Frustrating AGRICOLA

This is an article I wrote a few years ago, based on my experience with AGRICOLA playtesting my scenarios. I hope it may be of some use to others

FRUSTRATING AGRICOLA: DESIGNING SCENARIOS FOR THE EXPERT PLAYER

Techumseh, Sept 2009

INTRODUCTION

One of the most challenging things about scenario designing is trying to thwart the aggressive and ruthless style of some players. If the scenario is not designed with this in mind, it sometimes gives rise to rather unrealistic results. As much as possible, a historical scenario should place the actual historical options before the player, and reward a good historicaly valid strategy with victory, while preventing unrealistic or even historically impossible strategies altogether.

This obviously will make the scenario more difficult to win. That's not a bad thing, as most scenario fans are seasoned Civ players who enjoy a challenge, but you don't want to limit the fun to only those who are expert players. The article will conclude with a couple of suggestions to broaden the appeal of a difficult scenario.

Civ 2 is designed to provide a challenge for one human player competing with 6 AI controlled civs, all starting from scratch in 4000 BC and competing for six thousand years. The AI is designed with this in mind, as are the tools given to the human player. But in a scenario that represents a period of 100 years or less, and may have far fewer actual turns than a full game of Civ2, things can get out of balance. The AI reacts inappropriately, and the human player (HP) may have access to technologies and units that would be centuries away in a full game of Civ2.

In the hands of a skilled player these things are magnified, tipping the balance too far in favor of the HP. To counteract this, a designer could simply add more and stronger units to the AI, but that doesn't solve the problem of unrealistic human strategies. In fact, it will only encourage it. Far better to use design techniques to limit the HP to realistic options, and to enhance the capacity of his or her AI opponents.

This article is written primarily for the Test of Time version of Civ2, as it provides the greatest range of scenario tools of any version. Some points may not be applicable to other versions of Civ2, but many are. Regardless of which version you use, I hope the suggestions are useful, and increase your enjoyment of designing and playing Civ2 scenarios.

FRUSTRATING THE HUMAN PLAYER

1. Limit or control units and their functions

Many experienced Civ players use specific units en mass, especially those with special abilities. Large numbers of Settlers and Engineers are often used to transform the map into a mass of mines, roads, railways and fortresses. I suggest not allowing the HP to build Settlers or Engineers at all. If the HP needs 1 or 2 engineers, give them at start or by events.

Ditto trade units and diplomats. If trade units produce lots of gold (depending on what techs are given and the size of the map), you may expect mass trading followed by mass rush buying of powerful units. In one scenario the HP was only allowed to have trade units by events, one for each enemy trade unit captured. Events create trade units with the 'hides'commodity, so the payoff is generally low. Otherwise trade unit payoffs can be kept lower by avoiding using a very large map and by giving certain technologies, such as 'Flight'. See William Keenan's excellent tip 'Advanced Slot Properties' for specifics.

For a human controlled player, building diplomat or spy units can really upset the balance of the scenario, particularly if the HP uses extensive bribing of enemy cities as a strategy. Fortunately, by editing their functions in the Game Text file, you can control which functions a human-contolled diplomat or spy can perform. Eliminate 'bribe' for sure, and you may choose to only permit 'investigate city'.

The functions of AI controlled spies and diplomats are not limited by editing. They will still be able to bribe your cities, steal technologies that you don't want them to have, and can violate alliances you have established. For these reasons, it is strongly recommended that you restrict them to the human player only.

2. Watch cost/benefit carefully

An experienced player will exploit any cost advantage, ignoring any units that don't provide value for the shields or gold needed to build them. More powerful units, and those with special abilities, should cost a premium. For example, a unit that is twice as powerful as another should cost three times as much.

To prevent a massive siege train taking one city after another (a common tactic), consider making siege artillery units as expensive as possible, and add the missile flag, so they are destroyed after attacking. I believe this is realistic, as sieges are very expensive and time consuming.

3. Differentiate units

Maximize the different abilities of units, so the human player must build a variety of types. Differentiation of attack and defence units, those that can attack fortified cities, defend against air attack, those that can move through zones of control, treat rough terrain as road or move onto impassible terrain are all useful. Keep in mind that the AI may not be able to use these units in the appropriate combinations.

4. Limit mobility

The AI can't utilize high movement units as well as a human player. Obviously, fast moving armies allow the human player to conduct much more effective offensives. This isn't neccessarily a fatal flaw - in fact it can enhance the fun of the scenario, but you can balance higher movement with somewhat lower attack factors. This forces the human player to stop from time to time to let his units recover. You can also use 'moveunit' events for AI reinforcements to even things up a bit. (See 'Encouraging the AI')

Don't use railroads except in very limited ways, eg. as suburbs around large cities. They allow the unlimited movement of masses of units from one end of the map to the other, at no movement cost. As such, they are extremely unrealistic and are a huge advantage to a capable HP. NEVER allow the construction of railways.

5. Control diplomacy

For most scenarios, the Civ2 diplomatic model just isn't suitable. AI civs will break their allances with the HP and eventually attack. And they will often make peace after the HP takes one of their cities, instead of continuing to fight. It also gives and opportunity for the human player to shake down the AI civs by demanding gold, sometimes receiving huge tribute. This is unrealistic and undesireable.

For most scenarios, rigid alliances and hostilities are preferable. Keeping the civs from talking keeps them from changing their diplomatic status. Traditionally the 'Negotiation' event has been used, but the ToT 'talkermask' and 'listenermask' version has the advantage of allowing the 'talker' or 'listener' status of specified civs to be changed, using the 'Negotiator' event.


6. Limit improvements

A skilled and ruthless player will privatize (sell off) any municipal improvement that doesn't bring in money or increase production. Don't give cities any improvements not absolutely essential for the scenario, or they will be sold and used to buy units. Think carefully about including things like granaries, aqueducts and recycling depots. If a technological race is not a factor, don't add libraries and universities. Avoid adding Colosseums and Cathedrals to smaller cities that don't need them. If an improvement is not absolutely necessary, leave it out.

Since barracks and Sun Tzu's War Academy have an unrealistic effect, I don't usually include them. Historically, armies usually became proficient from battle, though there may be some cases where you want to represent a superior army at the start of the scenario. Adding one or more barracks at the start of the scenario may be desireable, but I suggest you don't allow them to be built.

Some improvements are critical, especially City Walls, which insure that cities are not wiped out from being captured and recaptured. To ensure that they do get rebuilt, make their build and maintenence cost as low as possible, even 0, so there's no reason for a player to not rebuild them.

ENCOURAGING THE AI

1. War vs. Empire-building scenarios

There's difference between a war scenario and an empire-building scenario. For the imperial scenario, you probably want to give the AI more freedom to make decisions about what to produce, including making more unit types buildable, such as settlers and trade units. But unless you specify that only specific civs are playable, anything the AI can build is also buildable by the human player. And that gives the human player an advantage that may have to be offset in other ways.

2. Know what your AI is up to

A designer should know, and control, what the AI is producing. This is often overlooked. Many players have had the frustrating experience of launching an invasion of a strong AI-contolled civ, expecting to have a real fight on their hands, only to find that instead of fighting for its' life, the AI civ is producing Superhighways or Sewage Systems in most of its' cities. This can be prevented by limiting the number of improvements that can be built, or controlling when their pre-req technologies can be researched.

Directly related is the difficulty in getting the AI to build all of the units you want it to, and in the desired ratios. This is very tricky, and in some cases requires much trial and error. See Leon Marrick's excellent guide "Advanced Scenario Design v.1.9" (section 2a), for an excellent expanation of how the (FW/MGE) AI decides which units do and don't get built. Varying attack/defense and unit costs solves most of the problem, but changing the role to air or naval superiority can also prod a reluctant AI into producing more types of units. Unit slots have some effect as well, eg. the musketeer slot, although this apparently is not the case in ToT. You can tell what units an AI civ will build by using the cheat menu to look at the build menu for the AI civ in one of its cities.


Check out what the AI is doing at various points in the scenario, and make sure it's building what you want it to. A diligent designer will usually play through a scenario several times from start to finish, stopping every few turns and looking at what the AI is doing. You can also play several turns of the scenario using the 'entire map' view, to see how the AI moves and fights. This can be a very long and tiring process, but it eventually stops, and you can see how much the AI depends on random movement.

3. Smaller is better.

Bigger maps favour human players. The bigger the map, the less effective the AI. Some designers can't resist 'gigamps', not realizing they are imparing the effectiveness of the AI in their scenario. The AI can sometimes mount reasonably competent offensives, providing that all units involved are within several squares of each other. Otherwise it will attack piecemeal, one or two units at a time, or not at all. For the same reason, cities should be close together, unless you're representing an unsettled area. Generally, cities should be placed so that their city radius touches or overlaps those of other cities. If your cities are far apart, your map is probably too big.

4. Use events

When it comes to moving units and fighting battles, the AI is no match for a skilled human player, so it must be supplemented by event-generated unit production and directed offensives. To provide a reasonable challenge to a good HP, AI forces should outnumber the human controlled force by about 3-1. Fortunately, the ToT events language allows multiple units (up to 255) to be produced by just one event.

It also allows multiple triggers, so you can set up counter-offensives and ambushes, which provides a welcome challenge to a good HP. Feel free to be as fiendish as you can - the Test of Time events give you enormous flexibility. See William Keenan and Cam Hill's excellent tips for the basics on ToT's events.

The AI's reinforcements should be directed towards strategic targets, not allowed to meander aimlessly. Make frequent use of the 'Moveunit' event to get AI units moving in the right direction. Like other ToT events, this event can be turned on and off by flags, representing multiple triggers. See Captain Nemo's excellent tip on the 'Moveunit' event.



INTERMEDIATE PLAYERS DESERVE SOME FUN TOO

There are a couple of ways to allow intermediate players in on the fun, without having to play the same scenario 20 times before they finally win. The first is obvious, reduce the difficulty level. However, in an event driven scenario, this may not be enough.

To provide a fun and winnable experience for intermediate players, you may wish to use different event files. Simply varying the number of event-generated units is the best way to make a scenario more or less difficult. After a little playtesting, you should be able to tweak the difficulty level by just adjusting a few numbers in your events file. Add a bat. file to switch the files back and forth, and you're set. Good luck!