically, it takes 3 or 4 luxuries for GF and WF to eliminate their luxury needs. However, by the end of turn 60, two luxuries would generally eliminate continuing luxuries needs – all tests end at 20% luxury or less by turn 60. SF has less than a 100gp lead and it is not likely to get too much further ahead if a luxury or two can be connected.

(I)FB1st2nd3rdrad
The story here is much the same as the (NI) tests. GF is crushed with a high average luxury rate, SF is the lowest and WF hovers in the middle. By the first luxury SF has eliminated most of its problems. By two luxuries WF has solved most of its problems, but it would take 3 luxuries to make significant headway for GF. As above, most problems for all starts are solved with even one luxury. It would be hard for GF or WF to gain on SF’s lead unless there is an unusual number of a luxury immediately available, though even one luxury would mitigate the damage and the gap is not likely to get bigger. The tradeoff here is that SF receives an extra 100gp head start.

 

[Raliuven]

This is the last planned installment for this battery of tests.

I have completed the FB1st2nd3rdrad tests. This is a world where there are food bonuses with reach of the first city site. There is a cow and wheat, both within reach of irrigation. There is also food bonuses located near a second and third possible city site – these bonuses are the same type and location as the NFB1stradFB2nd3rd test.
Here is a summary of the results:

1 Available Gold/ 2 Population/ 3 Food Production/ 4 Shield Production/ 5 Military/ 6 REX
GF(NI)NFB 1 B/ 2 A/ 3 A/ 4 A 5 A-/ 6 A
SF(NI)NFB 1 A/ 2 A/ 3 A/ 4 A/ 5 A/ 6 A-(A)
WF(NI)NFB 1 A-/ 2 A/ 3 A(A-)/ 4 A/ 5 A/ 6 A

GF(I)NFB 1 B+/ 2 A+/ 3 A/ 4 A/ 5 A(A-)/ 6 A
SF(I)NFB 1 A/ 2 B-/ 3 A/ 4 A-(B+)/ 5 A/ 6 A-(A)
WF(I)NFB 1 B+/ 2 B-(B)/ 3 A-/ 4 A-/ 5 A/ 6 A

The full analysis is in the report card. I will update the raw data sheets and report card.

(NI) test summary – by point count, SF and WF are tied for the win, though GF is close. Much of the specialization that has been seen in previous tests is washed away. SF’s major strength that pushes it ahead is its commanding lead in the treasury, though WF’s early scout is a tradeoff – SF or WF could have gone either way – a mixture of treasury or scouting could mean the lead in a GH rich environment at lower difficulties. GF does not have this choice – it needs warriors for MPs. In this set, high luxury settings kill GF, impair WF and are even significant with SF, though SF is at about 60% of GF’s luxury rate.

(I) Test summary – Each test had a very dynamic infrastructure growth. Abundant food made early workers viable and competitive and more workers were produced earlier – a double impact with industrial workers. By point count, GF is the winner. While SF maintains control of the treasury and WF leads in scouting, GF has a high population as well as strong food and shield production. GF has the highest gpt production, though it is crippled by its expenses/luxury rate. SF and WF remain relatively tied by point count. Specialization is again lost and all starts begin to draw closer together – trading a little here for a littler there.

I will be updating all of the data results, the report card and the luxury post to include these tests.