Ocean tiles / harbour / republic, cheese.

Can't beat the AI? Is it too tame? Discuss the best strategies!
Post Reply
case81
Posts: 8
Joined: Sun Aug 02, 2015 2:33 pm

Ocean tiles / harbour / republic, cheese.

Post by case81 »

I came up against an incredibly tough AI today, one who made incredibly fast research progress. I became suspicious when they took the Great Library from me by building it in 5 turns or so despite my earlier start. And then I just kept getting messages about how they built advanced wonders, one very shortly after another.

I wasn't doing so bad myself, but I was severely outclassed very quickly. That AI was just breezing through the game.

So I became curious and saved the game, then restarted and took over that AI to see how it got that massive advantage. It turns out it got over 700 research points per round (at game turn 150) by following a simple pattern:

- build a library
- build a harbour
- be a republic or a democracy (edit: democracy is better because corruption becomes an issue)
- harvest mainly ocean tiles.

This turns any ocean tile into a 2/0/3 field. Any worker on such a field feeds itself and brings home 3 trade points. The library is cheap and will double it all. This can net you about 30 research per turn pretty early. Due to the amount of food, if you just have a little extra from a fish tile for instance, the cities will keep growing, especially with a granary. This allows you to harvest more ocean tiles and get even more trade / research. All before even having a university. You don't even really need ANY land tiles for this to work except the city center. This also means you never need to bother with worker units - they're quite useless with this strategy since they don't work on water (he had 3 building roads for more trade.) On top of that, your resource tiles can't easily be blockaded by an enemy - you simply reassign the worker to another free ocean tile. Enemy land units have no chance whatsoever to meaningfully besiege you (and city walls were one of the few improvements he built.)

All the inland cities (the ones without enough ocean tiles) were set to permanently produce caravans. Most cities had 4 trade routes, giving minimum +4 trade for each, often +6 or +8. I guess the caravans served to accelerate wonders, as well.

All cities had horrible production outcomes except the capital, which had 15 production and mostly just built wonders (Hanging Gardens, Crusade, J.S. Bach, Darwin...)

Adding a university on top often nets upwards of 70 research per turn per (harbour) city.

Also, all cities were empty. He simply built no land units. I guess if the pirates showed up, he'd just buy 1 phalanx (minimum 9 defense with a city wall) and forget about it. He did actually build barracks, which makes it even harder. Soon after, he could upgrade that to a veteran musketeer (about 20 defense) and even a human opponent would have a very hard time overcoming that without battleships or howitzers. Unfortunately, he would have them first...

Only thing I could imagine that might help would be to mass ships and blockade his ocean tiles. But that costs a lot of shields if you're a republic, and probably causes unhappiness. He was also massing destroyers when I overtook him, probably to prevent just that. So... be a monarchy, mass frigates, and... you still wouldn't be able to blockade him effectively and he would eventually just kill you with cruisers or something.

-> Very, very nasty strategy.

Why do I think it is cheese?

- land tiles / geography become largely irrelevant, obsoleting a major element of the game; placement of the caravan cities doesn't matter
- workers and terrain improvements become largely irrelevant
- enemy loses the ability to effectively siege with land units, completely bypassing that part of combat
- using ships for a siege is ineffective because ocean tiles are a dime a dozen and all bring the same bonus
- rewards smallpox because ocean cities can overlap caravan/wonder cities with little consequence
- even a 1x1 island city can still produce excellent research
- even the need for production points to build units is partly bypassed by simply making a lot of money

In short, this strategy exploits the most abundant resource in the game (simple ocean tiles) and bypasses a large part of the game (anything to do with land tiles, siege warfare, city production points and even city placement) while at the same time being very easy to defend.

Hats off to the AI programmers, but... this seems exceedingly hard to compete with except by out-cheesing the AI. This was on Easy. When I saw what he had been doing, I groaned. :lol:

Edit: A possible counter would be to introduce corruption even under democracy (it's more realistic anyway, we all know democracies have their share of corruption.) Zero corruption everywhere really makes this strategy go through the roof.
case81
Posts: 8
Joined: Sun Aug 02, 2015 2:33 pm

Re: Ocean tiles / harbour / republic, cheese.

Post by case81 »

I was able to duplicate this strategy, getting upwards of 1000 research a turn and overtaking the AIs point-wise. This led to a situation where two civs have excellent research and a lot of money, and thus a stalemate.

I tried to throw battleships, rockets and bomber planes at his cities (even elite units), but he was restocking them with alpine troops quicker than I could kill them. Additionally my attackers would need to dock after every attack to replenish health, giving him time to restock his defense. Currently I'm building nukes because I don't see a better way. I assume he is building nukes too. The game is really tedious and stalemate-y.

Still wondering how to beat this "ocean trade" strategy without simply copying it (which seems to lead to a stalemate.)

Is it possible to run an early/mid game military campaign against such a monster research AI? Potentially alongside diplomats to steal his research. The trouble is, it seems impossible to overcome a city wall in the early game with just catapults or legionaries. So, mass troops? That would mean being a monarchy or republic at best, which automatically means losing at research because of corruption.

One thing I noticed is that an alliance is helpful, because the allied AI will actually build intercontinental trade routes for you, as well as share tech, which really helps. Could backfire later, though.

It looks like copying the ocean cheese strategy, having an ally and building the spaceship first might be the easiest way to beat this. It would be nice to beat it through warfare, but I can't see how. The game is massively biased towards defense and a rich AI can just keep buying units every turn despite not having good production.

If there is a rock/paper/scissors thing in this game, I haven't found it yet. I think there isn't. It looks like massive research is the way to win, which is unsatisfactory - I'd like the ability to be successful by warfare and superior production as well. But there is no equivalent to the ocean tile in the game that yields 2 food and 3 production. You could argue that you can build mines to get 3 production from a tile, but unlike a harbour and a republic/democracy, mines affect only 1 tile and hills only yield 1 food (irrigating a hill destroys a mine.) Thus, the ocean tile with 2 food and 3 trade has no equivalent. And because of this, the game seems to be biased towards a trade strategy instead of a production one, towards defense instead of attack, and towards using ocean tiles instead of land tiles.

The harbour building is really overpowered in my opinion, especially because the ocean tiles already are the best trade tiles and there is no equivalent that gives you additional food from the best production tiles in a city. Likewise, there is no government type that gives you a production bonus instead of a trade one (monarchy/communism gives you 3 upkeep for free per city, but you'll need most of those to pay for the martial law in your cities, thus it's not comparable to the trade bonus of a republic/democracy.)

If there was a research target (to counter the harbour) that gives you a 100% food bonus from all land tiles (hills...) that produce less than 2 food ("Industrial Agriculture"?), a wonder that raises the effectivity of supermarkets ("Agricultural Cooperative"?) and perhaps Communism gave you a +1 production bonus for all tiles that produce shields, then a production-heavy Communist warfare strategy might be able to counter a super-trade democracy. Because hills and mines could become as effective for production as the ocean tile is for trade while matching the harbour's food bonus. And weak production is the major shortcoming of the ocean tile strategy.

I might try my hand at writing a ruleset that does these things. Perhaps also allow veteran diplomats or spies to create +2 corruption in any city, even a democratic one.
louis94
Hardened
Posts: 270
Joined: Thu Apr 25, 2013 10:17 pm
Location: Belgium

Re: Ocean tiles / harbour / republic, cheese.

Post by louis94 »

case81 wrote:I tried to throw battleships, rockets and bomber planes at his cities (even elite units), but he was restocking them with alpine troops quicker than I could kill them.
From my experience, it is almost always possible kill the AIs by :
  1. Choosing a city in the periphery (not the capital), preferably not connected by railroad and not too far from you
  2. Massing a lot of attack troops nearby, harassing AI by air and sea
  3. When you can kill a lot of units, attack the city
  4. Leave minimal defence (the AI probably won't try to take the city back) and repeat
The exception would be an AI with lots of howitzers massed on an island with railways. You may still take a city, but defending it afterwards is very difficult. As you take cities however, the AI becomes less and less powerful, while you can more and more afford loosing units.

I hardly ever fight AIs if I don't have a strong technological lead or Howitzers (I prefer food and science to production). If I don't have that, I wait. The AI isn't good enough to be a very important threat if you have city walls and aren't far behind in research.

Louis
case81
Posts: 8
Joined: Sun Aug 02, 2015 2:33 pm

Re: Ocean tiles / harbour / republic, cheese.

Post by case81 »

Thanks for the reply louis94.

Yes, I was able to take AI cities by simply having a stack of e.g. catapults in each attack. It seems the key is to be patient and only commence your attack if your entire stack is assembled next to the city, healed, and with enough move. Of course the stack needs to be protected. In the early game, this works out OK. Enemy catapults can be taken out pretty easily by exploiting their 1 move limit - they can only attack 1 unit per turn, so if you do a kind of pincer attack with, say, two archers, the catapult will kill one of them but fall to the second.

I haven't tried to take a mid/late game city defended by a stack of alpine troops in this manner yet. Perhaps 6-8 howitzers might be able to do it, depending on the terrain. But howitzers come late in the tech tree, and until then, battleships will be the way to go. So I'd probably need to sacrifice a stack of battleships to take such a city, and have land units camped on a hill nearby to take it.

Concentrated attack does work. It was a mistake to have units take turns to attack. It needs to happen in 1 turn.

Anyway...

I actually went and implemented the "diplomats create corruption" mechanic today (it was fun to program.) Due to the fact that diplomats come early and corruption is really annoying, this works very well as a way to pester a trade AI in the early/midgame.

Image

Image

It can be prevented by defending the city with a diplomat. The cost is currently 20% of the revolt cost and calculated in the same way, meaning if the city is celebrating etc. it is very expensive. When I tried to corrupt Brasilia a second time, they wanted almost a thousand gold because they were rapturing. :D

The diplomat is lost after this and it currently adds 2 points of trade corruption even in a democratic city. Might force the enemy to build courts and rapture more often as well as defend their cities with diplomats pretty early.

The AI actually panicked and started building diplomats after this attack. :mrgreen:

Unfortunately this required C coding so just distributing a ruleset is not enough to get it.
louis94
Hardened
Posts: 270
Joined: Thu Apr 25, 2013 10:17 pm
Location: Belgium

Re: Ocean tiles / harbour / republic, cheese.

Post by louis94 »

case81 wrote:I haven't tried to take a mid/late game city defended by a stack of alpine troops in this manner yet. Perhaps 6-8 howitzers might be able to do it, depending on the terrain. But howitzers come late in the tech tree, and until then, battleships will be the way to go. So I'd probably need to sacrifice a stack of battleships to take such a city, and have land units camped on a hill nearby to take it.
Bombers are also very effective. They cost the price of a battleship, but are much more difficult to defend against (the AI rarely builds fighters). Mid-game, both dragoons and cannons are effective against musketeers. Fighting riflemen and alpine troops is more difficult, I'd say you need at least artillery to do that (but tanks are more powerful and can be researched quickly, so use cavalry+battleships).
case81 wrote: I actually went and implemented the "diplomats create corruption" mechanic today (it was fun to program.) Due to the fact that diplomats come early and corruption is really annoying, this works very well as a way to pester a trade AI in the early/midgame.

It can be prevented by defending the city with a diplomat. The cost is currently 20% of the revolt cost and calculated in the same way, meaning if the city is celebrating etc. it is very expensive. When I tried to corrupt Brasilia a second time, they wanted almost a thousand gold because they were rapturing. :D
How many turns does it last ? Adding corruption for only one turn is not worth loosing a diplomat + paying 150 gold. Maybe the price could be a function of the trade income lost (squared?), multiplied by some constant (4?) when celebrating. Also, does it survive revolution/owner change ?
case81 wrote: Unfortunately this required C coding so just distributing a ruleset is not enough to get it.
Well, send a patch and I'm pretty sure 3.0 will have it (I think it's too late for 2.6). It would be a great addition to the experimental ruleset !

Louis
case81
Posts: 8
Joined: Sun Aug 02, 2015 2:33 pm

Re: Ocean tiles / harbour / republic, cheese.

Post by case81 »

louis94 wrote: How many turns does it last ? Adding corruption for only one turn is not worth loosing a diplomat + paying 150 gold. Maybe the price could be a function of the trade income lost (squared?), multiplied by some constant (4?) when celebrating. Also, does it survive revolution/owner change ?
My plan is to make it last until the other player uses a diplomat to counteract the effects (only the effects of diplomat-induced waste), which would take something like 10 turns and consume the diplomat. Not implemented yet. Currently it just does pcity->waste[O_TRADE] +=2. I guess this could be refined and check how much waste there already is, and so on. I'm actually not sure if it survives a change to democracy.

Making it simply last 20 turns would be easier of course.

I also plan to have diplomat-induced corruption negate the benefits of trade routes, successivly, until either the effect runs out or enemy diplomat "cleans up the mess." This would really throw a spanner in the works of a mass-caravan-strategy. However the diplomatic warfare would have to be continuous to really harm an opponent, and like I said other diplomats could remove the effect.
case81 wrote: Unfortunately this required C coding so just distributing a ruleset is not enough to get it.
Well, send a patch and I'm pretty sure 3.0 will have it (I think it's too late for 2.6). It would be a great addition to the experimental ruleset !

Louis
I might just do that after some more testing/tweaking.


Edit: Of course a communist civ would have an advantage in this kind of diplomatic campaign (because of the ability to produce veteran diplos), which is pretty much what I'm loosely intending. The entire idea behind this was to enable a monarchy/communist/production-strategy player to stand up to a mega-trade democracy by a) hampering the trade (the corruption ability does that) and b) giving the Communist a production equivalent to the ocean tile thing (the ruleset I'm working on does that.)
Post Reply