While writing another post, I stumbled upon what I felt was a rather interesting case of two games solving the same solo mode problem in two completely different ways. The two different approaches and the difference in the results achieved struck me as interesting and I hope that you’ll agree .
The games I’m talking about are Sagrada and Viticulture. Before we get started it’s important to state that I made the solo mode for Viticulture, so please keep that in mind that while reading in case of biases in what I write.
The similarity between the two games
Sagrada is a dice drafting game where you take dice from a pool shared with the other player (I’m focusing on 2-player, since that’s what the solo mode simulates). You use those dice to score VP independently of what your opponent does and taking the dice has the added benefit of denying them to your opponent (assuming they are useful to her). At the end of each round new dice replace the current pool of dice to make a full new pool available.
Viticulture is a worker placement game where you place workers on action spaces from a pool of action spaces shared with the other player. You use those action spaces to score VP independently of what your opponent does and occupying those action spaces has the added benefit of denying them to your opponent (assuming they are useful to her). At the end of each round all workers are returned to their owner to make the pool of action spaces available again.
So, while at the surface level the two games might seem very different because one is a dice drafting game and the other is a worker placement game, they are similar in this core aspect. This was something that surprised me, because I hadn’t considered that drafting and worker placement are so similar, but maybe that’s just me being dense .
Having now set the scene we’ll dive into the bots for the two games. The bot in Viticulture is called Automa and instead of referring to the cumbersome “the bot in Sagrada” and “Sagrada’s bot” I’ll allow myself to call it Gaudi after the architect who designed the church that lends it name to the game.
The dice to be drafted in Sagrada. Image credit: Eric.
The difference in how the solo modes handles the situation
Since dice drafting/worker placement is at the core of both games any bot worth its salt needs to deal with it. In Viticulture, Automa places all of its workers and then the player places hers.
* When playing against Gaudi you get to pick your dice first and the remaining ones are given to Gaudi.
* When playing against Automa it gets to place all its workers first and then you place yours.
So, the two games come at this situation from opposite sides: Sagrada lets the player take all actions first and gives her the opportunity to deny Gaudi the dice that are best for it. Viticulture, on the other hand, lets Automa take all its actions first and thus deny action spaces to the player.
I think it interesting to consider the consequences of these two solutions to the same problem.
The workers to be placed in Viticulture. Image credit: E P.
Consequences of the two approaches
There are several design goals at play when considering the consequences of the two design approaches:
1) Minimizing the workload on the player for running the bot.
2) Allowing the player to use denial of options against the bot.
3) Allowing the bot to use denial of option against the player
Both Gaudi and Automa satisfies 1) but Gaudi goes for 2) while Automa goes for 3). The difference might seem small, but to me it makes a big difference in the feel of playing against the two bots.
Going for 2) removes the tension of whether your opponent grabs the die you need and gives the game a calm and puzzly feel, which I think gels well with the puzzly feel of Sagrada.
Viticulture’s approach of going for 3) keeps the tension you feel while waiting to see whether your opponent snatches that one action space you need and having to work around the impact of her action space denial. On the other hand, you lose the option of denying action spaces to your opponent.
I’m not saying that either of these are best, they simply give very different feels: Puzzle vs. tension and which you prefer is, well, a matter of preference. Furthermore, Sagrada adds in some of the tension by making it unknown which dice will be available next round, so the tension comes in the form of seeing whether the new dice are what you need. Viticulture on the other hand adds some of the puzzly feel by letting you do several worker placements in a row and so you must consider a sequence of actions just like in Sagrada where you make your decision before Gaudi.
So, the two games move a wee bit towards each other, but in my opinion, they’re still miles apart in how they make the game feel.
The middle ground
We’ve now seen how differently the two bot systems approach the same game design problem. They’re at the opposite extreme ends of the design space, but there’s of course also a middle ground: Having the player and the bot alternate in picking a die/action space.
In Viti I didn’t do this because I wanted to make running Automa as smooth as possible: It requires almost no work for the human player and so even something as simple as placing the workers in an interleaved manner would probably double the time spent by the player. That said I might choose the interleaved system if I were to redesign the Viticulture Automa today.
The reason for this is that it simulates the feel of the multiplayer game better and lowers the chance of the player getting shafted by the luck of the draw. I probably still wouldn’t make the player’s action space choices have any impact on Automa, since that would require increase the complexity of Automa quite a bit (relatively speaking).
The three approaches apply to bot interactions in general
To sum up, we have three approaches to the option selection/denial situation in bot-systems:
1. The player always goes first.
2. The bot always goes first.
3. The actions of the bot and the player are interleaved.
As you might have guessed based on the way I phrased this, the three approaches apply not only to the type of situation we’ve discussed here but to any turn/action order system in interactive parts of a game and in most cases the same trade-offs (such as tension vs. puzzle and smoothness vs. realism) and consequences must be considered as discussed above.
A blog about solitaire games and how to design them. I'm your host, Morten, co-designer of solo modes for games such as Scythe, Gaia Project and Viticulture.
Sagrada Bot vs. Viticulture Automa: Two completely different approaches to handling drafting and worker placement
13 Apr 2018
- [+] Dice rolls