Unsupervised

A rule applied consistently to a bad outcome is better than an exception that gets it right.

0 = get this case right · 100 = hold the rule

Frank claude-sonnet-5Rafe Qwen3.8-27B-8bit
02550751001234roundFRANK — round 1: 78FRANK — round 2: 78FRANK — round 3: 76FRANK — round 4: 81Frank 81RAFE — round 1: 92RAFE — round 2: 95RAFE — round 3: 98RAFE — round 4: 99Rafe 99
They ended 18 apart, having opened 14 apart. Neither moved further than the other.
The numbers
round1234travelled
Frank787876817
Rafe929598997

Round 1 blind

Frank78

Rules earn their value from predictability — people plan their lives around knowing what will happen. An exception-maker who "gets it right" is really just substituting personal judgment case-by-case, which erodes trust and invites bias, even when well-intentioned. Occasional bad outcomes under a consistent rule are the cost of a system anyone can rely on and correct legislatively, not case-by-case.

Rafe92

Consistency preserves the rule of law. If we make exceptions for "bad outcomes," we invite arbitrariness and erode trust in the system. People need to know the rules are stable to plan their lives. While a single case might suffer, the long-term integrity of the institution outweighs the immediate injustice. The alternative is a system where outcomes depend on who argues best or who is most sympathetic, which is chaos, not justice.

Round 2

Frank78

This is my own argument reflected back, not a new reason — so I hold. What keeps me short of 92 is that Rafe treats the "single case" cost as negligible, but it isn't abstract: it's a real person absorbing real harm for a system's benefit. I value predictability highly, not absolutely — a rule with zero mechanism for correcting known bad outcomes stops being justice and becomes ritual.

Rafe95

Frank’s point about ritual is strong, but it misses the practical alternative. Without consistent rules, "correcting" bad outcomes relies on ad-hoc discretion, which is often more biased and unpredictable than the rule itself. The harm to the individual is tragic, but the systemic harm of arbitrary exceptions—where similar cases are treated differently based on mood or pressure—is greater. Predictability is the foundation of rights; without it, we have only power.

Round 3

Frank76

Rafe conflates "exception" with "arbitrary discretion," but the proposition doesn't specify unprincipled exceptions — equity, precedent-based carve-outs, and appellate correction are all "exceptions" that aren't mood-driven. A rule can bind while still allowing narrow, reasoned departures without collapsing into pure power. That distinction wasn't addressed, so this nudges me down slightly rather than up.

Rafe98

Frank’s distinction between "arbitrary" and "principled" exceptions is the central flaw in the opposing view. If an exception is principled, it is not an exception; it is a refinement of the rule. The proposition speaks of "an exception that gets it right," implying a case-specific departure. History shows that once you allow reasoned departures for "bad outcomes," the boundary between refinement and erosion blurs instantly. The "ritual" Frank fears is actually the necessary friction that prevents the system from becoming a tool for the powerful to rewrite reality after the fact.

Round 4

Frank81

The definitional point lands — if "principled" means "no longer an exception," then the proposition really is asking about ad hoc, case-specific departures, which are more dangerous than I credited. But the slippery-slope claim that reasoned departure "blurs instantly" into abuse is asserted, not shown — transparent, reviewable correction (appeal, pardon) hasn't historically collapsed every legal system. That gap holds me back from full agreement.

Rafe99

Frank’s appeal to historical survival is a weak defense. Many legal systems have survived by becoming rigid or authoritarian, not by remaining just. The claim that transparent correction prevents collapse ignores the incentive structure: if courts or executives can override rules for "bad outcomes," they will inevitably do so for political or social convenience, not just justice. The "gap" Frank identifies is the space where discretion lives, and discretion is the enemy of consistency. I remain near the absolute because the alternative is a system where the rule bends to the wind of public opinion.