What makes a good simulation game?

I've been thinking a lot about simulation games lately (here's why, if you're curious). I had pretty much all the Maxis games as a yoot: SimAnt, SimEarth, SimCity, SimLife... and later on, I also played the Sims. I think there's much to be learned from comparing & contrasting these, but I'd like to get your thoughts on it.

To get the ball rolling, let's compare SimEarth to SimCity. On the surface, managing an entire planet sounds like enormous fun. To some extent this could be considered a god game. And in fact I did pour quite a few hours into it, but in practice, it was nowhere near as fun (nor as successful in the market) as SimCity. I think the reason for this is simple: when a SimEarth game is going well, there is nothing for you to do. Your planet evolves on its own, and doesn't benefit from your meddling. So you are pretty much a spectator, most of the time.

In SimCity, on the other hand, there is always stuff to do: until you lay some roads and zone some zones, nothing can happen at all. Then you have to make sure your zones are supplied with water and power, and carefully place municipal buildings, manage the budget, put out fires, and so on... without active guidance, at best your city will stop growing, and at worst it will be plagued by crime, fires, pollution, etc. There is still the automatic-simulation element (buildings appearing or being abandoned) that you can't control, but this is balanced by a lot of things you must control. So you stay busy and feel productive.

SimLife was close to SimEarth, in that evolution pretty much happens, though I would always try to monkey with things by dragging creatures here or there for example. Those efforts rarely paid off, though. So again it was hard to feel like you were a productive participant.

SimAnt, on the other hand, was closer to SimCity (or maybe we should say it the other way around, since SimAnt was the first). There's always plenty to do, and without your guidance, your ants won't do much... until you get to the "multiple squares of the yard" stage, when you discover that the yard squares you're not actively managing do just fine without you, often better than the one you're in control of.

So. To summarize, I guess I'd say that simulation games must strike a fine balance between automatic evolution & events, and giving the player plenty of tasks they must do to advance.

But what do y'all think?

 

It seems to me you are basically tackling the "how much micromanagement?" question.

I think different types of players have different answers so what one finds fun another won't. Ie. I am more of a "planner" and I am happy to queue up items, pan around the world, chat with people, and think about the next major thing I want to do while my order queue completes. Others want to click 50 times a minute dragging things around and ordering units to and fro(I am using RTS equivalent since that is what I am working on).

Personally I think the answer is more freedom rather than less, but with optional computer assistance to level the playing field. Others will argue you should design your game for one type of play and just perfect it, and perhaps that is correct instead.

As an example(sorry going back to RTS), lets say I always want X number of resource harvesters out getting resources. Traditionally you would have to monitor their progress and if one died go to the unit building and create another and set him up to go gather again.

I find this tedious when I am busy conducting a battle halfway across the map as I like to focus on one thing at a time. Some players like that challenge and want to micromanage it. My solution is to set a threshold of minimum ore harvesters and just spawn more if it drops below a certain number, perhaps even have them start auto searching to the closest ore by themselves. The option still exists to do all of this by hand of course and not use the auto-spawn. It could be slightly slower and more inefficient than if I did it manually but if it is "good enough" it would at least ensure my economy is semi-functional until I can shift my focus back to home after the battle, instead of sitting there doing nothing and putting me light years behind my opponent who is a master clicker/multitasker, all due to a lone wolf that went wild on my harvesters.

 

I think having an active hand in the development of the main subject of the simulation is essential for a great simulation. That means being able to direct what is happening, and micromanage it beyond that.

In SimCity, you could not only manage the zoning of your city, but if you didn't like a building, bulldoze it! Something else will pop up in its place. It's micromanaging, but it's part of the experience for many people.

I think it's also important to start small and end up epic. Going from that flat plain of grass to a roaring megalopolis was a great experience for me, marred only by the time I was forced to spend just watching it... Or letting it run overnight on automatic.

Complexity is a good thing in a simulation. SC tried to remove things like power lines and sewers, and many protested. It's like a kiddy's sandbox now in comparison. Don't even get me started with the neighorhood-sized cities in the new one.

This is a personal preference, but I don't like resources that run out. If you build your city on coal production and it runs out, your city could die. That's fine for a scenario that you could run in a few hours, but if you're building for the long-haul, that's just not fun.

Each simulation has things that are vital to it. SC's traffic and people movements, for instance. It was so messed up in the new SC that people practically rioted. Traffic and people are such important parts of the city that I can't believe that they tried to simplify them *and* managed to screw them up so badly, too.

Personally, I like upgrades to come with a plot. SC introduces them when the city has need of them (and later when you buy them) but I liked it a lot better when there was a story behind everything and doing well meant I got to see more of it.

 

Yes, exactly, it's a question of micromanagement. Though that term itself has a bit of a negative connotation, implying that players are likely to want less of it. I think SimEarth and SimLife show that it's possible to go too far the other way, and not offer enough management.

Civilization is another good example, I guess because it sits sort of halfway between a traditional strategy game and a sim game. Some versions of Civ don't let you queue up your production, what was fine at first, but I eventually got annoyed with it because I pretty much always wanted to build things in pretty much the same order, and towards the end game, you spend a large portion of each turn going around just telling each city to build the next thing on the list.

A simple production queue would have helped this quite a lot. But here's another idea: use machine learning to train an AI city manager to predict your choices. Whenever a city comes up for its next production, auto-select the predicted choice so you can just smack OK to move on. Then, have an "auto-produce" checkbox somewhere that prevents the dialog from coming up at all, which you can use once you have confidence in your automatic city manager.

I have a related idea for High Frontier: most buildings will spawn automatically, but you can click on any one to bring up a little info panel. On that info panel will be "thumbs up" and "thumbs down" buttons. By clicking these, you reward or punish certain types of buildings (based on purpose, style, etc.). So, if you want a futuristic city with lots of sweeping tensegrity spires, you reward those, and get more of those automatically. If you prefer a retro look with lots of brick vine-covered buildings, you reward those instead.

I guess I'm hijacking my own thread here a bit, but I do think there are lots of opportunities for some fairly simple machine learning to really balance out the whole micromanagement issue, by learning from the player's choices.

 

When you can do, some cool wacky stuff in the simulation game. Alot of simulation games, sometimes don't have many fun things to do, except do things that are expected in a sim game, which sometimes makes them feel basic, and a bit barebones and limiting. But since hard drives with larger storage space, and dlc have been a normal thing for over a decade, implementing some cool wacky dlc stuff in a sim game, shouldn't really be a problem nowadays.

 

I can't argue with this, but I will point out that this is looking at it from a very "game design" perspective. To broaden the perspective a bit, consider that Will Wright, who (I believe) designed all of those sims, actually doesn't consider them to be games.

Instead, Wright sees his stuff as something similar to a "modern Montessori toy". These are basically toys which are designed to teach something through explorative play. There are no explicit win or lose states (which is why it's not a "game"), but in the sim games there are often player goals which are given which take on a similar role. For instance, in Sim City your people might complain about something and the player takes on the goal of fixing it, or someome might request something and the player takes on the goal of providing it.

I haven't played SimEarth, but I wonder if a part of the reduced engagement compared to Sim City is that the thing it's simulating doesn't naturally give rise to such directed goal setting? In Sim City it's an integral part of the simulation that people will complain about or request things. That's how cities work. In Sim Earth is there any similarly integrated mechanism that results in clearly defined goals being suggested to the player?

Personally that makes a huge difference to my own engagment in a game. If I have goals my mind focuses on finding solutions to meet those goals. If I don't have goals then it doesn't really matter how cool and sophisticated the underlying systems are, I'll tweak knobs for a bit and see what happens and then I'll lose interest as soon as I have a superficial grasp of the underlying patterns. There's a huge difference in engagement (for me) between messing around and seeing what happens vs. having a specific, target state to reach and having to figure out how to get there using what control, influence and resources are available to me.

Anyway, all of that is to say that if you think about it less as a "game" and more as "toy + goals" you might come up with different insights into the design of these things.

 

A good simulator game is a fun game that convinces players it’s a simulation regardless of reality.

 

A good sim in my view is one that will be as close to reality as possible. An even better will have the option for the player to choose the reality level.
Otherwise one may be stuck doing something completely useless and boring (inspecting the tires before a take off in an airplane sim for example), instead of doing what he/she wanted to play the sim for.