I try not to argue exact language much. I try to keep my approach to language as a descriptivist. Language evolves in terrible and funny ways. Metroidvania is an awful term, but it's a term with history I will use long before I use the term Search Action(a genre name that has the same appeal as calling hotdogs Intestined Scraps). Sure, some phrases like Backtracking are too misguided for me to tolerate, but Quality-of-Life isn't like that. So called QoL changes can mean a lot of of things to a lot of people, but I'm going to talk about a way I see it used a lot. Maybe the most common way(The usages I get exposed to are surely extremely biased).

We often treat Quality-of-Life changes like freebies. Changes that are only an improvement. Changes that don't actually change the actual game (the definition of which will vary wildly depending on who you ask). Changes that just make things nicer, more enjoyable, smoother. Changes that simply add up over time to make something a better game. Less menuing, automating repetitive tasks, smoothing out randomness... generally just adding consistency where the controls, UI, or meta-systems are fighting you. They just make things better, right?

There are no free lunches

Here is the point where I am tempted to say something catchy like "there is no such thing as a Quality-of-Life change". Unfortunately, that's a bit too reductive. Instead I will say: if you're a designer, you should pretend that there is no such thing as a Quality-of-Life change. Because for every change that might actually be a strict improvement, 10 changes happen with consequences that went unconsidered. That doesn't make these changes even necessarily a net negative, but it's important we at least know what is happening. If a change seems so obviously good, we should at least be a little suspicious of it!

I feel like the consequences of supposed "QoL" changes are most obvious in competitive games. When input buffers got introduced in Fighting Games, people said they would just let players get to the real game faster, and that is true..ish(There is no such thing s as the 'real game).

... But they also changed the game, changed what was rewarded. The modern way we talk about frame data still doesn't apply to older games, even when we bring younger players and modern perspectives to them(They DO change from these two things, but in different ways). Being assured of a first-possible-frame input for any normal fundamentally changes the risk reward of pressure. Removing things like variable wake-up timing strengthens okizeme and setplay. It increases the pressure to be optimal because consistency is realistically achievable.

Starcraft had a similar arc going into Starcraft 2. Larger control groups, elements of auto macro, and better path finding... hell, even changing how many workers you start with all seem like obvious positive "Quality of Life" changes... but then it quickly became obvious that the game was just fundamentally different. Early on, the so called "Ball of Death" became the end game strategy, easily moving large sections of an army in one huge control group. While the game had plenty of room for skill and strategy, tactics took a deep hit, and years of patching later the game is still found lacking in those departments compared to Brood War.

The changes made to SC2 or modern fighting games weren't bad ones. In many ways they were necessary ones for the future of both genres... but neither really fully came to grips with what those "Quality of Life" features were really doing. They tried to adjust, but never nearly enough. SC2 tries to add little bits of micro to tons of units, but none of that strains people to the point where shit can just go wrong. SC:BW professional games are filled with mistakes at every moment. Commanding an army is hard and SC1 captures that struggle on all levels of play.

Some modern fighting games try to address these issues, but can't compare to old fighting games, which are filled with drops on simple BnBs, weird scrambles, and a degree of instability that has to be strategically planned around. Extremely new games like SF6 try to add new elements, or Just Frame(... 'Just frame', but its like a 3-5 frame window) bonus tricks to try and help out. They've tried to move a lot of skill into situational resource awareness too. These are neat ideas, but ones that came almost a decade late to the genre. Nothing has even come close to mixing up the rigid situations that arise from modern frame data. Things often feel less like a fight and more like strategic stock investments.

Single Player games seem like they'd be a free pass but even small changes can encourage unfortunate player behavior. A classic is how old school PC saving worked. Quick Saves and Loads, with the absence of any other save system, basically subtly encourages save scumming, tedious encounter optimizations and just... playing really greedily. You could resist this, but the line between this and "using the autosaves between levels" was so strong that a lot of people would give in. It took the modern, perpetual save systems to eventually course correct and let people relax a little. You want just enough control and convenience that the player becomes okay with failure.

Often when modernizing old games there is a temptation to remove certain old design models... Item pickup weight limits, or a lack of automap or whatever. Yet the tension of playing Demon's Souls and not having enough item load to carry an item out was, good or bad, an experience. In games without any kind of limit, people just compulsively pick up everything and horde. Is this the behavior you want? Maybe. I don't know if I wish I had pickup limits in Elden Ring, but I sure thought about my inventory a whole lot less.

Things go even farther to systems like auto loot processing where the question begins to become why? Why are we optimizing mechanics to the point they're basically gone from the game(I hear this is what the history of Monster Hunter is like, but having not played any I don't feel comfortable ACTUALLY writing about it)? Things stop being a matter of QoL changes and more a question of genuine design goals. What's the point of intentionally introducing complications only to eventually sand over them with QoL changes? We get modern metroidvanias with color coded blocks, and color coded attacks, with color coded map icons because of "QoL"... but then all we do is reduce games to a check list. Have we just replaced the fantasy of exploration with busywork?

Our job is to make problems for the player to solve, not make problems and then immediately solve them for ourselves. If we have desire to do that, then the initial problem probably shouldn't exist at all... Probably.

Even button-press reducing changes can introduce undesirable behavior. Do you want the player to be swapping their class or equipment around all the time, for every encounter, now that you've made it easier to do so? Maybe! But that isn't Quality of Life, that's a genuine design question. If you don't want them switching around so wildly, what's a more pleasant way than crummy UI to stop them? Does including a quick restart in your roguelike encourage restart-scumming? What are they looking for? Can you minimize whatever variance to discourage that behavior?

So I'm not going to sit here and say "Don't say Quality of life!" like I will with "backtracking"(No, for real, it's a fucking useless, meaningless term 99% of the time)... but instead look at QoL problems like any other. What is your change encouraging, and what is it discouraging? Were there any positives, even if unintended, from the old behavior? How could that be preserved? If it can't be, is the sacrifice worth it? Or, to a bigger extreme, if it might not be worth it, has your change gone far enough? Maybe you don't need to preserve what you had, maybe you need to obliterate it with a new, powerful idea. Not a QoL change. Real change.