Concentrated Investing with Adversarial Thesis Review
Summary
This example describes an AI trading team modeled on concentrated investing. A quality researcher selects a high-quality large-cap company, an activist bull develops the case for catalysts and value creation, and a short-seller challenges the thesis on leverage, governance, accounting, competition, and valuation. A portfolio manager then checks account and order state and decides whether the surviving idea merits a single concentrated position, capped at 25% of portfolio value.
The workflow also specifies a source check using Pershing Square’s SEC company feed before holding a position, with Yahoo daily prices. It describes broker-connected execution and a backtesting option, but provides no performance results or evidence that the process improves returns. The document is an implementation example, not a validated investment method; its research roles and source checks do not establish that a candidate is sound, and concentration leaves the portfolio exposed to losses if the thesis fails.
Key ideas
- The workflow selects a small number of understandable, high-quality companies for concentrated research.
- A bull case is challenged by a bear case covering business, governance, accounting, and valuation risks.
- A portfolio manager makes the final position decision and limits one position to 25% of portfolio value.
- The example checks an SEC company feed before holding and uses daily prices from a market data provider.
- The document gives setup and backtesting instructions but reports no strategy performance.
Tags
Full text
# agents example bill ackman concentrated Bill Ackman Concentrated AI Trading Team ======================================== .. meta:: :description: This strategy is inspired by Bill Ackman and Pershing Square-style concentrated investing: do deep work on a small number of understandable, high-quality businesses. .. image:: ../docs/assets/ai-trading-team-workflows/bill-ackman-concentrated.png :alt: AI trading team workflow for Bill Ackman concentrated style investing :width: 100% This strategy is inspired by Bill Ackman and Pershing Square-style concentrated investing: do deep work on a small number of understandable, high-quality businesses, make a strong bull case, invite a brutal bear case, and then act with conviction if the thesis survives. The team is intentionally adversarial. The quality researcher finds the best candidate, the activist bull looks for catalysts and value creation, the short-seller bear attacks the thesis, and the portfolio manager decides whether one concentrated position is still justified. How the team works ------------------ * ``quality_researcher`` finds the best high-quality large-cap candidate. * ``activist_bull`` argues for catalysts, pricing power, and value creation. * ``short_seller_bear`` attacks leverage, governance, accounting, competition, and valuation risk. * ``portfolio_manager`` is the dedicated trading-and-risk agent. It verifies account and order state, then holds or sizes one surviving idea to at most 25% of portfolio value. * The source proof fetches the live SEC company atom feed for Pershing Square, CIK 0001336528, before it holds. Yahoo supplies the daily prices. Run it with a broker -------------------- The file defaults to broker-connected execution. With Alpaca, it runs in paper mode unless you set ``ALPACA_IS_PAPER=false``. .. code-block:: bash export OPENAI_API_KEY='your-key-here' export ALPACA_API_KEY='your-alpaca-key' export ALPACA_API_SECRET='your-alpaca-secret' export ALPACA_IS_PAPER=true python lumibot/example_strategies/ai_trading_team_bill_ackman_concentrated.py Backtest it ----------- Use the same strategy class and change ``IS_BACKTESTING = False`` to ``IS_BACKTESTING = True`` in the runner: .. code-block:: bash export OPENAI_API_KEY='your-key-here' python lumibot/example_strategies/ai_trading_team_bill_ackman_concentrated.py Example code ------------ .. literalinclude:: ../lumibot/example_strategies/ai_trading_team_bill_ackman_concentrated.py :language: python
Shown in full with attribution under the source's licence. Licence: GPL-3.0
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.