Hacker News Reader

Comment by devit | Hacker News Reader

While Civilization games are fun for new players, ultimately they are quite flawed:

- In Civ 1-3 infinite city sprawl is optimal, which is tedious and unfun

- In Civ 5-6 one-unit-per-tile creates gimmicky non-strategical combat, forces you to move units individually and is very hard for the AI to use

- All the original versions have horrible AI (so bad that you could reasonably request a refund), although the Better AI mods for Civ4 and Civ5 have somewhat passable AI

- While the early game is really fun since you have few very impactful decisions, all versions get very tedious by the mid/end-game where you need to make a ton of minor decisions for every city and unit in the empire

The last problem is especially hard to fix, and I don't think any similar game has solved it.

View on HN

I always liked Alpha Centauri the most out of all these games.

I don't think it has any of these problems except for maybe the last (you can automate most repetitive actions, but the control is somewhat too coarse).

Alpha Centauri is one of my all time favourite game. Out of curiosity, does anyone know if it is possible to replicate the Human Hive strategy from Alpha Centauri in other Civ games?

> In Civ 1-3 infinite city sprawl is optimal, which is tedious and unfun

Guess I'm in the minority then, having built a 100-city empire in Civ2 and enjoyed every step of it. (Yes, I sometimes wonder how I managed to get married.)

As someone who has played a few hundred hours of the base civ6 game, I agree with the combat needing to be improved. The biggest flaw imo, is the AI makes terrible decisions. To overcome this, the harder difficulties give the AI a huge resource material advantage. This makes attacking early suicidal. Instead, the best strategy I have found is to turtle up until you unlock the bomber.

This bomber has a gigantic range. Forgot exactly what it is, but something like 15 tiles. Comboed with spies to give you vision of the enemy, and a few bombers utterly destroy everything. The solution to every problem becomes spies + bomber. The enemy close to getting a science victory? Put a spy in their city and bomb the spaceport. Going for a domination victory? Put a spy in their city and bomb all the units.

Civ 6 seems like a game where it would have been good for them to add hooks for people to add their own AIs easily.

I suspect the built in AI does okay if you place units in a limited area (which is likely how they tested it), but it struggles getting armies around the full game map with complex borders.

In my experience, the Civ 6 AI doesn't do combat well in limited areas either. And it hardly ever uses planes and bombers. I really hope Civ 7 will be focused on AI innovations.

There are mods already that improve the AI. So it is feasible.

Yes, but there are severe limits as the DLL is a black box. They can only tweak the inputs.

I've always thought the easy fix to the last problem would be to allow cities to be set to "auto" mode, and let the AI decide what to make. You can already get this effect by just always choosing what the advisor suggests.

The AI is certainly not very smart, but sometimes I have cities I really don't care about the minutae of: they are already fulfilling their strategic role and can suffer being mismanaged by the AI for a dozen turns.

The Advisor is kind of lousy at deciding what to build though.

This is also why I don't let workers automate, because they'll chop down my jungles before I have a chance to research Universities.

I've always wanted a really simple scripting or rules engine tools for my workers to automate them.

Just a simple if-then block like:

if (floodplains) { build farm }

if (forest) { build lumbermill }

if (jungle) { do nothing }

It wouldn't be perfect, but is obviously some sort of crappy version built-in to them that often does not do what I want, to the point that automating workers is counter-productive.

Or I'd love to be able to plot out a road/rail tile by tile and queue it up. Or heck, just a general worker queue would be an improvement even if manual.

if (jungle) { build tradingpost }

For building roads you can use the route-to mode which works fairly well. The only caveat is that you can't tell the worker to build roads (because they're cheaper) if you have researched railroads, even if the city has a harbor so it already has the production bonus. Or it's a city state and you're only building it for the diplomacy boost.

Civ3 and Alpha Centauri both have separate Build Road To... and Build Railroad (or Magtube) To... automation modes.

>In Civ 5-6 one-unit-per-tile creates gimmicky non-strategical combat, forces you to move units individually

For me, I think this is a good thing. It makes combat much more tactical, and there is less of the "death ball" effect. The AI criticism is fair, but I do think that 1-unit-per-tile is the way to go ultimately.

"Death Ball" was just a problem in Civ4 and ONLY Civ4.

Civ2 / Civ3 had no death balls, because your entire stack lives-and-dies together. A stack of 20-musketeers in Civ2 could die in a single combat.

In Civ 6 you can create corps/armies with 2/3 units. Still one unit per tile but a bit better than the Civ 5 system.

> - In Civ 1-3 infinite city sprawl is optimal, which is tedious and unfun

FreeCiv (based on Civ2) meta is based on Rapture growth to size 8, 12, and 20+... allowing you to reach gunpowder in less than 50 turns (EDIT: 75 turns reliably, depends on the map of course).

You can only rapture with max happiness, which puts a hard cap on the number of cities in a Rapture Strategy to 13 cities (Republic / Early game) and 14 cities (Democracy, researched later).

Republic/Democracies get the following bonus on "We love the king days": +1 city growth per turn, as long as the city has +1 food AND retains "We Love the King" status. It took many years to develop a strategy that can sustain Rapture reliably, but Rapture beats infinity city sprawl in the long-game.

Infinity City Sprawl is still a strong rush-down strategy if you go Monarchy, but you will absolutely reach gunpowder 2nd. You have to hope you can win with a rushdown with Knights (and probably Trireme / Lighthouse, since Monarchy has that +Production bonus after all). The multiplicative bonuses from libraries / markets make Rapture far stronger than ICS in the long run.

Rapture-based FreeCiv games are looking at Marines / Transports + Destroyers soon after turn 100 or so (0AD), if you let the Rapture player do their thing. Obviously, a bit slower if you're harassing them, but its a big threat. You absolutely won't win the long game with the older infinite-sprawl strategy.

> The last problem is especially hard to fix, and I don't think any similar game has solved it.

FreeCiv's governor is a constraint-solver, automatically solving the ideal amount of food / shields / trade in all cities to your specification.

https://freeciv.fandom.com/wiki/City_manager

Its a bit difficult to learn how to use, but all that micromanagement disappears when you use the governor. Since its a constraint-solver, you are provably given the optimal mix of food / shields / trade that you request from the system.

You need to spend an afternoon with the city manager, figuring out when +1 food / high gold or +1 food / high production builds are useful. Or Max Food / equal production, etc. etc. But once you've got your plans in place, its just a few clicks to automate the management of your cities in FreeCiv.

----------

If a game is good, but tedious, you can solve that with AIs / constraint solvers / etc. etc. that assist the player.

Ex: I think the "ideal Starcraft" engine wouldn't remove Mutalisk stacking or Dragoon Dancing... but instead provide a better UI so that the human can more efficiently utilize those techniques and play the game at a higher level. Similarly, Civ games may get tedious, but you can program miniature AIs to remove a gross amount of that tedium.

Reply on news.ycombinator.com