I first learned about role-playing games in secondary school, about the same time I was getting serious about comics. I went through the standard first few stages of the hobby. Dungeons & Dragons (not Advanced) held my attention for way longer than it should have, considering how minor my interest in fantasy was. Sci-Fi came along in the form of Traveller. That was cool, but I was more of a Star Wars kid, so its tone was too dry for me. I went through a West End Games' Star Wars RPG phase, a Ghostbusters RPG (also by West End) phase and probably half a dozen others before I finally found my brand in the form of R. Talsorian's first-edition Cyberpunk game.
By this point it was 1989. The game was set in the far-off reaches of 2013, and was a close match for the kind of science fiction I was into. The rules were clunky in places, but I followed that game through two more editions and dozens of expansion books. Cyberpunk was my second home. I forgave it every structural inconsistency, every rules contradiction, every weird digression into Japanese giant robot stuff. I ran Cyberpunk games all the way through university and well into life out in the world. All the time, although I guess I didn't realise it then, I was learning to tell stories. I was building characters and worlds. I was learning to do my research, I was teaching myself to respond to feedback, course-correct my plotting and strengthen my weak points. For better or worse, my entire approach to writing is grounded in my role-playing years. I still own every game, expansion and sourcebook I ever used, and I still refer to them once in a while. Cyberpunk in particular is still worthy of attention, as much for what it did wrong as what it did right.
The man most directly responsible for the game was Mike Pondsmith and, despite all the first-edition copy-and-paste errors that somehow made their way into every successive edition of the game, he's still kind of a hero of mine. He had a definite vision for approaching storytelling, and expressed it in what may be the smartest words ever said by a Microsoft employee.
Mike Pondsmith says "Don't Play Video Games."
I'll clarify. In the words of Maximum Mike himself, "video games are predictable. You start on one level, beat the henchmen monsters, defeat a few traps and take down the Boss Monster". In his Cyberpunk referee's guide, "Listen Up, You Primitive Screwheads" he talks through a few basic options to avoid the boss monster problem. He suggests skipping it for a few chapters, or throwing the boss fight in at the middle of the chapter rather than its end. Whatever it takes to break the lather/rinse/repeat cycle that virtually every how-to guide on storytelling seems to insist on.
Quick side-note: I've been reading a lot of books on writing technique lately, more for entertainment than guidance. In fact, about the only useful thing I've learned is that success in the storytelling business seems to be entirely dependent on slavish devotion to the ugliest and most simplistic thematic and structural clichés. As for stylistic factors, when one of the world's most respected writers tells you never to use adverbs or the passive voice in your work - but tells you this using BOTH IN THAT VERY SENTENCE - then it's probably time to look elsewhere for advice. Also, he calls it the "passive tense" - which no-one should ever do while within my earshot and dick-punching range.
Back on-target: Mike Pondsmith's boss monsters. Something about his rant struck a chord with me back in 1994, whose echoes I can still hear today. It was always probably a little unfair to the video game industry to accuse them so broadly of formula-farming, but it's hard to find much fault with the actual point he was making. The build-up of knocking down minor obstacles on your way to the big one is so much an accepted storytelling principle that its absence is a lot more powerful than its execution at this point. Seven Psychopaths was the stand-out film of 2012 for me, for pretty much that exact reason. That film was a better examination of storytelling technique, convention and subversion than any how-to I've ever read. Not only does the boss monster climax get thoroughly deconstructed, but the film flat-out tells you it's going to do it.
More recently, I was struck by so many reviewers' reactions to Captain America: Civil War. According to a lot of people, that film blows its boss fight half-way through and leaves us with a limp climax. That reaction just totally fucking baffles me. To me, Zemo has one of the most personal and comprehensively explored motivations in the whole MCU. His actual plan was kinda batshit, obviously, but in getting the big fight out of its system part-way through, Civil War gave itself the opportunity for an endgame that actually meant something. We got a closer, much more personal struggle between the heroes, and a stronger closure between Black Panther and Zemo.
Hell, even video games get it right once in a while. While I have a hard time forgiving Blood Omen: Legacy of Kain for presenting me with my first experience of the irrelevant binary-choice ending that still infects video game storytelling to this day, its sequel Soul Reaver essentially presented a final-act boss fight from the boss monster's point of view. I'd never seen that before, and I'd be hard pressed to pinpoint a game that's done it that well since.
Anyway, I'm approaching rant velocity again. Basically, my point is don't read books on writing technique. Writers can't teach you how to write your stuff. They can only teach you how to write theirs. Worse still, it turns out a lot of them are just parroting a formula they learned from someone else, and using the tired "do as I say, not as I do" thing to justify their own deviations from a blueprint that clearly doesn't even work for them.