Monkey Flip Millionaire

Play goes through cycles of
 - Consider options
 - Choose option
 - Implement choice
 - See the results
There should be multiple decide-act-observe cycles moving at the same time.  The "act" might be one thing that has differing effects along multiple cycles, and the cycles should probably operate at different paces.

So, like, in Super Mario Bros number one, the player is almost always Implementing and Seeing ('jump this pit' / 'did I manage to get across it?') but there's a bunch of Consider/Choose cycles perpetually (and occassionally) cropping up.  The timer - are you going to knock This Block Right Here to check for a sky-vine or hidden 1up, knowing that burns your limited allotment of time?  The invincibility star is heading for that pit - will you rush in and risk losing control, falling into the pit?  The hammer bros are jumping around on those two platforms / the ground, are you going to bypass them or bop the blocks they're standing on to take them out?  

In Bard's Tale you've got to think through what classes you want to build your party out of, and are perpetually provided with the option to go back to the starting position and swap someone out.  The synergies are sufficiently complicated that there's a solid sense of obviously wrong choices and potentially correct ones, but seeing if they're correct takes actual play and, potentially, corrective action later.  Once you go into Skara Brae you've got the Day/Night timer ticking, distance from the healing Temples to consider, whether to burn your magic user's MP to win the fight sooner or take a few extra blows with your front-line fighters, whether to spend all of that GP to resurrect a level three character or to start a new one, whether to finish the dungeon or go back and level up early / get new spells, etc etc.  These consider/choose/implement/see loops are almost all tied to the PC's position on the maps, specifically their location relative to a small cluster of buildings in central / central-east Skara Brae, so the "Choose/Implement" part are usually monolinear but the various variables considered and the results are on different cycles.

Prey (2017) is brilliant with this - it primes the player to be thinking about the Trolley Problem, then reskins it over and over. Do you help the two people stuck in an escape pod get to Earth - there's a very, very small chance there's a doomsday-producing threat in there with them, but the alternative is to kill them / let them suffocate over hours.  Consider, Choose, Implement, See.  There are lots of contrasting missions available simultaneously - which is a priority?  Consider, choose, implement, see.  A recycling grenade is a great weapon, but could be used to make a ton of recyclable materials for printing ammo for other weapons (or the level-up currency!), how do you use it? C/C/I/S.  When you have enough level-up currency, do you do one of the normalish humanist options or one of the stronger/maybe more fun but very much suspect alien options, knowing it might cause issues later?  You have no more ammo for your pistol - throw it in the recycler for a few more shotgun rounds?

In all these examples, there are choices to be made constantly that affect payoff cycles.  Ideally, a given choice should affect multiple payoff cycles.  The obvious clicker example is "level up Clicks Per Second or Points Per Click?".  Some games give more points per Unit X once a certain number of Units are bought, but because of cost scaling the individually-weaker-X becomes more expensive than the stronger Unit Y, so the player has to choose maximum Right-This-Second currency acquisition or play a slightly longer game in the hope that saving up forthe group-enhancing option will have a better payoff.  C/C/I/S.  

A big threat, one that I think your game fell prey to, is a particular CHOICE being the obviously correct one. This can be overcome by introducing an element of player skill (time that pit jump in Super Mario Bros) but can also be handled via complexity (figuring out whether to level the weaker X or grab more Ys is a multi-variable equation people won't fully crank out in their heads) or by having known unknowns that make the correct choice "fuzzy" (will I find more pistol ammo before I find another pistol, and will I need those extra shotgun shells anyway?)

To sum up - the player is always doing SOMETHING, or else we're failing as game designers.  The SOMETHING should (usually? always?) be considering their options or implementing their choices.  Complexity in the C/C/I/S cycles that descend into chaos makes a game seem arbitrary, and insufficient complexity makes it seem like engaging with the game is pointless.  There's a sweet spot where payoff / punishment loops are fed by the same implementation (and therefore tradeoffs exist), where the player is acting on partial information (joy of discovery), and where there's an element of player skill involved.