Old World Designer Notes #3: One Unit per Tile

The following is an excerpt from the Designer Notes for Old World. The game, a historical 4X set in classical antiquity, released on July 1, 2021, and is available for purchase here.

The big change that always gets mentioned when going from Civ 4 to Civ 5 is one-unit-per-tile (1UPT), which is interesting as 1UPT is purely a mechanical – as opposed to thematic – change. (The introduction of religion in Civ 4 or city states in Civ 5 would be examples of thematic changes.) The change was intended not to make the role of combat different in the game but to make it more interesting. Jon Shafer, Civ 5’s lead designer, was inspired by turn-based games, such as Panzer General which were laser-focused on combat mechanics and benefited from the visual transparency of no stacking.

(Many of these tactical games also used hexes – clearly superior to squares from a tactical perspective – which helped the series finally adopt them. I had considered hexes for Civ 4 but held back from a long-held prejudice inside Firaxis that hexes would scare off mainstream gamers. Indeed, many have commented that Civ 5 finally made hexes safe for video games. I was content to move Civ 4 back to a top-down chess-board grid and away from the visually misleading isometric squares of Civs 2-3, where horizontal and vertical distances were separated by a factor of two.)

Generally speaking, opinions were divided over (although largely in favor of) the success of one-unit-per-tile. Most felt that 1UPT did increase tactical depth but many others began complaining about the “carpet-of-doom” which dulled potential depth by filling the world with so many units that simple movement became difficult. The term carpet-of-doom is an ironic reference to the “stacks-of-doom” of Civ 4 where an AI could invade with 20+ units stacked on a single tile. Few people called for the return of stacks-of-doom, but critics pointed out that carpets-of-doom were just as bad.

I watched the debate with interest, glad perhaps that it wasn’t my problem to solve this time. I felt that Jon had made the right choice with a radical change to the combat system – each new Civ needs to give itself a fundamental reason to exist – and I also wasn’t surprised that it was not necessarily an easy process. From my own experience, I felt that the main issue with 1UPT in Civ 5 was limited unit mobility and a lack of space between cities for maneuvering. Tactics-focused games like Panzer General tended to have a preset number of units which were chosen to fit the size of the map, a design luxury not available in a 500-turn 4X game where a dedicated player could have hundreds of units by the end of the game.

Although I saw that Civ 5’s 1UPT had become the mainstream conception of a tile-based 4X game, Old World actually went through roughly a year of development WITH stacked units. Early on, the simplest way to prototype the game was to maintain a strict 1UPT rule (as stacking has the major downside of making the UI much more complicated). However, I knew that at some point we would need to experiment with stacking to at least allow military and civilian units to share the same tile. (Civ 5 & 6 both allowed a military and a civilian unit to share the same tile, and everyone in the world considered that a good thing, a rare moment for game development!) So, when I added the code to allow multiple units per tile, I decided to try to find a happy medium between encouraging 1UPT in practice while allowing stacks when necessary for managing large armies.

My theory was that if stacking was possible but also potentially very dangerous, Old World would have the benefits of both systems. Thus, our movement system allowed infinite stacking, but the cost would be that, when attacked, every unit in the stack would be damaged just as if it was the primary defender in non-1UPT Civ games. There would be no “top unit” which would shield the units underneath it from damage. This system fit very well with the lack of defensive retaliation in Old World – attacking a tile only damaged the unit on the tile, so attacking a stack meant damaging every unit in the stack.

In practice, the system worked great. Players generally kept one unit on a tile, both to avoid suffering extra damage as a defending stack and to maximize the number of tiles covered by Zones-of-Control. Plus, players often faced an interesting decision during combat when deciding whether to stack an extra unit onto a tile to get a second attack, especially if they needed the extra attack to kill the defender. Doing so would be very tempting, but the cost would be exposing stacked units to multiple hits per attack during the enemy’s turn. Tempting players to make an impulsive decision for a short-term gain but at the risk of a bigger loss later is solid game design. I had cracked the problem of how to combine the tactics of 1UPT with the convenience of stacking. I was very pleased with myself!

The only problem was that I seemed to be the only one who was pleased. The rest of the team ranged from tolerant of the system to subdued hostility. In due course, another mutinous mod appeared (see the section on Orders). I agreed to give it a try, and to my surprise, I sort of seemed to prefer the simplicity of 1UPT. Unlike the rest of my writing on Old World’s design, I’m not sure I can quite articulate why 1UPT worked better than the stacking system I had implemented. I can still explain what was great about giving the players a reason to stack and a reason not to. Look at the interesting decisions! Marvel at the trade-offs! (I am reminded of Sid joking that the way to respond to negative feedback is to explain to players that “You just didn’t know you were having fun!”) In reality, I had bumbled into the best system for Old World with some strong encouragement from the team.

The simplest explanation for why 1UPT works in Old World is that it‘s actually a perfect fit for the game’s unusual combat system, which splits combat resolution across multiple turns by requiring multiple attacks to kill a unit and minimizing retaliation damage from the defender. Allowing a strong defender to damage an attacker significantly was never going to work with this system because doing so is akin to giving the defender a free attack on the attacker’s turn (and without spending an Order). Defenders always tend to be at an inherent advantage, both in real life and in strategy games, so anything that boosted their advantage even more would be a mistake. At its best, the lack of retaliation damage pushes players to go on offense, to take risks, to take an active role in combat, which is inherently more fun. A game where you take an active role killing enemies is much more engaging than one where the AI annihilates itself against your stationary defenders.

Because Old World expects combat to last multiple turns, allowing the player to overload a combat front with extra attacks (via stacking) was sapping away an important part of our unique combat system. Counter-attacks (and retreats) are meant to happen during the other player’s turn, and giving the player more opportunities to focus fire via stacking to kill a unit short-circuited that dynamic. 1UPT has multiple other advantages over stacking (including a cleaner UI, more cohesive fronts, and simpler combat rules), and we never looked back after re-implementing it.

In many ways, the Orders system, limited city sites, and one-unit-per-tile all need to be viewed as part of a single holistic system where each part buttresses the other. The enforced distances between city sites ensure that there is enough space between each city for a battle to be fought, instead of ICS-style cities inconveniently clogging up the battle lines. The Orders system prevents 1UPT traffic jams because the game allows a unit to make multiple moves with a single click and to covertly stack on another unit as long as it’s not the final destination. (Civs 5 & 6 do the same thing by allowing units to pass through each other, but because they are still limited to one move per turn, the feature is of limited use.) 1UPT, on the other hand, helps balance some of the extremities of the Orders system by making it more difficult to kill a unit via stacking. Each system cannot be viewed in a vacuum, and debates over the costs and benefits of each need to be made within the context of the game as a whole.

3 thoughts on “Old World Designer Notes #3: One Unit per Tile

  1. Pingback: Soren Johnson on 1UpT in Old World | CivFanatics

  2. As an engineer I have thought AI would have better time with softened 1UPT like penalizing stacking as described in this post. It is interesting to know that it was not well received. Was it because it was too easy to unintentionally overlap units due to path-finding? Or perhaps they simply didn’t find it as an elegant solution?

    Anyway, great post. Thank you for sharing your experience!

Leave a Reply

Your email address will not be published. Required fields are marked *