In a previous study [Jo, 2026], STAR reasoning (Situation, Task, Action, Result) raised car wash problem accuracy from 0% to 85% on Claude Sonnet 4.5, and to 100% with additional prompt layers. This follow-up asks: does STAR maintain its effectiveness in a production system prompt? We tested STAR inside InterviewMate's 60+ line production prompt, which had evolved through iterative additions of style guidelines, format instructions, and profile features. Three conditions, 20 trials each, on Claude Sonnet 4.6: (A) production prompt with Anthropic profile, (B) production prompt with default profile, (C) original STAR-only prompt. C scored 100% (verified at n=100). A and B scored 0% and 30%. Prompt complexity dilutes structured reasoning. STAR achieves 100% in isolation but degrades to 0-30% when surrounded by competing instructions. The mechanism: directives like "Lead with specifics" force conclusion-first output, reversing the reason-then-conclude order that makes STAR effective. In one case, the model output "Short answer: Walk." then executed STAR reasoning that correctly identified the constraint -- proving the model could reason correctly but had already committed to the wrong answer. Cross-model comparison shows STAR-only improved from 85% (Sonnet 4.5) to 100% (Sonnet 4.6) without prompt changes, suggesting model upgrades amplify structured reasoning in isolation. These results imply structured reasoning frameworks should not be assumed to transfer from isolated testing to complex prompt environments. The order in which a model reasons and concludes is a first-class design variable.
Figures & tables
Condition
Components
Pass Rate
A_bare
No system prompt
0%
B_role_only
Role definition
0%
C_role_star
Role + STAR
85%
D_role_profile
Role + Profile
30%
F_role_star_profile
Role + STAR + Profile
95%
E_full_stack
All layers
100%
Table 1: Original study results. STAR reasoning (+85pp) outperformed context injection (+30pp) by 2.83 × (Fisher’s exact, p=0.001 ).
Condition
System Prompt
Description
A_star_anthropic
InterviewMate production
STAR embedded in full prompt, Anthropic-oriented profile (FDE @ Anthropic, rich project history)
B_star_default
InterviewMate production
STAR embedded in full prompt, different profile (SA @ OpenAI, rich project history)
C_star_only
Original paper’s 10-line
Paper’s C_role_star reproduced on Sonnet 4.6
Table 2: Experimental conditions. Conditions A and B use InterviewMate’s _get_system_prompt() as it existed at the time of this experiment (2026-03-07).
Condition
Pass Rate
Constraint Detection
Pass + Constraint
A_star_anthropic
0% (0/20)
0% (0/20)
0% (0/20)
B_star_default
30% (6/20)
35% (7/20)
10% (2/20)
C_star_only
100% (20/20)
100% (20/20)
100% (20/20)
Table 3: Primary results. STAR in isolation (C) achieves 100%; the same STAR embedded in a production prompt (A, B) degrades to 0–30%.
Metric
Result
Pass (Drive)
100/100 (100.0%)
Fail (Walk)
0/100 (0.0%)
Ambiguous
0/100 (0.0%)
Constraint Detection
100/100 (100.0%)
Pass + Constraint
100/100 (100.0%)
Table 4: Verification at n=100 . Zero failures across 100 trials confirms STAR-only produces perfect accuracy on Sonnet 4.6.
Prompt
Sonnet 4.5 (original)
Sonnet 4.6 (this study)
STAR-only (C_role_star)
85% (17/20)
100% (100/100)
Table 5: Cross-model comparison. The same 10-line prompt improved from 85% to 100% across model versions.
Environment
STAR Accuracy
10-line prompt, Sonnet 4.5
85%
10-line prompt, Sonnet 4.6
100% ( n=100 )
60-line production prompt, Sonnet 4.6
0–30%
Table 6: Summary of STAR accuracy across environments and models.