A game is nothing more that a series of patterns. To keep the player's interest, around the time that the player masters a pattern, a new pattern is introduced.
Most games are nested play loops. For example, in Halo you learn how to move, then you learn how to explore, then how to fight, then how to manage your fighting resources, then how to work tactically.
Most games spike to the largest loop as quickly as possible. In Halo, they explicitly explain how to move and explore, because these aren't considered to be "critical play loops". This is largely because most of the people who will play Halo mastered these basic play loops long ago, on older games. These people have also largely mastered the fighting and tactical parts, too, but Halo introduces enough variation from the norm that the pattern of these play loops takes some time to master.
Thent he game takes the largest play loops and alters them. In Halo, they put you in new levels (well, somewhat), face new enemies with a new level of combat resources. This alters the pattern of the play loop enough that the player takes some amount of time to re-master this new variety.
However, any given play loop has only so much "area". For example, there's only so many ways you can face the same enemies with the same gun before you get thoroughly sick of it. Every player has different thresholds, and that threshold can be raised or lowered based on how invested in the game the player is, but there is no doubting that a play loop only has so many variations before a player gets sick of it.
Some play loops are larger than others, and it is relatively easy to extend play loops. For example, in Halo, they expand the main play loops by giving you a host of different enemies, a host of different guns, and a host of different methods of encounter, and a play space which changes with every enemy killed and every bullet fired. The area of the play loop is essentially determined by the number of ways these can be recombined in any given play through the loop.
Of course, the area of a play loop may or may not be accessable to the player at any given time. Although it is definitely possible to blow away enemies with a rocket launcher, that is not an option you have in the first level, because there are no rocket launchers. This is a standard approach: the area of the play loop is "bumpy", with high and low spots. The players naturally gravitate towards the highest highs (or the lowest lows, however you want to look at it). This means that, given the option, they'll never climb the "pistol" peak if always allowed the "rocket launcher" peak. So, even though the "pistol" areas in the play loop exist, they functionally do not, because few will enjoy being there.
The answer is, as I mentioned, restricting access. If you can't get a rocket launcher, you can't climb the "rocket launcher" peaks. Therefore, you climb the highest peaks available. Perhaps that is a pistol. Perhaps that is something else.
Restriction happens in all dimensions. When the game tosses specific enemies at you, it is forcing you to look a specific subset of its whole area, and you have to figure out how to climb the local mountains. It could allow you to navigate this dimension, but most players would feel extremely lost, since all they can see are slopes rising all around them. By confining the possibility space of any given play, they allow a player to see challenges without feeling lost.
What does this have to do with Outpost Kaloki? For that matter, what does it have to do with Frequency, Katamari Damacy, and Tetris?
As I mentioned at the beginning, most games have "critical" play loops and "noncritical" play loops. They give a quick tutorial for the noncritical loops to move players quickly into the critical loops. Really, that's what a genre is: it's a pack of pre-mastered play loops.
But the games I just mentioned don't do that. Most of the games the real gamers jabber about - the expert's experts - they have no extraneous play loops. They are clean, lean, playin' machines. Marble Madness. Dance Dance Revolution. That really cool racing game whose name I can never remember. NetHack.
These games have two or three play loops, all of which closely follow Pattern Adaptation Control methodology. Instead of ramping up to a new play loop when the first is mastered, they shunt back down to a newly altered original play loop.
In Katamari Damacy, there are two play loops: maneuvering and collecting. They are tightly bound together: when you collect something, it changes your movement, and you have to move to collect. In Tetris, there is only one play loop: dropping blocks. But it feeds back into itself.
Outpost Kaloki shows the strength of this kind of design, while also highlighting the differences. Unlike the other games mentioned, Kaloki is a "normal" game. That is, it depends on half a dozen play loops and balancing half a dozen resources. It's core play loop is not one which feeds back into itself in any meaningful way.
But Kaloki has carefully structured missions which restrict play space to a tiny fraction of the full play space. It "stripes" its way across its own play space, and with each mission there is a very different play loop. So even though it doesn't have the tight feedback loops of the memorable games of the industry, it SIMULATES these tight feedback loops.
This works great... for a while. But if you were to map Kaloki's play space out, you would see it has a very distinct and specific shape. This means that Kaloki only has so many "interesting permutations" it can show the player before, quite simply, it runs out of new space.
Therefore, they take very small slices. The new play loop is usually almost identical to the last play loop, just a few more things to manage. They make up for this with charming dialogue and fun characters, but without those characters, Kaloki would be a much weaker game.
Can you see the difference between Kaloki and Katamari?
Kaloki and Katamari both have a tilted play space: "bigger is better" is always true.
But Kaloki has a STATIC play space. All its hills and bumps are preprogrammed and change very, very little. Kaloki changes your experience by giving you slices of that static terrain, guiding you across it in new ways.
Katamari has a CHANGING play space. In addition to being wholly defined by the layout of any given level, the relative terrain of the play space changes every time you hit a bump. Therefore, Katamari changes your experience by letting you try ANY slice of that dynamic terrain.
That's the key to the prolonged popularity of all those games I mentioned: the terrain of the play space CHANGES DURING PLAY. It's a REACTIVE play space.
Kaloki really isn't. That's not a bad thing, but it means Kaloki can only hold someone's interest for a few hours before it simply runs out of new patterns for the terrain.
1) In a memorable game, all play loops are critical play loops.
2) In a memorable game, the play space changes during play.