Many developers underestimate how trivial it is to code a sandbox environment for any engine whatsoever - letting the player decide the way they want to play the game.
Of course, it requires a lot of testing, but an actual developer attempts to test every single possibility they can think of and have people test what they won't think of, which is usually a lot regardless of how long they've been at it. People can be extremely creative on how they play/break (either accidentally or deliberately) a game and people can also have unexpected fun (or things that really make the game annoying).
Testing and putting yourself in your audience's shoes while forgetting you made it helps you find if your game will or won't click. If you didn't enjoy it because of something, you just have to ask yourself if it's an user problem or a concept problem.
Just a few example for RPGs since I think that was the topic.
Too much dialogue? Add skips and auto-forward, use logs and notes for important info (or just if someone wants to re-read things later, once they're done with the main goal, in the case of a RPG, for example - flavor dialogue is good but it'll slow things down, have the NPCs get to the point and be funny but not lame, extra stuff for talking to them is the extra detail but it must not be a mandatory read). Too much combat? Reduce encounter frequency, change to combat on contact (this is a preferred way as the player can either go berserk and run into everything or casually go past).
Boss battles too hard? Add some quick and witty ways to win the battle for very underleveled people, and for extra challenge, randomize them so people will have even more fun coming up with their own ways instead of following a walkthrough. Make it relatively obvious but not too obvious, it's good for it to be fun or unexpected if you can make it that way.
Item drops are a pain? Treasure chests. Unexpected findings. Hidden locations. Secret areas. Give people who don't want to grind a chance to find something if they put the effort but have terrible luck. Alternately, you can guarantee drops just like a pity gacha does if your audience is intended to be grinders, or if there's a funny advantage to grinding (hidden dialogue, unexpected encounters - you name it, it's your game, be creative).
Mini-games too hard? Put yourself in the shoes of someone who absolutely can't solve a puzzle for their lives and try to solve them. Fix what appears to be way too hard, don't make it too easy, make it intuitive but challenging and fun to solve. Mini-games should be optional and add a fun element of sorts. If they're mandatory, you have to ensure everyone can solve them in a timely manner without having to rely on a guide, but it's a good thing if they have to think. If it's too easy, it'll be boring. If it's too long, it'll ruin the game.
I could go on and rant about every other aspect, but you get the point. Have common sense, play and test as the end user. Forget you made that game. Behave as candidly as possible, and once it's stable and comfortable to test, have other people try it too because you obviously won't be able to get a feel of how everyone else will see it. Have them tell you everything about their experience, what they ran into, what they hated, what could be done better. You don't have to make the best game ever - you won't. Make a game you would enjoy yourself. Show people you want the game to be something worth playing and you want to work hard on it. And if others end up liking it too, then that's just a bonus.
(RMMV/RMMZ games are almost all open and moddable, the engine runs in JavaScript while being fed JSON and it's pretty intuitive - even though it has some massive flaws - which makes it accessible to almost everyone including people who haven't touched code. It's actually quite easy to revamp a game that sucks because of a few oversights, it all depends on whether the entire game was made in a lazy way or it's just a few things that completely ruin the experience for a good bunch of people.)
For HTML games, well, that's even easier. Documentation is all over the place, you just have to look a bit.
...Now if I could apply my logic to my own posts, that would be wonderful, wouldn't it?