Thursday, November 18, 2010

Endgame Slowdown in 4x Games

Most 4x games share a fatal flaw -- endgame slowdown. Everyone who's played civilization has experienced this issue: at the start of the game your turns take 5 seconds, and by the end of the game you're spending 3 to 4 minutes per turn. This starts to significantly delay the positive feedback that you get from making good decisions, forcing you to make a decision (I'm going to attack Catherine), and then only see the result an hour later.

Some games have no mechanisms to combat this issue, but Civilization 5 came up with an interesting solution, by giving special benefits to small empires. These small empires (1 - 4 cities or so) get to advance along an alternate tech tree at a faster pace than larger empires do. Additionally, when conquering an enemy city you can choose to incorporate it fully into your empire, or turn it into a puppet state. Puppet states still provide you with resources, but are uncontrollable and don't count towards your total number of cities. In this way, a player can continue to only manage a handful of cities all game, all while building, conquering, negotiating with other leaders, etc. Since they released Civ 5, I haven't been able to play through one game with a large empire. I've tried, but every time I give up after a few hundred turns and start a new game with a small empire. I don't think I'll ever go back.

For Deep Field, I wanted to have a mechanic that similarly reduced the amount of endgame slowdown. I decided to incorporate a 'mothership' token, which essentially forces the player to only take action at a specific location. If they want to colonize a new star system, they have to move there and colonize it themselves. If they want to build buildings, they have to fly to the system and queue the buildings up. If they want to attack an enemy star system, they have to transfer some of their fleet to their mothership, fly to the system, and initiate the fight. This system means that a turn takes the same amount of time during all periods of the game.

The mothership also had additional effects on the flow of the game -- at the start of a game you spend most of your time colonizing, building, and exploring. By the end of the game, most of your time is spent fighting enemies and conquering planets. Additionally, it forces interesting strategic choices and encourages players to play by feel rather than working out the perfect set of moves for any particular situation. I feel that overall the mothership mechanic has been a big success.

4x Lesson Learned
Endgame slowdown is something to try and avoid. No one wants to spend 10 times as long to see the result of their decisions at the end of the game compared to the beginning.

Monday, November 8, 2010

Active Opponents in Short Games

When I started work on Deep Field, I wanted to make the enemy empires fully simulated -- they would explore, colonize, grow, declare war, etc. As I built out the A.I. for the enemies, I started to realize that it felt fairly shallow. Battles were always front-loaded, with all of the fleets of the respective empires clashing in one massive confrontation near the border. After this battle, whichever side had more ships left would quickly steamroll over their enemy's solar systems, ending the war.

When the player was on the losing end of this, it felt unfair -- the player didn't have time to react to mistakes that they had made. Obviously there are circumstances where the player deserves to lose completely for making a truly egregious mistake, but most of the time mistakes should be equated with setbacks, not with losing completely.

To make things worse, it felt shallow when the player was on the winning side as well. There was no give and take to empires fighting. When you win one fight and then defeat the enemy empire wholesale, it's fairly boring.

Eventually I realized that the better option was something like the campaigns from RTS games. In Starcraft 2, the enemies in the campaign level aren't fully realized A.I.'s. They start with a huge amount of structures, units, and resources, and then the player whittles them down over time. This makes the game almost feel like a bit of a dungeon crawl. That's the feel that I realized I needed, and so I changed the game to have enemy A.I.'s that attack you every now and then, but don't colonize new stars. The player defeats larger, more powerful opponents one at a time, as they progress through the dungeon of nodes.

Flash 4x Lesson Learned:
Fully realized enemy empires feel shallow in short-form games. The game is decided on a knife-edge, which isn't as interesting as a slow steady advancement as you defeat chunks of the enemy's empires.

Monday, November 1, 2010

Local vs Global Resources

Near the start of my work on Deep Field, I didn't have any upkeep mechanics for income. This meant that both A) you could have any number of ships, since your empire constantly produced them, and B) Your income tended to skyrocket faster than you could use it. I tried solving this issue with increased colonization and alliance costs, but it always felt a bit off. Additionally, the lack of a cap on the number of ships you could have meant that you tended to flip from being underpowered to overpowered, if the enemies left you alone for 30 turns too long.

Overall, it just felt bad. I realized at the time that there's a significant difference between how you should deal with local and global resources.

One of the core skills to learn in most 4x games is judging when to invest resources in a new colony, city, etc. These new resource producing objects cost a fair amount, both immediately, and over time: they take many turns to build up. Eventually they'll be worthwhile, but at the start, they do very little for you. However, if you can use a global resource to start or speed up the development of these new colonies, you can focus your whole empire's output on a small set of locations. This leads to significant and honestly unbalanceable increases in player power.

Deep Field has two local resources: Food and Construction, and two global resources: Income and Research. Research is internally balanced because I set the cost of every single researchable technology by hand. Income needed to be balanced, and I think I've done a decent job. Using a slider, you can set what percent of your Income you want to go towards maintaining fleets. Going from 0 to 100% fleet maintenance should take the player about 40 turns (if they have equal Income and Construction).

This means that they have to front-load the majority of the building time. If they're under heavy attack at 0% fleet, it's likely that they will just lose their entire empire. A player needs to stay at say... 30% to have a hope of defending themselves. And as they go up in difficulty, this number rises and rises. Still, they always want to keep it as low as possible, so they can make more money to start more colonies. This interplay makes for satisfying gameplay, and the short game lengths means that it isn't too frustrating when you gamble and lose.

Tying together the fleet cap and income, and putting in upkeep for buildings, stars, and fleets, has dramatically improved gameplay.

4x Lesson Learned:
Global resources need to be capped or limited in some way. A player's empire will change in size and output dramatically over the course of the game. Allowing them to focus the entirety of their empire's power in one location will end up being incredibly difficult to balance.