diff options
| author | YurenHao0426 <Blackhao0426@gmail.com> | 2026-04-07 22:38:57 -0500 |
|---|---|---|
| committer | YurenHao0426 <Blackhao0426@gmail.com> | 2026-04-07 22:38:57 -0500 |
| commit | 44614df2f4382e567b986bc6dbe5b3091072461e (patch) | |
| tree | 21e2763e93a09d223c03f0cc43aa33481b4f35d1 /results | |
| parent | 0c245f5683cceba448d20d9dfc2090adb3503f14 (diff) | |
Add protocol decision-utility ablation table
Builds on the 5-method audit JSON. For each method, evaluates 7 reporting
strategies (S0=acc only, S1=+Γ field standard, S2-S5=+single diagnostic,
S_full=full protocol), and emits the verdict each strategy would have
reached.
Result: 3 of 5 methods (DFA/SB/CB) are walked back by S_full but NOT by S1.
Each of (a)scale, (b)floor, (d)frozen is independently sufficient for
binary detection of those 3 failures. Diagnostic (c)stability adds
sub-mode discrimination (drift vs noise) but not new positive detections.
This is the §3 protocol decision-utility evidence.
Diffstat (limited to 'results')
| -rw-r--r-- | results/protocol_audit/ablation_decision_utility.json | 69 |
1 files changed, 69 insertions, 0 deletions
diff --git a/results/protocol_audit/ablation_decision_utility.json b/results/protocol_audit/ablation_decision_utility.json new file mode 100644 index 0000000..06f4b66 --- /dev/null +++ b/results/protocol_audit/ablation_decision_utility.json @@ -0,0 +1,69 @@ +{ + "table": { + "bp": { + "S0": "no walk-back (acc only)", + "S1": "no walk-back (looks fine)", + "S2": "no walk-back", + "S3": "no walk-back", + "S4": "no walk-back", + "S5": "no walk-back", + "S_full": "no walk-back" + }, + "dfa": { + "S0": "no walk-back (acc only)", + "S1": "no walk-back (looks fine)", + "S2": "WALK-BACK: (a)scale", + "S3": "WALK-BACK: (b)floor", + "S4": "no walk-back", + "S5": "WALK-BACK: (d)passive", + "S_full": "WALK-BACK: (a)scale + (b)floor + (d)passive" + }, + "state_bridge": { + "S0": "no walk-back (acc only)", + "S1": "no walk-back (looks fine)", + "S2": "WALK-BACK: (a)scale", + "S3": "WALK-BACK: (b)floor", + "S4": "WALK-BACK: (c)drift", + "S5": "WALK-BACK: (d)passive", + "S_full": "WALK-BACK: (a)scale + (b)floor + (c)drift + (d)passive" + }, + "credit_bridge": { + "S0": "no walk-back (acc only)", + "S1": "no walk-back (looks fine)", + "S2": "WALK-BACK: (a)scale", + "S3": "WALK-BACK: (b)floor", + "S4": "WALK-BACK: (c)drift", + "S5": "WALK-BACK: (d)passive", + "S_full": "WALK-BACK: (a)scale + (b)floor + (c)drift + (d)passive" + }, + "ep": { + "S0": "no walk-back (acc only)", + "S1": "no walk-back (looks fine)", + "S2": "no walk-back", + "S3": "no walk-back", + "S4": "no walk-back", + "S5": "no walk-back", + "S_full": "no walk-back" + } + }, + "strategies": { + "S0": "headline acc only", + "S1": "+ \u0393 (field standard)", + "S2": "+ diagnostic (a) \u2016h_l\u2016", + "S3": "+ diagnostic (b) \u2016g_l\u2016", + "S4": "+ diagnostic (c) stability", + "S5": "+ diagnostic (d) frozen baseline", + "S_full": "full protocol" + }, + "summary": { + "missed_by_S1": [ + "dfa", + "state_bridge", + "credit_bridge" + ], + "trustworthy_by_S_full": [ + "bp", + "ep" + ] + } +}
\ No newline at end of file |
