Lessons learned from Fallout Shelter

I have been playing Fallout Shelter for about a week now. While it’s been fun, that enjoyment is facing diminishing returns, and I’m pretty close to uninstalling it and setting the game down for a while. It’s worth talking about some of the reasons why, since they represented issues I want to avoid in Frasgird.

First, Shelter faces the same challenge of many 4X or sim games: it becomes a exercise in micromanagement, and ultimately a tedious one at that. What I’m finding with Shelter is that as my vaults grow in size, there is no way to automate or abstract away common tasks. I have to click on each and every room that has completed its production; I have to click on each and every character who has leveled up. Shelter should at least have an ‘update’ button that would perform all these actions across the entire vault with a single click; better still, it should have an automate option that allow such harvesting to take place automatically.

Second, and this is tied to the first issue, there’s no growth in your capabilities and powers as the Overseer of the vault. Part of this is because there really is no “Overseer” character being modeled — there are only the other inhabitants of the vault, who go where you drag them to and stay there until you move them again or something kills them. Now, you could argue (with some reason) that your ‘growth’ as Overseer is reflected in being able to run a vault that is getting larger and larger. But most of what you’re doing there is simply tapping on rooms and characters; your major decisions are deciding what type of room to build next, when to get characters pregnant, and when to send characters out into the wilderness (or bring them back).

Which brings us to the third issue: there is no game escalation, that is, no increase in scope (beyond a somewhat bigger vault) or stakes. Part of that is by design — I strongly suspect Bethesda saw Shelter as more of a ‘sandbox’ game, though they didn’t give you a lot of tools to play around with. But there is no arc here and, as noted in my previous post, no victory conditions at all.

Fourth, I really feel no attachment to the characters. Part of this is because there are so many (a vault can hold up to 200), but it goes beyond that. When I play XCOM: Enemy Within, I really focus on building a team of a dozen or so characters. I always use the ‘Not Created Equally’ (varying initial stats), ‘Hidden Potential’ (varying growth in stats), and ‘Training Roulette’ (somewhat randomized training tree) options; the result is that each new team really does feel unique, and the members feel like individuals. In Shelter, it’s really just the armor and weapon that you equip a given character with that matters.

Finally, Shelter has a few technical (vs. game design) issues. For one, Shelter’s touch-based interface makes it difficult to perform critical actions. When there’s a threat — raiders from outside, roaches and fire from inside — you often need to drag characters from nearby rooms to combat it. However, since you also touch the screen to move scroll around the vault, you often find yourself sliding the vault around the screen instead of moving the character you thought you selected to the point of threat. Another is that Shelter gives you no easy way (that I’ve found so far) to review the armor-and-weapon status of all your characters. That information is not in the character list window — it only appears when you bring up an individual character’s window — and there is no way to cycle through characters in that window.

So, what does all this mean for Frasgird? I think it suggests the following as game design goals:

  • The game should move through phases of increasing scope and concern. In the early portion, you may need to make lower-level decisions and even micro-manage a bit, but as the game goes on, that should be abstracted away so that you can focus on larger issues. At the same time, you should have new opportunities unlocked — for example, multi-building combinations that can be constructed and managed as a unit, with higher and/or more valuable production.
  • Your character, Elare, should likewise grow in capabilities and influence as you successfully move through the game. I am reminded of Space Rangers 2 (another one of my favorite games), which allowed you to increase your ‘command’ attribute as the game went along, which in turn meant that you could recruit one or more other ships to follow yours (so long as you were paying them). Likewise, as you succeed in growing the colony, you should be able to attract characters to follow you.
  • Characters should feel unique and valuable, not interchangeable, and they should be different enough so that you have to adapt your approach to fit the characters you have available. For example, in XCOM, I may end up with a Sniper who has the Field Medic trait (3 medpaks instead of just 1), or a Heavy who has the Sprinter and Lightning Reflexes traits, and so I end up using them in a different way. Likewise, in the Fall From Heaven 2 mod for Civilization IV, character advancement was sped up from plain vanilla Civ4, so you ended up with more powerful characters that could make a difference — but whom you also guarded more carefully.

More as I think about it.

Filed Under: Game DesignMainOther Games

About the Author

Webster has been doing game design since 1980, but only has one actual published game to his credit -- Sundog: Frozen Legacy (Apple II, 1984; Atari ST, 1985). This is his second.

Comments (1)

Trackback URL | Comments RSS Feed

  1. […] play it as a roguelike game (high odds of failure; permadeath of character; as I mentioned in a previous post, I tend to focus on building interesting teams. But there’s a problem: once I’ve built […]

Leave a Reply




If you want a picture to show with your comment, go get a Gravatar.