Strategy vs Variety

The constant credo that I want to expound about the games I want to create is "never the same game twice", and that has been a very concerning problem for me.  How do I make a game that gives the player a brand-new, novel experience, every playthrough?

One of my theories as to why players will replay a game is because it allows strategy, but strategy can only occur if you give the game an element of predictability as well.  Strategy relies on the fact that your brain can make guesses about the future, and you'll react to each situation as they arise to solve the problems, much like how you are supposed to walk on the outer ring of a level in Slender: The 8 pages, and then move into the inner ring.  People who have played that game understand that they have to check all of the possible spawn points of the eight pages, and be able to conserve their flashlight's battery and their running stamina.  The problem with that game-making style is that after a while it becomes a robotic repetition of following the same steps you did previously to complete the game, and after a long enough time doesn't give the player enough replayability.

Games that gives you a set of iterated problems with specific solutions (like enemies which may be immune to one form of attack but vulnerable to another) gives the player some variety, but in the end is just a game of rock, paper, scissors that relies on your ability to predict which move your opponent will make.  I'm reminded of Starcraft, Diablo, and Skyrim when I say this, though Skyrim's and Diablo's RPG elements add another aspect of variety in their gameplay: specialization.

Specialization is a fantastic tool, and if enough specialization options are available to the player, gives the player a multitude of options to address each challenge presented.  The problem with this style is that it again is a game of rock, paper, scissors, only this time you are constantly choosing "rock, rock, rock, rock", and sooner or later you will encounter a critter who chooses "paper", making your life interesting.  This reminds me of how I typically play Skyrim, a rogue/sniper/thief.  You take the least amount of damage, and if you are patient enough, you can eventually wear down every enemy into death.  The only problem I encountered is when I had to go head-to-head with a magic-user who could heal themselves: if I brought their health down far enough they would just heal up, so I had to do enough damage to them quickly enough while getting them to spend all of their mana before hand to prevent them from healing, or have a few poisons or weapons specifically reserved for when I encountered those enemies.

Despite the fact that getting hit with "rock" while choosing "scissors" is a bit vexing, it actually doesn't demean or limit gameplay.  It just means that a variety of creatures were presented and you finally encountered one that exploited your limitations.  Given a long enough time scale, you would always eventually encounter a challenge that countered your strength.

Rogue-like games are interesting because they give the player a variety in environments to explore, but also prevents the player from developing a strategy because there isn't enough predictability to develop a strategy.  If another player and I are both playing Terran, and the other player is building tanks, I understand that spamming marines isn't the best bet for me.  If I'm playing a rogue-like game, then I have little to no idea as to what to expect as I progress from level to level.  Being unable to guess my enemy (in this case the game environment) lends a sense of realism to games: life is unpredictable, and so is this game.  You can also add RPG elements to a rogue game to encourage specialization, giving the player a variety in environments and specialization choices, but the problem with that is that you'll create a game that is more about casual, arcade style playing.  Not only that, but rogue games may or may not give you the correct equipment drops to allow you to progress to the final stage of the game (I'm looking at you Faster Than Light).

So how do you create a game that gives the player enough variety to keep the player interested, but enough predictability that the player can develop a strategy?  One possible solution is just to create enough content and then barring the player from taking alternate routes to find the solution.  The problem with this style is that given enough playthroughs, players will eventually come across all of the content.  On top of that, the human brain desires novelty, and barring a player from exploring alternate paths makes the player feel as if they failed to complete a challenge, much like how one will find a waist-high fence that "just so happened" to barricade the player from going down the rest of that one hallway.  So, creating copious amounts of content is not the solution either.

So the question remains: how to create enough variety to keep the novel experience, but enough predictability to allow the player to develop a strategy?

Curious question indeed.