feat(2048): add 'why is this so addictive' notes page

Author: cablehead

examples/2048/notes/content/what-is-2048.md

@@ -77,10 +77,7 @@ head and rich enough to surprise you. Each merge is a tiny dopamine
 hit. The board is small enough that every move feels consequential.
 And it's free -- no install, no account, no ads.
 
-Next: [the ones worth playing](./best-of-2048) -- the ancestor, the
-canonical version, and a variant worth your time.
-
-# Best of 2048
+## The ones worth playing
 
 **Threes!** -- the careful 14-month original. Pay for it. It's the
 version made by people who hoped you'd put it down sometimes.
@@ -102,8 +99,7 @@ free, in any browser tab.
 parody variant where merges resolve by absurdist British game-show
 logic instead of arithmetic. Worth a few rounds for the joke alone.
 
-Next: [how this is built](./in-nushell) -- the same game on top of
-event-sourcing, SSE, and view transitions.
+Next: [why is this so addictive?](./why-is-this-so-addictive)
 
 # In Nushell
 
@@ -147,3 +143,5 @@ The source: [examples/2048 on
 GitHub](https://github.com/cablehead/http-nu/tree/main/examples/2048).
 Browse the [/design](../design/) component viewer for the live mirror
 of every server-rendered piece.
+
+Next: [back to the rules](./the-rules), and round again.

examples/2048/notes/content/why-is-this-so-addictive.md

@@ -0,0 +1,62 @@
+# Why is this so addictive!?
+
+I ported 2048 to Nushell over a weekend, for fun. Then I noticed I
+kept playing it. Not testing it. Playing it. I'd close the tab to get
+back to work and find it open again before I'd decided to. That's when
+I stopped and asked: what is going on here? Has anyone actually written
+about the pull? Turns out a _lot_ of people have.
+
+The most-quoted answer is from [Judy Willis](https://radteach.com/), a
+UC Santa Barbara neurologist the press talked to in 2014. She named
+[two dopamine boosts](https://web.archive.org/web/20150311024653/https://www.popularmechanics.com/culture/gaming/a10341/why-the-2048-game-is-so-addictive-16659899/).
+The first is prediction: every move is a guess at what the board does
+next, and brains (in her phrase, "like bookies and psychics") love
+making predictions, so the dopamine arrives just for guessing. The
+second is that the game stays challenging but achievable, which keeps
+that reward flowing instead of tipping into frustration.
+
+But the part that actually got me is something her account skips: the
+[fake sense of progress](https://en.wikipedia.org/wiki/Incremental_game). Every new high tile lands
+like a level-up. Your first 256, then 512, then 1024, each one a new
+personal best, new territory, a little "I'm finally getting somewhere".
+But nothing has actually changed: same 4x4 grid, same four swipes, same
+game you were playing back at tile 16, just a bigger number. And
+because every step is a doubling, the jump to 512 feels every bit as
+exciting as the jump to 256 did. The climb doesn't lose steam. The
+progress is just a number going up, one of the stickiest things a game
+can dangle.
+(An entire genre, the [idle and clicker
+games](https://www.vice.com/en/article/cookie-clicker-wasnt-meant-to-be-fun-why-is-it-so-popular-8-years-later/),
+is built on nothing else.
+That's its own rabbit hole, and its own page, later.)
+
+The rest pile on top:
+
+- **Variable-ratio reward.** Each move's payoff turns on a random spawn
+  (nine times in ten a 2, otherwise a 4) in a square you didn't pick.
+  Rewards on an unpredictable schedule are the [hardest to walk away
+  from](https://en.wikipedia.org/wiki/Reinforcement); it's what slot
+  machines run on.
+- **Free restart.** No login, no install, no score to keep, no "are you
+  sure?". "One more" costs nothing, so you take it.
+- **Near-miss.** Only [about 347,000 of the first 42 million
+  games](https://www.buzzfeednews.com/article/hillaryreinsberg/why-this-free-puzzle-game-is-the-most-addictive-thing-on-the)
+  reached 2048, under one percent, so nearly every game ends just
+  short. And just short
+  is the addictive part: gambling studies find a [near miss makes people
+  keep playing](https://pmc.ncbi.nlm.nih.gov/articles/PMC2658737/) more
+  than a clean loss does, even though it's still a loss.
+- **The unfinished board.** The [Zeigarnik effect](https://en.wikipedia.org/wiki/Zeigarnik_effect):
+  we remember interrupted tasks about twice as well as finished ones.
+  2048 almost never gives you a finished one. Lose, and the game ends
+  mid-climb, before you got where you were headed. And even when you
+  win, you're not finished: 2048 isn't the end, you keep going for 4096,
+  8192, up to a [theoretical maximum of
+  131072](http://www.science4all.org/article/2048-game/). So it ends
+  unfinished either way.
+
+If you'd rather play the version its makers actually hoped you'd be
+able to put down, [Threes! and the others are over in the
+backstory](./backstory).
+
+Next: [how this is built](./in-nushell).

examples/2048/serve.nu

@@ -426,6 +426,8 @@ let design = source design/serve.nu
                     (SPAN {class: "callout-desc"} "the basic rules"))
                 (LI (A {href: ($req | href "/notes/backstory")} "2048 is a broken game")
                     (SPAN {class: "callout-desc"} "how a clone of a clone ate Threes!"))
+                (LI (A {href: ($req | href "/notes/why-is-this-so-addictive")} "why is this so addictive?!")
+                    (SPAN {class: "callout-desc"} "the hooks, and the rabbit hole"))
                 (LI (A {href: ($req | href "/notes/in-nushell")} "in Nushell?")
                     (SPAN {class: "callout-desc"} "how this is built"))))
             (DIV {class: "preview"}