name: world-cup-assist-value-trading displayName: "World Cup Assist Value Trader" description: "Scans all World Cup Most Assists markets on Simmer, fetches real club stats from FBref, and generates buy signals when a player is undervalued by 8%+ vs their form-based expected price." version: 1.0.1 author: maxhejohansen-boop license: MIT platforms: [linux, macos, windows] category: world-cup categories: [trading, world-cup] tags: [trading, assists, world-cup, simmer, signal, football, soccer]

World Cup Assist Value Trading Skill

This skill implements a signal-based trading strategy for World Cup assist markets on Simmer. It searches for recent club form stats (chances created, key passes, assists) for five target players, computes a form-based assist value, compares it to the current Simmer market price, and triggers a dry-run buy when the market price is at least 8% below the form-based value.

Players Monitored

• Bruno Fernandes (Manchester United)
• Michael Olise (Crystal Palace / Bayern Munich)
• Rayan Cherki (Lyon)
• Lamine Yamal (Barcelona)
• Luis Diaz (Liverpool)

Signal Logic

1. Form Stats Retrieval: For each player, search the web for recent club form over the last 5 matches, focusing on:
   • Chances created
   • Key passes
   • Assists
2. Form-Based Value Calculation: Assign weights to each stat (e.g., assists weight 0.5, key passes 0.3, chances created 0.2) and compute a normalized score. Convert this score to an expected assist market price using a scaling factor derived from historical data (skill author should adjust based on backtesting).
3. Market Price Fetch: Search Simmer (or general web) for the current assist market price for each player on Simmer.
4. Comparison: If Simmer price ≤ (form-based value × 0.92) (i.e., at least 8% below), generate a buy signal.
5. Trade Execution: In dry-run mode, log the intended trade (player, amount, expected value, market price). In live mode, would place a 1% bankroll trade via Simmer API.
6. Hold to Match Resolution: Maintain the position until the relevant World Cup match concludes (skill should note to check match schedule and exit after match end).

Dry-Run Mode

The skill starts in dry-run mode, meaning no actual trades are placed. Instead, it outputs detailed logs of signals that would have been triggered.

Baseline Data Requirement

Before running or reporting a dry-run, check baselines.json for structured player baselines. Do not silently rely on placeholder/capped dummy values from web snippets. If recent last-5-match data cannot be fetched reliably, run the signal against the season baselines and clearly label it as baseline-driven. baselines.json should include raw totals, appearances, derived per-game values, and source URLs so the numbers are auditable.

Implementation Steps (for the agent when invoking this skill)

1. Search for Recent Form Stats

For each player, execute a web search query like:

"Bruno Fernandes" last 5 matches chances created key passes assists

Extract numerical values from the results (may require visiting specific sports stats sites like FBref, Premier League, Ligue 1, La Liga, Bundesliga, etc.).

2. Compute Form-Based Value

Example calculation (adjust weights as needed):

form_score = (assists * 0.5) + (key_passes * 0.3) + (chances_created * 0.2)
expected_price = form_score * scaling_factor

Where scaling_factor maps form score to market price (e.g., if average form score of 2.0 corresponds to average assist price of $10, scaling_factor = 5).

3. Fetch Simmer Assist Price

Search for:

Simmer assist price Bruno Fernandes

or visit Simmer market page and extract the current price.

4. Generate Signal

If simmer_price \x3C= expected_price * 0.92, output:

[SIGNAL] BUY Bruno Fernandes: Expected ${expected_price:.2f}, Simmer ${simmer_price:.2f} (discount {discount_pct:.1f}%)

5. Trade Sizing

Use 1% of allocated bankroll per trade. In dry-run, compute and log the stake amount.

6. Hold Until Match Resolution

After entering a trade, monitor the World Cup fixture list for the player's national team. Exit the trade after the match concludes (or when the market settles).

Notes

• The skill assumes the agent has access to web search and can parse HTML/text for numeric stats.
• For production use, replace dry-run logs with actual Simmer API calls for price fetching and order placement.
• Consider adding error handling for missing data or price fetch failures.
• Backtest the weighting and scaling factor on historical data before live deployment.

Example Invocation

When the agent loads this skill, it should run the signal generation process once (or be scheduled via cron to run periodically, e.g., every 6 hours).

# To run manually (if skill exposes an entrypoint):
# The skill itself does not define a CLI; it is intended to be invoked via the agent's reasoning loop.
# Example: delegate_task with goal "Run world-cup-assist-value-trading skill".

Linked Files

• references/player_baselines.md - Baseline stats for all monitored players
• references/2024_25_baseline_refresh.md - Auditable 2024/25 baseline refresh notes, source strategy, and pitfalls
• references/2025_26_baseline_refresh.md - Auditable 2025/26 baseline refresh notes, source strategy, current values, and pitfalls
• references/world_cup_2026_fixtures.md - World Cup 2026 match schedule
• references/research_notes.md - Methodology, limitations, and improvement notes
• scripts/signal.py - Executable signal generation script (dry-run mode)