Gary & Dave Tinkered Their Way to Success.

Having the privilege of knowing people who were there when Dungeons & Dragons has changed how I see the early editions of the game, as well as not-D&D RPGs of the same era, and greatly informed my current opinion on game design as a skillset. (And no, the FleshMech crowd at The Gaming Den aren't where I'm going.)

It is clear now that Michael Monard's summary of "We made up some shit we thought was fun." was the way things worked. They iterated their way to what D&D would become. You can do that for your game, especially the early iterations; make something, play it, see what works and fix what is not working. Rely on spot-rulings over written rules, and provide only enough mechanics to cover what can't be spot-ruled reliably.

We see this in the game's introduction of what is now the Core Four classes. Fighter was first, as befits a wargame derivative. We now see that Magic-User was second and Cleric was third. Thief came with the first supplement, and wasn't in the original set of Original D&D. When you consider the tinker-style iteration going on, this makes sense.

Fighters wear any armor and use any weapon or shield. Clerics trade some of that for the Val Helsing suite (Turn Undead, Healing & Blessing spells, etc.). Magic-Users trade almost all of it for Real Ultimate Power (if you live long enough). Thieves trade some of it to become Stealth Gods (eventually). Defining other classes in terms of how they differ from the Fighter (which is our baseline) works just fine, especially in hobbyist tinker-mode game design. So it will be in other mechanic designs; pick a baseline and make other choices deviations from that.

Specific things follow a similar pattern. Originally, all Hit Dice were 1d6 and all weapons did 1d6; this was itself an iterative shift from the wargame where a model had one Hit and did one Hit, and Heroes had and did more. Spells are similar in their development, with the Fireball from the previous post being a notable example.

Don't worry about fucking it up. Worry about finding the breaks, figuring out how it broke, and fix that. Later, rinse, repeat until you get what you're after. Make it work, then make it work as intended, and only after it's reliably working as intended do you bother making it pretty.