009

experiments/009_catalyst_ascent/hypothesis.md

@@ -0,0 +1,72 @@
+# Experiment 009 Hypothesis — Private Until After Play
+
+## Question
+
+Does an enjoyable qualitative build become meaningful when later problems are sensitive to the capabilities it creates, or does added resistance merely turn a pleasant toy into compulsory throughput work?
+
+## Why This Experiment
+
+Experiment 008 separated two judgments that earlier tests often blurred: the player enjoyed building, but the build felt meaningless. Longer mature fields made the effects and modest movement refinements easier to see, yet the player believed nearly any selection—or no selection—could clear every population. The four-choice cap therefore excluded options on paper without producing a felt tradeoff.
+
+The likely issue is not simply low difficulty. More enemy health could make upgrades mandatory while leaving every build interchangeable. Catalyst Ascent instead retains the exact six-component causal vocabulary and introduces qualitative enemy capabilities over one continuous run:
+
+- **Wards** absorb eight hits inside a short opening window, reset partial progress, and later re-form. Focus ruptures and Conduit lances can breach them directly, while sufficiently dense products can strip them.
+- **Renewals** continuously regenerate, rewarding concentrated damage, chained effects, or amplified secondary output rather than intermittent baseline fire.
+- **Broods** keep creating small bodies, potentially feeding kill-triggered Bloom while increasing pressure if ignored.
+
+The player chooses after fields one through seven, then carries the complete build through fields eight through ten. This preserves the positive authorship and visible cascade from 007/008 while allowing the environment to outgrow baseline capabilities. Failures restore the current field and build so execution mistakes do not erase the construction history.
+
+## Competing Interpretations
+
+1. Capability-sensitive enemies make construction both enjoyable and instrumentally meaningful; the player anticipates what the build can handle, notices a limitation, and adapts through one of several causal routes.
+2. Escalation compels continued play but creates no additional curiosity, repeating Experiment 002's distinction between persistence and reasoning.
+3. Ward, Renewal, and Brood labels prescribe obvious counters, reducing choices to loadout replacement as in Experiment 003.
+4. The exact qualitative effects and visible power trajectory sustain interest, but enemy capabilities add little; spectacle and composition remain sufficient without consequential exclusion.
+5. One universally dense effect network defeats every capability, so the experiment again measures throughput rather than meaningful construction.
+6. The environment finally makes a choice consequential, but a mistaken build creates a soft lock or feels punitive rather than recoverable.
+7. The player enjoys selecting and watching a build but still does not care whether it succeeds; capability relevance alone does not supply valued stakes.
+
+## Evidence Priorities
+
+Strong evidence:
+
+- anticipating a later capability while choosing, then confirming or revising that prediction;
+- noticing that the current build changes which target, firing pattern, or movement strategy is viable;
+- identifying two causally different routes through the same enemy capability;
+- wanting to replay specifically to test a different route, rather than merely to finish or maximize a number;
+- a limitation generating a new build question without making the current run feel wasted;
+- the complete build continuing to reveal consequences in fields eight through ten.
+
+Ambiguous evidence:
+
+- completing all ten fields;
+- dying, clearing faster, or selecting seven upgrades;
+- using the textual Ward/Focus or Ward/Conduit relationship;
+- long play caused by durable enemies;
+- large effect counts or spectacular screen clearing without a reported or behavioral question.
+
+Failure evidence:
+
+- one obvious package answers every field;
+- enemies merely take longer to kill;
+- a capability is unreadable or is confused with ordinary health;
+- the player feels forced to take a named counter with no meaningful alternative;
+- a mistaken build makes progress impossible and recovery requires restarting the whole run;
+- building remains enjoyable but still feels unrelated to anything worth accomplishing.
+
+## Analysis Guardrails
+
+- Treat the player's “meaningless” report as evidence, not a final diagnosis. Test whether consequence sensitivity is the missing cause rather than assuming difficulty is the answer.
+- Do not infer meaning from necessity alone. A compulsory upgrade can be strategically empty.
+- Do not use clear rate, deaths, duration, or upgrade count as stand-alone proxies for fun.
+- Separate anticipation, observed causal attribution, revised understanding, and post-hoc explanation in both telemetry and questioning.
+- A selected counter is only meaningful evidence if the player perceived alternatives or learned why it worked.
+- Preserve the possibility that building is intrinsically enjoyable while the shooter objective remains emotionally irrelevant.
+
+## Feedback Questions
+
+Ask only after reviewing telemetry:
+
+1. When, if ever, did the build start to feel like it enabled something baseline fire could not? What did they believe caused that difference?
+2. Did any enemy capability change a choice or tactic, and did it feel like a problem with multiple possible answers or a prescribed counter?
+3. Did a limitation create another build they wanted to try? At what point were they ready to stop?