My approach to making games is like a funnel, or a knock-out competition. Only the strongest concepts move on to the next round.
I keep a text file of ideas for new games. These are just notes to myself. I have plenty more ideas floating around in my head.
Some of my ideas are amazing. A very small number of these ideas become prototypes.
Only when I'm bursting with enthusiasm for an idea I have, do I start working on it. That enthusiasm will rapidly drop off, so it needs to start out extremely high.
I make the important decisions about how the game will work. This will likely take a number of hours of brainstorming, and scribbling down notes and ideas. I need to get this right, or the game is doomed.
My design process is a kind of a condensation. Before any actual creation or work begins, I start with a cloud of ideas, regarding what the systems of the game are. These ideas are very basic, and systems can easily be added, replaced, or removed at the stage.
From this point, I start to answer some of the bigger questions about the game. What is the objective? What are the game objects and scores/resources?
The cloud of ideas begins to condense down. I don't want to make strong arbitrary choices, so I just condense the most obvious bits. It's like a Sudoku puzzle. I want to make sure I'm getting it right, so I just "write in" the stuff I'm most sure of. I don't want to make a mistake here. As systems become stronger and clearer, other systems can safely form strong links to them (or be removed from the game.) This makes those systems themselves stronger, and the game quickly crystallises.
I create the first prototype.
Despite it being the first prototype, it is also the final form of the game. It will be clearly recognisable to someone who plays the finished product. Systems will change their internal working, but they will perform the same function, and, most importantly, be in the same arrangement. It becomes almost impossible at this point to add or change fundamental systems of the game. It has already condensed.
I then do a playtest.
At this point, the game either succeeds, or it doesn't. I create the first playable prototype, and play it. The game will either be interesting and fun, but flawed, or it will just "not work". If there's fun and interest, the flaws can be fixed. If the game doesn't work, or isn't fun, it's doomed. Removing a system at this stage causes me to have to re-evaluate all other systems, and the whole game turns back into a nebulous cloud of possibilities again. I'm really just designing a new game. This is the point at which most of my games died. Some of my earlier games failed to consider player interaction at the start. When I went to add it later, it simply didn't fit the game. I struggled with the game, but eventually it went to my "game graveyard".
For the small number of game experiments that shine on their first playtest, the next step is simply a refinement of the game, not a redesign. The whole core is sound, and a game can be built upon it.
Even then, many of my games reveal deeper problems, within their first ten or so revisions. These are problems that I try to fix, but can't. These are not things that make the game bad, or stop it from working — they just stop the game from being amazing, or reveal themselves to be too complex to solve. I abandon most games during this period.
It's a terrible mistake to work on something that doesn't have a great spark. It might be a great idea, but I have many other great ideas.
A great idea often does not equal a great game.
My best games were fun on the first playtest. They needed a bit of work, but the final game was clearly the descendant of that first prototype.
For many others, I couldn't let go of the idea, and I altered part of the game core. This is effectively creating a new game, and I've never had this be a success. Even worse, some of my earlier games staggered on with a fundamental deficiency in their core (such as there being no interaction between the players.) These games all eventually died.
The biggest mistakes I've made in game design were the games I put dozens of revisions into, and then abandoned.
If you're putting 100-200 revisions into each game, killing games after a handful of revisions is the sane thing to do.
Even when I've completed games, only a minority have been published. Sometimes, I simply made an error in the development of the game, and couldn't see it. (This was much more common in my earlier days.) Other games I simply couldn't find a publisher for.
Other games were completed early in my design career, and just aren't good enough.
As of 2024, I've created (in whole or in part) 35 games. Five were completed and proposed to publishers, and two of those were signed for publication.
If you don't have this kind of ruthless, industrial approach to game design, you'll never beat the people who do.