28d • General discussion
First Meeting Recap
Hey everyone, we had our first intro meeting last Friday. About 5-6 of us were able to make it and had a cool conversation! We left with some takeaways for ideas on how to contribute to the group.
We figure
- we can do presentations on our algo projects and have them peer reviewed by the group
- we can work on a project that tackles a pillar of algorithmic trading as a group.
let me know what you think would be best!
Attatched are the meeting notes:
Meeting Purpose
Introductory call to connect on algorithmic trading projects and challenges.
Key Takeaways
- The "Edge" Debate: The group debated where a retail edge exists. Nik argued for using high-resolution MBO data to detect institutional order flow before patterns form, while Surya advocated for simple, robust price-action strategies that avoid parameter optimization and overfitting.
- Data Quality is Paramount: Alpaca's free data is unreliable (UEX feed covers only ~3% of the market), making its $90/mo SIP feed or a premium vendor like DataBento necessary for accurate backtesting.
- Alpaca's PDT Trap: Alpaca's margin-account-only structure forces all users into Pattern Day Trader (PDT) rules, requiring a $25k balance to avoid trade limits—a critical, often overlooked detail.
- Surya's High-Return Strategy: A simple, 15-min price-action strategy shows >30% annual returns on Tesla since 2021. The group will analyze its logic to understand why it works on specific stocks and how to generalize it.
Topics
Member Intros & Project Status
- Jared: Developer (ex-MSFT/IBM) focused on a prop firm (FTMO) challenge and local algo development.
- Zach: Scientist learning Python for quant finance; building a backtesting engine but struggling to find a profitable strategy.
- Dennis: Medical genetics grad starting a data science Master's; seeking a clear roadmap for algo trading.
- Nik: Data analytics veteran building complex options/futures strategies; paper trading results don't match backtests, prompting a shift to simpler futures strategies.
- Surya: Python developer with a backtested strategy showing >30% annual returns on Tesla since 2021.
The "Edge" Debate: Order Flow vs. Price Action
- Nik's Argument: Order Flow & MBO DataThe edge is in using high-resolution MBO (Market By Order) data to detect institutional order flow (e.g., iceberg orders, exhaustion) before patterns form.This approach aims to "ride on the backs of the giants" by following their moves, not competing on speed.Challenge: MBO data is expensive (Nik has spent ~$500 on DataBento).
- Surya's Argument: Simple, Robust Price ActionThe edge is in simple, repeatable price-action patterns that work across market conditions.Warning: Avoid parameter optimization, which leads to overfitting and strategies that fail in live trading.Surya's Strategy:Candle Pattern: Identifies support/resistance based on a 3-candle pattern (middle candle's high/low is an extreme).Entry: Long on a close above resistance; short on a close below support.Stop-Loss: The low of the previous candle, creating a very tight stop (~0.1–0.5%) that protects capital in non-trending markets.Performance: >30% annual returns on Tesla since 2021, including 9% in Jan 2026.
Data & Brokerage Infrastructure
- Data Sources:Alpaca: Free UEX data is unreliable (covers only ~3% of the market). The $90/mo SIP feed is required for quality data.DataBento: A trusted source for granular data (tick, MBO). Cost-effective for small downloads but expensive for large datasets.Nik's Data Assets: Willing to share his DataBento downloads (3 years of NQ/ES futures, several months of mega-cap stocks) to help others.
- Brokerage APIs:Alpaca: Margin-account-only structure triggers PDT rules, requiring a $25k balance to avoid trade limits.IBKR: Requires a $500 Pro account to access its paper trading gateway.Schwab: API is usable but requires frequent manual OAuth token renewal.
Group Collaboration & Future Direction
- Proposed Project: "Best Process Framework"Nik proposed a collaborative project to map out a robust, developer-centric framework for each component of an algo trading system (backtesting, execution, risk management).Goal: Create a flexible foundation to avoid the need for refactoring later.
- Future Meeting:Surya will present the logic for his high-return strategy to the group for peer review and analysis.
Next Steps
- Surya:Research the feasibility of using order book data to predict price bursts before investing in granular data.Present the logic for the high-return strategy in a future meeting.
- Nik:Share research papers on MBO data analysis with Surya.Share his DataBento downloads (futures, mega-caps) with interested members.
- All:Discuss the "Best Process Framework" project in a future meeting.
5
3 comments
First Meeting Recap
powered by
skool.com/strat-quant-career-bootcamp-3467
A study group for software engineers to land a 200k+ job | Free for first 150 members.
Suggested communities
Powered by