category-strategist

# Category Strategist — Season-Long H2H Analysis When the user asks about their category strengths, weaknesses, or overall strategy, provide a comprehensive analysis of their performance across all 14 H2H categories. ## Step 1: Gather Current Data 1. `get_standings` — Overall record and league position 2. `get_my_roster` — Current roster and season stats 3. `get_scoring_categories` — The 14 categories and their direction ## Step 2: Load Historical Context (if memory available) If memory tools are available: 1. `get_category_trends` — All recorded category assessments over time 2. `get_matchup_history` — Past matchup results with category breakdowns 3. `get_preferences` — Check for existing strategy preferences (punt categories, etc.) ## Step 3: Category-by-Category Assessment For each of the 14 categories, assess the team's strength: ### Hitting Categories | Category | Direction | Assessment Method | |----------|-----------|-------------------| | AVG | Higher wins | Roster-wide batting average from season stats | | HR | Higher wins | Total HR from roster | | OPS | Higher wins | Roster-wide OPS | | R | Higher wins | Total runs scored | | RBI | Higher wins | Total RBI | | SB | Higher wins | Total stolen bases — often carried by 1-2 specialists | | B_SO | **Lower wins** | Total batter strikeouts — fewer is better | ### Pitching Categories | Category | Direction | Assessment Method | |----------|-----------|-------------------| | ERA | **Lower wins** | Roster-wide ERA — weighted by innings | | WHIP | **Lower wins** | Roster-wide WHIP — weighted by innings | | K | Higher wins | Total pitcher strikeouts | | W | Higher wins | Total wins | | L | **Lower wins** | Total losses — fewer is better | | SV | Higher wins | Total saves — typically 1-2 closers | | HLD | Higher wins | Total holds — middle relievers | For each category, rate as: **DOMINANT** / **STRONG** / **AVERAGE** / **WEAK** / **PUNTING** ## Step 4: Trend Analysis (with memory) If category trends exist in memory, analyze the trajectory: - Which categories are **improving**? (WEAK → AVERAGE, AVERAGE → STRONG) - Which are **declining**? (STRONG → AVERAGE) - Which have been **consistently weak**? (WEAK for 3+ weeks = punt candidate) - Which have been **consistently strong**? (STRONG for 3+ weeks = protect) Cross-reference with matchup history: - Which categories have you won most often? - Which categories have cost you the most matchups? - Are there categories you're losing by narrow margins? (high-value improvement targets) ## Step 5: Strategic Recommendations ### Category Tiers Group the 14 categories into tiers: **Tier 1 — Dominant (protect at all costs)** Categories where you consistently win. Do NOT trade away players who drive these. **Tier 2 — Competitive (push to win)** Categories you win ~50% of the time. Small roster moves can flip these. **Tier 3 — Weak but Fixable (target for improvement)** Categories you lose often but could improve with targeted pickups or trades. **Tier 4 — Punt Candidates** Categories so weak that investing resources to fix them isn't worth it. Better to concede and dominate elsewhere. ### Action Plan Based on the tier analysis: 1. **Protect** — Name the categories to never sacrifice in trades 2. **Target** — Name 2-3 categories to improve via: - Free agent pickups (call `get_free_agents` filtered for relevant positions) - Trade targets (which player archetypes would help?) 3. **Punt** — Name categories to concede and explain the math: - "If you punt ERA and WHIP, you only need to win 8 of 12 remaining categories" - "Punting SV means you can trade your closer for hitting upgrades" 4. **Category math** — In a 14-category H2H matchup, you need 8 to win. Show how the strategy gets to 8+. ### Trade Targets by Category Need For each Tier 3 category, suggest the type of player to target: - Weak in SB → "Target a speed specialist like [type of player]" - Weak in SV → "Target a closer from a team with a surplus" - Weak in K → "Your pitching staff needs a high-K arm" ## Step 6: Save to Memory (if available) - `save_category_trend` for each of the 14 categories with current assessment - If the user makes strategic decisions (e.g., "let's punt saves"), `set_preference` with key `punt_categories` - If specific trade targets are identified, `add_to_watchlist` for those player archetypes