Believability instead of realism
I would argue that realism is not bad per-se. However, realism should not be your objective, and it is not good excuse - by itself - for introducing game mechanics. Instead the place of realism in video games is as a particular esthetic, which you may want or not, but that is another topic.
Instead, you should aim for immersion. And believability is your main tool for that. It means that you should not aim to create something that matches our reality, but something that players can believe about your fictional world.
This is also a good principle when creating animations or character designs. Your world can have crazy creatures that cannot exist in reality, and nobody cares. But if their physicality is wrong, they lose believability, most people will know it does not look right.
Often times excluding things helps believability, because once you portray something, players can ponder if that makes sense. Meanwhile, if the soldier at war does not have the need to go to the bathroom, it will not break immersion because everything else keeps the player engaged.
Your base line for believability is popcorn/fridge logic:
You know. You’ve just come home from a movie, you had a great time, you go to the refrigerator to get a beer, you open the door, and you say, ‘Wait a minute …’
-- Jonathan Demme
However, sometimes you include things because it helps believability.
Toilets and complexity
The Sims is certainly the main example, up to the point of being nicknamed "The Toiled Game" within Maxis. The Sims has toilets, because if the game did not have them, you would notice (you build the house, after all). If it had them but sims did NOT use them, it will not make sense. Thus, The Sims needs toilets.
They also covered what happens if the sim can't go the toilet (accidents that needs cleaning). They added on top that toilets get dirty, and that they could break and flood the place.
I also want to point out that The Sims added sobrenatural elements.
However, I want to talk about a game that uses toilets much better: Prison Architect.
Players would notice if there were not toilets to put in the cells.
Why would players place toilets in the cells? Prisoners need to use the toilet.
If they do not have access to a toilet, not only you will get accidents, that needs cleaning... but also it increases the chances of prisoners to riot.
Furthermore, a toilet needs an active water pipe connection to work, meaning that you need to ensure that there is enough water supply and connected (the pipes can be destroyed).
If they break a toilet, while still connected to water, it will flood the place.
Furthermore, prisoners may hide tools, drugs and weapons in the toilet (which they can be smuggled by other means). Which they will use in their riots, gang fights and attempts to escape. Do not forget to search the prisoner’s cells.
Oh, but prisoners will take advantage of your toilets to dig tunnels to escape. Search for tunnels regularly, put guard patrols and outer walls when you can.
The developers has gone up and beyond asking the questions about their world. Furthermore, they have managed to create interesting game mechanics around the answers to those questions. Moreover, the game – or at least that aspect of the game - is more engaging thanks to that.
As you can see, sometimes you need to add things for believability. And that building complex systems on top of that can be engaging, even with toilets.
It is, in fact, a good idea to turn this kind of problems (test players ask why there are not toilets) into fuel for creativity. If you have to include toilets, what interesting mechanical implication that might have for the game?
For instance, the sims can die. This is a possible out come of neglecting their needs. How can they make death mechanically interesting? They become ghosts, and go on scaring other sims, of course. Realistic? No.
Is it ok to aspire to create something like that? If that is your vision, sure.
However…
Find a minimum viable product
Let me tell you about software engineering. We know that the complexity of the software increases exponentially with the number of requirements. We also know that the number of possible bugs increases with the complexity of the software.
I will refer you to “What We Actually Know About Software Development, and Why We Believe It’s True” by Greg Wilson.
This means that if you can remove requirements, not only you create a software in less time and with less bugs, but also the improvement will be super linear.
This is a good argument to take your design and start removing things. File them for version 2.0 or whatever. Another argument is that you will be able to find the core of your design: the minimum viable product. Ship this first.
Oh, and any bugs you leave alive build up your technical debt, they charge interests (the longer you take to fix a bug, the more expensive, and that is exponential).
I am not telling you that you should not go for a crazy design; I am telling you that you need to start small. Plan how you will add stuff on top, and grow over time.
Back to The Sims, they called it “The Toilet Game” because it required doing unappealing tasks such as going to the toilet and cleaning it. In addition, executives initially rejected the project. However, the developers kept working on it until they created something that executives approved. It required years to archive it, of course, but look the success that The Sims has.
Growth and design, by examples
I have told you to find a minimum viable product, and to make that product grow. You can plan how to extend it ahead of time, which will save you a lot of trouble. However, once you do, you will have effectively coupled those systems to your core design, and that can bring trouble.
Examples:
The Sims. They had many expansions, which means many systems for which they had to plan. That is a form of Accretion (look it up). That kept the evolution of their core mechanics in check. The Sims 4 was a return to rework the core of the game, and that is a good thing to do.
Civilization. Again, it had accretion, and people were used to - and enjoyed - the complexity of all the interlocked systems. However, that complexity meant a high entry point for new players that had to learn a lot to be competent in the game. Again, the developers did a return to the core mechanics, and Civilization V shipped with an overhaul of the core principles of the games (down to the grid and movement), there was some backslash from veteran players, but it served as entry point for many new players who were not following the franchise from the start. Then they went to improve it and add some of the complexity back in Civilization VI.
Windows. Microsoft sells licenses and support to companies. Companies that has business critical software running on Windows. If the developers of that software used deprecated API functions, or relied on undocumented behavior, and as a result did not work in the new version of Windows… who do the company blame? Microsoft, because Windows was the thing that changed. They call support – that they are paying for – and Microsoft has to solve the problem. As a result, Microsoft developed a strong tradition of backward compatibility, Windows core grew in complexity, and the system started to become slower in each iteration. Until you got Windows Vista. Windows 7 was them going back to rework how they organized things. Windows 8 was them going too far, and trying to appleal to the mobile market at the same time.
English. The spelling make little sense. Look at why the letter “e” sounds different “here” than “there”, and countless more examples. The abstract is because of “historical reasons”. As English takes words from different roots, and as those words evolve their pronunciation change. Then either you change the way you write it to match the pronunciation (breaking backwards compatibility with old documents that will no longer make sense to people learning with the new version of the word) or you let the spelling and pronunciation diverge. Furthermore, people has been finding ways to shorthand writing since the time of monk scribes. Nuthin new m8. Sometimes people did try to standardize. We have to thank the printed press for pushing the process. Yet, that is just spelling, "the structure of language grows, as a reef of dead metaphors" -- The Unfolding of Language by Guy Deutscher.
I am telling you that systems evolve, and that accretion happens. Yes, aim small. Yes, create a simple and elegant minimum viable product. However, plan and design the system in such way that it can grow, preferably avoiding strong coupling. Yet, every decision you do, cuts some path you didn't taken. If eventually you need to go back and rework, be ready.