With the improvement to the city-building and management aspect of the game, I feel that the middle-late stages of the game are now in excellent shape. IMO, the big problem is now with the initial/early game for one large reason:
Production is limited by city level and starting materials, making city spamming the only viable way of increasing overall production at a level to compete with the AI.
Let me illustrate the problem with some starting location images and explanations.
Let's take a look at our first position, shall we?
OK, so we have two production, four grain and a whopping 2 essence! In addition, we have a mana shard near our starting position (it's a life shard, but as soon as our city's ZOC overlaps it, it will turn into a death shard that we can use) and a river (for dock upgrades). Not a bad location huh?
One small problem:
It takes forever to build anything useful! As you see, our starting 2 materials gave us about 9 production which is rather paltry, even at the start of the game. We could build a slave militia quickly for some defense (this is Magnar's unique trait: slave units) but our city is not in extreme danger at this point. We could also start building buildings, but that would be a poor investment (the average building time for upgradable buildings is ~ 13 turns, which is just as long as the pioneer).
Shall we chalk this one up to an average start? Well, let's take a look at a different starting scenario:
Again, another 4,2,2 start. This time, I will ignore the essence and instead move towards a spot that has guaranteed materials (highlighted in red). After settling, here are the production values:
For the higher production cost units, the build time has halved; however, this was expected since we started out with double the starting materials (each material provides 4.75 production). Another thing to note: the 3 turn minimum build time for our slave militia has kicked in. We actually over produce its production cost (19 production per turn and a 15 production unit cost) but it still takes us 3 turns to build. If we built that unit first, we would waste 42 production! Please ignore the "slave spear" unit without a head; it is a left-over from a design I made in a previous beta version.
Now, our pioneer takes only 6 turns to build and building improvements take an average of 6 turns to build themselves.
Conclusion: Starting materials influence base production values on a linear scale, but unit and building construction times have hard turn limits placed on them at the high end.
OK, so that was a pretty dry investigation, I admit, but let's look at the consequences, shall we?
Consequence 1: the hard limit on unit build times (can never be lower than 3 turns) automatically favors a higher quantity of cities rather than the quality of individual cities.
Consequence 2: since starting materials influence base production values, and city levels influence gildar production and research points, expanding immediately is the best first economic/military/research decision.
Consequence 3: if you are not able to easily expand early (which is more frequent than not), you fall behind in research, gildar and production.
Consequence 4: since natural resources are now built by cities directly, weak starting locations cannot harvest them effectively.
Consequence 5: the AI is very sensitive to the overall empire score. If it senses any weakness, it is likely to declare war without warning.
Looking at my two starting locations, let's just see how production scales up until turn 30:
Start1 (4,2,2): 9.5 production
Start2 (3,4,0): 19 production
Turn 10: Start1: 95 production (0.88 pioneers)
Turn 10: Start2: 190 production (1.76 pioneers; expands first!)
Turn 15: Start1: 142.5 production (1.3 pioneers)
Turn 15: Start2: 285 production (2.63 pioneers; second expansion possible!)
Turn 30: Start1: 285 production (2.63 pioneers)
Turn 30: Start2: 570 production (5.3 pioneers)*
* Note: pioneers are actually more expensive than several starting buildings. This means that up to five pioneers/buildings could be constructed by the second starting location by turn 30. If next to a forest, this means a logging camp can be constructed for more production. A merchant also makes a great starting building to offset the wages of a starting champion. So starting location 2 easily beats the first location.
Now, let's see how many slave militia we can create in the same time:
Turn 30: Start1: 10 militia
Turn 30: Start2: 10 militia
The hard cap equalizes the disadvantages of the two cities, but this is only applicable to Magnar, since only he starts with cheap, low production slaves! Higher production military units would (obviously) still favor start2's bonuses.
Overall conclusion: Players must pick a starting location with at least 3 food, 3+ materials and 1+ essence in order to remain competitive. Also players must have an easy expansion slot. Anything less cripples early game production.
Now, let's talk about some possible solutions to this. Here are my ideas:
1. Reveal the "hard" fog of war (the black, "unexplored" terrain) from a larger area around the starting position.
Rationale: This would allow a player to make a judgement on easy expansions after a map is recently created.
Problems: This does nothing to address the problem, but rather defers it by making it easier for a player to conclude that there are no easy expansions before clicking Ctrl-N.
2. Add a new starting building to the first city founded by the sovereign. This building would give production and research bonuses inversely proportional to the number of cities that the player/AI owns (ie. more cities, less bonuses). For example: +10 production (-5 for each additional city).
Rationale: This would reduce the influence of variable start locations on early game production and competitiveness. It would also encourage players to stick with the random map that they are dealt (rather than hit Ctrl-N 4-5 times (like I do)). Also, this would make the decision to expand optional, as each expansion does not improve your early production as much (it actually lessens production in your starting city!)
Problems: This would require balancing to make the starting city just powerful enough to allow the player to forgo expansions for a while.
Regardless of which solution you choose, I think that the 3-turn hard cap on unit/building creation should be reduced to 2 turns at the least. Three turns is too restrictive on the early game and the creation of great quantities of cheap units is not that effective in tactical combat (yet) to make the three turn cap justified.
Well, this has been a long-enough post as it is! If you have read through most of it, congratulations on making it this far! If you have your own suggestions or comments, please feel free to post below.
#3. Rebalance the cost of pioneers to make them more expensive relative to the initial buildings. I'd also make them slower (the base unit should move one square per turn) which would make them more vulnerable to monster attacks.
Good analysis. I haven't seen the point mentioned that the pioneers cost as much as the early buildings, which I think makes training pioneers initially a far better strategy than building anything.
One thing that I did not see in the calculation was Enchanted Hammers. Now of course, you did not have it for Magnar, so that is fair in this case analysis. Still, a 4/2/2 is great if you cast Hammers to double production. Your overall observations are spot on.
I would also point out that at least in my games I cannot find a point to settle another city until many many monsters are cleared out. So I build logging camp or merchant first. Playing as Magnar specifically, I would take a 4/2/2 and go Inspiration. Then I would go after as much research as possible and look for a hero with Earth I. Second enchantment would start out as Meditation. So now I have some good mana coming in and a jump on research. It would not be the fastest start, but as soon as I either found an earth wizard or finished researching Leatherworking and Training, I would then have a strong advantage. Even one slave spear and leather unit is powerful in the early game with Magnar. The next step would be leveling Magnar to get growth and then choosing Oracle at level 3. Now I have 3 Essence. I then start producing mass amounts of slaves and a few pioneers. All new cities will be fortresses of 4 materials or better. That is how I would handle it. Doesn't always work out, but then trying is the fun of playing.
I would fix the pioneer spam problem by making pioneers cost twice as much to start. Then in the Civilization tree I would make Administration half the cost of pioneers. So now you need a tier 2 tech to expand at the rate one can currently expand. The opportunity cost of pioneers is doubled on turn one to reflect the scarcity of people early on. The AI would then need to change its value of expansion until Administration is researched.
Very true. Without Earth magic (enchanted hammers), you are limited to the base values. Thank you for the confirmation.
This is a good starting strategy; however, I see a couple of issues: (1) it is very dependent on random chance (champion with earth, fortresses with 4 mats, expansions in general) and (2) you are quite vulnerable without some base defender units (though no-moreso than in other builds).
I think that the major issue is that an expansion at the earliest time point is absolutely critical rather than being a luxury.
I think that this is a good option, but the AI would have to take this into account as well. As it is, the early game appears to be a race against the AI to secure city spots and resources. If the AI focused more on the minor city improvements after this change, then it could be viable.
Ohhh! I like this very much. Very elegant solution, as it forces players to focus on developing their initial city to get to a somewhat expensive tech before really expanding.
I do also like the idea of inversely scalable buildings, though. There is still no good strategy for those of us who would rather build 2-3 cities and work on quests/diplomacy/magic victories. You will invariably fall behind the computer who is generating more money, more research, and having more opportunities to build things.
Toss on top of this that armor is still largely king in combat, and you have to waste research up the warfare tree to even reliably defend your cities, which slows down all other non-conquest victories.
The new focus seems to be on a desperate expansion early on. I don't like that as much. There are a few factors making the "more cities is better" approach not work. It relies on good map balance. This sort of thinking assumes that each player has the opportunity to expand that is relatively equal to everyone else. That has never been the case. Two of even the weakest cities are still better than one bountiful city in this new beta. That works fine if we all have an equal chance to expand, but in many cases I get a mediocre start and have to go 12+ tiles to find a place to settle that is not covered in Obsidian Golems. By the time I have that city, all the areas inbetween are taken by the AI, who can settle next to an Obsidian Golem. The next season they invite the Golem over for tea, while I am attacked by a roving umberdroth that passed up countless AI outposts to find me. That is another problem with the current design, the AI is not attacked by the world as much. They are not harassed by roamers or attacked when settling next to a lair in general. Humans are in general the target of the Monster AI. That means we have to dance and fight our way to an expansion, while the AI can move once per turn to all the best lands. This is all compounded by the fact that sometimes there just aren't any expansions near my start. And then you have to consider how the human games this design. I generally don't build outposts because I know I can go to war with the AI and take them. Those outposts come with constructed resources. I save myself 50-100 turns just by taking undefended outposts. And then I look for a good AI city, create a modest army and attack. The AI is so busy expanding, I can usually find a nice expansion city with 100+ turns put into it. Wait out the Unrest penalty and now I have a level 3 city without all the fuss of earning it. Train two more armies and my faction power goes above my new enemy's. So I can buy peace at a fair price and look to the spoils of war. The AI can neither defend against this, nor accomplish it for itself.
There are ways to make the new design work. You could fix all these problems. That is likely the best way to go at this point. Still, I would do this:
-Increase Pioneer cost by 100%
-Make Administration half Pioneer cost.
-Give Towns +1 Gildar per level; Fortresses +1 level for Trained Units per level; Conclaves +1 Research per level, keeping the current bonuses of research for city levels.
-Make each starting point offer at least a 5/3/1 or 9 points worth of tile resources; require each starting location to have 3 random resources within a radius of 2 from the city.
-Give each Sovereign History Trait a unique city enchantment or ability.
-Make stamps for expansion bases that are always within reach of each player; set Wildlands to be as far away from starting locations as possible.
-Make sure the Monster AI is playing fair.
-Set each expansion stamp to have lots of baddies to fight.
-Teach the AI to always clear all baddies before settling in an area; don't let it settle near super strong monsters.
-Teach the AI to only put up an outpost when it can dedicate units to defend the area and to build the outpost defenses.
-Give outposts a city militia unit that is upgraded to be a proper guardian.
-Make cities very hard to conquer until later in the game. There needs to be a sense that we need catapults to take cities.
-Fine tune the Sov start to allow for a longer and more interesting RPG feeling. Sovs need armor, weapons, magic, and friends from turn one.
I will of course want to mod in a one-city option anyway, but the devs should look at doing all of these to make their design more functional.
An excellent analysis of [at least one of] the incentives behind expansion. Just like seanw3, I, too, am troubled by the AI's apparent focus on Pioneer Cloning Vats in the early game to the exclusion of much of anything else.
I'm personally partial to an alteration in the taxation curve for new cities such that they begin by draining, say, 5 Gildar, don't break even until they reach Level 2 (so any taxation on a city of 50 population will return 0 Gildar), and don't hit the current taxation return curve until they reach level 3. GalCiv 2 made colony spam viable because it was possible to make sure there were a lot of planets per player, even at maximum settings. The minimum distance in FE combined with the general infrequency of viable building sites on "Balanced" maps means that it doesn't work so well here.
We're still wrestling with an AI that calculates that it's in its best interests to seed cities like so many grains of wheat across the map-- and in some sense, the AI is probably right. The fact that it chooses to rush the cities puts pressure on the player to compete. But we're better tacticians and strategists than the AI is, so for us, "competition" often means "taking tons of poorly defended enemy cities" rather than "spamming our own pioneers." I'd prefer if neither were the case; there needs to be disincentive, for players and AIs both, from expanding too fast, and that disincentive needs to be a sizable economic hit of one sort or another.
Why not just have Pioneers reduce the population of the city that built it by X (50ish?) That way you cannot just immediately build pioneers since there are not enough people to do so. Once you do have enough people, the question is do you want to set your first city back in population and delay leveling the city in order to expand, or do you want to keep the population and therefore the production centralized. The material cost of pioneers would then be secondary to the population cost.
Spamming pioneers would quickly deplete the population of even a modestly sized city. Losing pioneers would be losing precious population that took time for your prestige to build up.
I think you should actually watch the AI play before making too many complaints. I saw that it took 200 turns for the AI to go to war and 300 turns for the AI to kill off the first player. I think that the problem is the faction power is not accurate enough since it tends to underestimate human performance. However, I never have trouble staying above the AI.
The AI does spam pioneers, but it doesn't really spam cities. The first pioneer always makes an outpost nearby their starting city. The first thing they build is a pioneer, and they usually get a 4/2/1 starting position. The next thing they do is build any resources connected to their city and then send out a second pioneer to establish a city. If they dont find a good spot they spend many turns escorting the pioneer around looking for one.
The AI does like to create lots of cities, usually getting 3 by turn 60. At that point they usually slow down and focus on other things (or do a terrible job scouting out new areas to settle). One AI will shoot forward at this point by expanding like crazy and putting down cities wherever possible. This AI will usually be the strongest and usually win against the other AI.
So the AI does City spam, but not really at the beginning of the game. You really just need to watch out for which faction seems to be doing the best and wipe them out early, or team up against them with other AI's.
City spamming does have a flaw in that your population goes up really slow. If you keep fewer cities they can get to level 3 and 4 pretty quickly, if you put a bunch down they will never get past level 3. Maybe we just need stronger city level up bonuses.
Also the minimum build time is less than 3 because you can buy things. If you have lots of production you can buy them in 2 turns and get gold back, allowing you to alternate between buying on turn 1 and buying on turn 2 without losing money. I do thing that this bug is silly and should be fixed, but for now it solves the issue for me.
I think the biggest problem with the AI's expanding strategy is that it highlights how little aggro the AI gets from building right next to monsters and lairs.
So what I suggest is:
1. For starters, for any city or outpost location that the AI puts down, make sure that it has made sure there is an area clear of monsters and lairs (except weak monsters) equal to the starting ZoC plus two more circles around the city or outpost. This will lessen the human's agastness (yea, I made that word up ) at the AI being able to settle normally right next to umberdroths and obsidian golems.
2. Consider having monsters that are medium or tougher threat (or any monsters protecting lairs) starting off stationary, but always targeting cities or outposts that disturbed them. In other words, no matter who the owner is, if a monster or lair is disturbed, there should always be monsters heading for (and attacking!) the city or outpost closest THAT DISTURBED THEM (probably defined by being next to them by any units or when a town's/outpost's ZoC goes over a monster or lair). Honestly, that should be feasible with Frogboy's AI coding capabilities. Even if it means that the monsters don't move until they are disturbed so the monster AI can determine which player disturbed them. Hey, you could even make the monsters hold a grudge - each monster or lair once disturbed only goes for the player that disturbed them until any such time the player is wiped out, when they would go back to stationary status. Plus, if the medium or higher threat monsters don't move until they are disturbed, it makes the game MUCH easier to balance than currently, when players have already agreed that there is just too much variability in viability of even good starts, based on how aggressive (or not) the monsters turn out to be.
3. Make Master Scouts protect against units only as Stardock said they were strongly considering, so anything owned by a player that isn't a unit DOESN'T get protected from attack (so cities can be attacked deliberately, and outposts and resources can be attacked randomly). Cities or outposts are not very stealthy, you know, from the fact that they can't move.
Plus there are a lot of good suggestions in this thread. The 3 turn minimum is a tough one. If Stardock want to keep this, they need to make sure that for starters, the negative gildar cost is fixed, and any rush buy cost is based on 1. production left to do, and 2. if a unit or building WOULD get completed in less than 3 turns then a flat fee of 50 gildar per turn remaining is added to the cost of rush buying any production that would normally complete in 1 or 2 turns. I think that is a simple fix that would work.
I like where this thread is going.
Based on the OP analysis how about this?
The new outpost are powerful enough you could limit pioneers to that function alone. City building should be done by the Sovereign, which is in line with the lore. The sovereign could build a set amount of cities at the start then an additional city after a set number of turns (e.g. Two at start then a new one every 100 turns). This will help pacing, control city spam, and make city defense crucial.
Certainly. My comparisons are highly dependent on the AI's abilities and that tends to vary greatly from game to game. I just finished a game where I had an optimal start and simply rolled over all AI players except for one. I also had two games where I tried an inoptimal start (low mat start and (in one game) I was unable to expand for 100 turns. I was on a "barren" peninsula, sadly) and was rolled by two AI's in the middle of the game. They just had too many stacks of units.
Sure, faction rating is not accurate, but the problem is that the AI treats it as if it is the end-all-be-all measure of your potency. Having a faction rating 20 points above one AI is enough to make them offer you tribute. The inverse is enough to make them declare war on you. Whether that is a problem depends on your abilities, I admit.
The AI does like to create lots of cities, usually getting 3 by turn 60. At that point they usually slow down and focus on other things (or do a terrible job scouting out new areas to settle). One AI will shoot forward at this point by expanding like crazy and putting down cities wherever possible. This AI will usually be the strongest and usually win against the other AI. So the AI does City spam, but not really at the beginning of the game. You really just need to watch out for which faction seems to be doing the best and wipe them out early, or team up against them with other AI's. City spamming does have a flaw in that your population goes up really slow. If you keep fewer cities they can get to level 3 and 4 pretty quickly, if you put a bunch down they will never get past level 3. Maybe we just need stronger city level up bonuses. Also the minimum build time is less than 3 because you can buy things. If you have lots of production you can buy them in 2 turns and get gold back, allowing you to alternate between buying on turn 1 and buying on turn 2 without losing money. I do thing that this bug is silly and should be fixed, but for now it solves the issue for me.
I have played some more and I do see this in action. The AI's placement of outposts and cities does leave things to be desired. I have also noticed that there is no minimum distance between outposts for the AI (or it breaks the rules).
As for the minimum build time, you are correct that you can rush production on everything, but it depends on if you hit the hard cap or not whether or not you will be charged gildar. I agree that the negative cost bug should be quashed immediately.
There are many great features available to you once you register, including:
Sign in or Create Account
