Tuesday, April 6, 2010

Tower Defense Design Proposal

The following's from an email I recently sent to a fellow who's started messing around with a flash tower defense game. I don't normally share this kind of "game design consultant" work here; I probably should.

Let's say we go with a generic fantasy premise, just for the purpose of this description- I can handle the fiction and any writing needs. The "Overworld" gameplay in between rounds is similar to the Last Stand 2. You see this overworld map/matte painting-esque image depicting the ancient, overgrown ruins of a once-magnificent city, with a variety of locations marked out on it. A bar across the top depicts the monster types that're stalking you and will attack that night- clicking one displays information regarding their strengths and weaknesses. There'll typically be one you haven't fought yet; their info consists of a black silhouette along with some cryptic lore containing hints as to how you fight them.

Any unexplored locations are described via similar cryptic hints, giving the player some indication as to what benefits might be found by searching there (especially whether they've got the upgrade that's going to be effective against a given incoming enemy- you have access to the day's search results when choosing your battle loadout). Each location also contains a note as to how defensible it is (there's a number of factors we could use to actually alter the challenge enemy waves pose, lots of potential for each location having unique quirks). You'll have to actually go to a location to get the description that has concrete details.

In other words, you're simultaneously choosing where to search for upgrades and which stage you'll be fighting the next wave on; naturally, the most dangerous locations will contain the coolest upgrades. I like the idea of a player who's had the last battle go badly being wounded/vulnerable, but because of how the game's set up their response is to retreat to a safe location and just rest in preparation for the next night (rather than just hitting reload).

Next, there's the actual battles. This concept and the overworld concept were brainstormed together, but neither needs the other. I'll skip a bunch of micro-level design ideas and just focus on the tower setup, as sketched out in those cruddy webcam pics.
  • On the left side of the 1st pic, the UI displays 9 items.
    • The 3 tower types you'll have throughout the game- +Damage, +Range, and +Chain (AoE is alot tougher to balance, even before mana effects come into the picture). The display also shows the random positions of the 3 link nodes on the next tower of that type you'll place.
    • The 3 mana generator types you've chosen for that round. In addition to providing a bonus effect to any tower they're linked to, they generate mana at a rate equal to (X/[1+# of linked towers currently firing])- 1 active tower reduces its mana generation to half capacity, 2 active towers reduces it to a third, etc. The current mana supply for each type would be listed next to its icon.
    • The 3 (?) spells you've chosen for that round. You use mana to cast these- I expect you'd click the spell, then click the mana type you'd use to pay for it, with each spell having one or two favored mana types that can cast it for a third of the normal price. (This'd be an extra motive to explore in the overworld- you've found this great spell, but not the associated mana generation rune, and it sounds like it might be in one of the most dangerous stages. . .)
  • Every X seconds, you can place a new tower or generator.
    • If you want some mana types to be more powerful, they could incur a larger delay before you can place your next item.
    • Since linking towers enhances both of them, you'll often want to start by mostly placing towers
    • X could get longer as the round went on.
    • In the 2nd pic, our hypothetical player started by placing a +Range tower (with the node links the UI displays in the 1st pic), then a +Damage tower, and then a generator for Death mana.
  • Node links can cross open spaces (and go under the monster path) so long as there's another placement point on the other side.
  • Each tower has its own starting upgrade, plus up to 3 more upgrades from the generators and towers it's linked to. (So in the pic, both towers are getting +1 range, +1 damage, and whatever Death's benefit is- say, bonus magic damage equal to 10% of the target's max HP, making it good vs waves of a few tough targets. Don't worry about bosses, they'd- whoops, I'm getting into micro-level design.)
    • We can screw around alot with how many nodes a tower gets and how many upgrades they can have max, along with a number of other sub-matters like whether links should be two-way for towers.
    • Multiple upgrades of the same type stack, though the exact progression depends on the type for balance reasons. A +Range tower linked to one other +Range tower shoots 4 times as far as the base distance; if all 3 of its links are to other +Range towers, it can shoot enemies on the other side of the map (though it'll be dealing the base damage with no side effects).
So the experience of playing the game consists of placing towers and generators in interconnected networks, and using the spells to influence things. (As for how much influence those spells would have. . .imagine if instead of having to click to pick up the sun in PvZ, every [40/sunflowers] seconds you could make a few clicks to either deal 3 peas' worth of damage to a zombie or make one lane fire down a neighbor for 5 seconds).

I probably don't have to explain this, but one of the design priorities here is elegance- rather than giving the player forty options, give him a few meaningful choices that combine to produce the same flexibility. I'd guess that there'll be no more than a dozen mana types, and half of those are 'bonus options' (a normal playthrough will find 2-3 by the endgame). But the potential for combinations between different upgrades means that your 3 chosen mana types are going to play in a way that's noticeably different from other combinations.

No comments: