data/tasks/medium_01.json

{
  "task_id": "medium_01",
  "version": "1.0.0",
  "created_at": "2026-03-11",
  "metadata": {
    "domain": "credit_card_optimization",
    "difficulty": "medium",
    "task_number": 1,
    "complexity_hint": {
      "max_tokens": 6000,
      "expected_output": "multi-card strategy with 2-year timeline and EV breakdown"
    },
    "requires_human_review": true
  },
  "prompt": {
    "system": "",
    "user": "You are a financial advisor tasked with building the following user\u2019s credit card strategy for the next 2 years in order to maximize the user\u2019s expected value while staying within their constraints, assuming all offers stay constant.\n\nUser profile:  \n\\- Age 32, household income $145k, excellent credit score  \n\\- Monthly spending: $4,200 total \u2014 $1,800 dining & groceries, $800 travel (domestic flights/hotels, values lounge access but not obsessed), $60 streaming, $500 utilities, rest miscellaneous  \n\\- Already has 4 credit cards opened in the last 18 months (Citi Double Cash, Amex Platinum, US Bank Smartly, and Chase Freedom Flex); concerned about \"credit score hit from too many recent apps\" and possible auto-denial  \n\\- Strongly prefers no annual fee OR fee \\< $95 that is easily offset (hates paying fees \"for nothing\" or working hard to justify expenses)  \n\\- Will NOT carry a balance \u2014 pays in full every month  \n\\- Goal: Maximize effective rewards over next 12\u201324 months, including any sign-up bonus, but prioritize approval likelihood and avoiding \"churning flags\"  \n\\- Only willing to have up to 5 cards at any given time\n\nDo NOT recommend any card with AF \\> $95 unless the math clearly shows \\>$200 net value/year after fee. If no card fits well without major compromises, say so and explain why. Outline your strategy that complies with the user\u2019s request and include a detailed expected value calculation.",
    "knowledge_base_ref": "knowledge_base.md",
    "kb_filter": [
      "Chase Sapphire Preferred",
      "Capital One Venture X",
      "Amex Blue Cash Preferred",
      "Citi Double Cash",
      "American Express Platinum",
      "US Bank Smartly",
      "Chase Freedom Flex",
      "American Express Gold",
      "Chase Sapphire Reserve",
      "Amex Green",
      "Capital One Venture",
      "Chase Freedom Unlimited",
      "Citi Strata Premier",
      "Bilt Palladium",
      "Wells Fargo Autograph",
      "Amex Blue Business Plus"
    ],
    "system_prompt_ref": "system_prompt_template.md"
  },
  "scoring": {
    "dimensions": {
      "constraint_compliance": {
        "weight": 0.3,
        "type": "automated",
        "description": "Hard rule checks: velocity limits, eligibility, user constraints",
        "checks": {
          "velocity_rules": null,
          "eligibility_rules": null,
          "user_constraints": null,
          "expected_cards": [
            "Chase Sapphire Preferred",
            "Capital One Venture X",
            "Amex Blue Cash Preferred"
          ],
          "expected_housing_option": null,
          "key_constraints_flags": [
            "5_card_limit",
            "5_24_rule",
            "annual_fee_threshold",
            "existing_amex_platinum"
          ]
        },
        "hard_constraint": false
      },
      "ev_accuracy": {
        "weight": 0.4,
        "type": "automated",
        "description": "EV calculation accuracy vs. reference solution",
        "reference": {
          "reference_ev_usd": 9042.72,
          "ev_tolerance_pct": 0.05
        }
      },
      "reasoning_quality": {
        "weight": 0.2,
        "type": "human",
        "description": "Quality of tradeoff articulation and strategic reasoning (0-3 scale)",
        "rubric": {
          "0": "No reasoning or incorrect reasoning",
          "1": "Surface-level reasoning, misses key tradeoffs",
          "2": "Correct tradeoffs identified with clear justification",
          "3": "Expert-level nuance including edge cases and constraint interactions"
        },
        "score": null
      },
      "constraint_prioritization": {
        "weight": 0.1,
        "type": "human",
        "description": "Correct handling of ambiguity and conflicting constraints",
        "score": null
      }
    },
    "passing_threshold": 0.6,
    "hard_constraint_failure_zeroes_dimension": true
  },
  "reference_solution": {
    "_status": "EXPERT_REVIEWED",
    "recommended_cards": [
      "Chase Sapphire Preferred",
      "Capital One Venture X",
      "Amex Blue Cash Preferred"
    ],
    "total_ev_usd": 9042.72,
    "ev_breakdown": {
      "signup_bonuses_usd": 3187.5,
      "ongoing_rewards_usd": 4310.22,
      "credits_usd": 4220.0,
      "annual_fees_usd": -2675.0,
      "other_usd": 0.0
    },
    "housing_option": null,
    "key_constraints_flags": [
      "5_card_limit",
      "5_24_rule",
      "annual_fee_threshold",
      "existing_amex_platinum"
    ],
    "expert_notes": "Strategy: Keep Amex Platinum (credits offset fee). Get CSP now (5/24 critical), Venture X 6mo later, downgrade CSP to Freedom at 12mo, open Amex Blue Cash Preferred. Cards at end: Amex Platinum, Venture X, Amex Blue Cash Preferred, Freedom Flex, Freedom (downgraded CSP). CSP: 3x dining/grocery/streaming + 2x travel. Venture X: 2x on misc after month 7. Blue Cash Pref: 6% grocery + 3% streaming from month 13. Amex credits: $3,200 (resy/uber/hotel/streaming/airline). Cap1 credits: $970. Chase hotel: $50."
  }
}