A turn-based artillery match has a problem that action games hide: the score is visible, permanent, and public. If a player is behind by 60% health on turn six, nothing is happening on screen to distract them from that fact. They are one bad throw away from deciding the match is over and closing the app.
I build a game in this genre — Cannon King: Wind Tank Battle, a mobile turn-based artillery game with 48 tanks, 20 destructible maps and 53 stages, made by one person. The most useful design question I worked through was not "how do I make the loser feel better". It was: what makes the next shot still worth calculating?
Four things did most of the work.
1. The wind is a shared, visible variable
Wind changes every turn, and it is printed for both players. That single decision does more for comebacks than any rubber-banding I could write:
- A player who is behind can be handed a tailwind that turns a desperate long shot into a realistic one, and they can see that before firing. The hope is not a hidden dice roll; it is a number on the screen.
- The same wind that helps them changes the leader's next turn too. Wind is symmetric, so it does not feel like charity. Nobody can say the game helped them.
- Because it varies turn to turn, no lead is stable in the way a health bar's lead is.
The unit of hope in this genre is "this shot is possible", not "the game is being nice to me". Wind delivers the first without doing the second.
2. Terrain is a resource, not a backdrop
Craters persist for the whole match. That turns the map into something both players spend and shape:
- Digging under an opponent takes their high ground away without dealing damage. A player who is behind on health can still be ahead on position.
- A crater also works as a wall and as a ramp, so a losing player's shots start landing from angles the leader has not been defending.
- In the late match, both players are playing on terrain the losing player partly authored. That is a real form of agency that does not require a stat advantage.
The design rule I follow: if an action can change the shape of the battlefield, it counts as progress even when it deals zero damage. Players who understand that stay in matches they are losing on health.
3. One turn should be able to change two things
A comeback needs a shot that is worth more than its damage number. In practice that means designing for shots that resolve a threat and create one at the same time: clear the ground under an opponent, and you have removed their height advantage and opened a new firing line in the same turn.
This is also why the item loadout is chosen before the battle instead of bought mid-match. Four slots — teleport, shield, double shot, heal and the rest — are a plan the player writes in advance, and a plan is something a losing player can still execute correctly. Buying your way out mid-match turns a comeback into a transaction; committing to a loadout turns it into a decision.
4. Remove the modes where comebacks are impossible
There are Free-for-all, Team and no-item matches. The no-item mode matters more than it looks: it is where a player can lose because of aim alone, without also losing to an economy they have not learned yet. Being able to name the reason you lost is what keeps a player practicing instead of quitting.
For the same reason, every tank in the roster carries its own shell and skill rather than being a bigger version of an earlier one. If the only difference between two tanks is a damage number, the losing player's correct move is always "grind for the bigger number". If the difference is a skill, their correct move is "pick the tank that answers what just happened to me", which is available immediately and for free.
What this costs
A symmetric, legible comeback system makes match length less predictable. Some matches run long because a player keeps finding a real shot. That is the trade I accepted: I would rather have a match that is decided on turn twelve with both players calculating than one that is decided on turn three and played out anyway.
None of this is hidden difficulty scaling. Every number that helps the losing player — wind, terrain, the shop-free roster — is visible to both players on the same screen. The design goal is a match where being behind changes your plan, not your odds.
Cannon King: Wind Tank Battle runs on a web build wrapped in a Capacitor shell for Android and iOS, which is how one person keeps shipping both stores from one codebase.
- Google Play: https://play.google.com/store/apps/details?id=com.cpcai.potanwang
- App Store: https://apps.apple.com/kr/app/id6827588104
- New players: the code POTANKING2027 in settings gives 5,000 credits and 10 premium draw tickets, once per account, valid until 2027-12-31.
If you build in this genre, I would like to hear which of these four you would cut first. I am @cpcai_studio on dev.to or [email protected].
Top comments (0)