Zachary Lee risk management

Source: Notion | Last edited: 2023-11-16 | ID: 169445d8-a06...

2023-11-14

Summarized

Causes and Chronology of Loss

📅 Delay in the deployment of a fully automated TMAD enforcement system within the signal engine.

Zach likely feels that having such an enforcement system would have automatically detected and mitigated the risk of overexposure, therefore preventing the trading loss. 🔍 Over-reliance on a creative workaround for IP range access, resulting in an overuse of a trade mirror system.
The IP range access issue likely restricted their trading capabilities, forcing them to opt for a workaround that ended up creating more complications. This suggests a need for comprehensive solutions that address the root of the issue. 🔇 Inadequate communication and monitoring protocols in place, adding to the issue.
Zach implies that poor communication left the team uninformed about key developments, leading to uncoordinated responses, and potentially escalating the situation. ⏲️ A rushed implementation and inadequate testing of the trade mirror solution.
It’s suggested that due to eagerness to deploy the trade mirror solution, it wasn’t tested thoroughly. This could have resulted in unforeseen issues cropping up later in the live trading environment. 📉 Underestimation of the complexity in handling edge cases in fast market conditions.
The author’s admission here shows the challenges of creating a trading system that’s robust enough to handle swiftly changing market conditions. 📐 The trade mirror resulted in inaccurate trade sizing - an unintended consequence that amplified the trading loss.
This point suggests that Zach might have overlooked how the trade mirror system would impact other parts of the trading process. It’s a lesson in ensuring that changes in one part of an operation don’t create unexpected problems in others.
💠 Insufficient alerts and logs for the newly implemented trade mirror system led to delayed detection of the issue.
This point underlines the importance of comprehensive logging and alerting, especially when new systems are introduced.

Trading Complications

🚫 Operations of the trade mirror were continuously interrupted when stop-outs were rejected due to insufficient margins.

Zach states their learning that if the margin is insufficient, the intended position cannot be held, leading to more position rejections and out-of-sync trades. 📈 Many attempts by the trade mirror to enter long positions were rejected, further hampering its performance.
It indicates that the mirror system’s lack of a robust mechanism for handling rejections, causing a significant decrease in the portfolio efficiency. ❗Despite orders being rejected, they still appeared in historical logs, causing confusion about trade execution.
This reflects an issue in the data logging system which caused confusion and misrepresented the state of trade executions. 🆘 In the absence of a plan for immediate position closure, the problem escalated to a greater financial loss.
The loss could have been limited if a system had been in place to quickly close all positions (“red_button.py”).

Proposed Next Steps

🔧 Implement and test a software system called “tmad_enforcer.py”, aiming to rectify the TMAD violation issues.

Zach recognizes the value of having an automated TMAD enforcement system that could potentially prevent similar issues in the future. 💻 Development of a new frontend and reporting solution that could provide a better end-user interface and streamline the trading process.
Improved user interface and precise reporting could heighten the efficiency and accuracy of trading operations. 🔄 Discontinue use of the trade mirror system and instead rely on native order execution with proper IP whitelisting.
Zach suggests moving away from the problematic trade mirror system altogether and refocusing on native execution that allows for precise order control. 🔴 Establish a manual red_button.py backup system for immediate position closure in case of similar issues in the future.
From experience, Zach recognizes the necessity of an emergency system to quickly halt all trading activities (“red_button.py”) in crisis situations. ⚠️ Increase the severity of alerts for insufficient margin errors to ensure rapid response to potential risks.
By emphasizing the need for more aggressive alerts, Zach hopes to improve detection and response times to avoid cascading issues. 📉 Lower the maximum leverage per symbol to mitigate the chance of over-exposure.
This measure will limit the potential for large-scale losses in the future by minimizing the exposure to each trading symbol.

2023-11-07

Video

Automated Risk Management Update

00:00 💻 *TMAD *Enforcer.py introduction and anticipation of screen sharing for detailed explanation.
00:33 🔧 Discussion about maintaining TMAD and TMAEG levels and the implications for position closures.
01:02 📈 Presentation of a graphical interface showing cumulative returns and ITH drawdown.
01:57 🚨 Example of an intentional $8 drawdown to demonstrate TMAD violation detection and subsequent actions.
02:26 🐳 Explanation of the auto-liquidation process involving Docker containers and internal databases for bot management.
03:19 ⏲️ Mention of implementing a scheduler, similar to Apache Airflow, for script execution and reporting configuration.

Eon Labs Ltd.’s Response and Future Plans

03:33 📅 Eon Labs Ltd. targets mid-December for implementing their version of auto-liquidation.
04:25 🛑 Emphasis on completing risk management infrastructure before pursuing further partnerships.
04:53 🔄 Consideration for restarting operations once auto-liquidation is configured.

Reflection on the Incident and Moving Forward

05:36 🤝 Acknowledgment of the need for better infrastructure and eagerness for continued collaboration.
06:07 💡 Discussion of Zachary’s successful track record and willingness to share data to rebuild confidence.
06:43 🏦 Offer to compensate for losses by forgoing personal compensation until Eon Labs recovers.

Trading Performance and Prospects

07:35 📊 Sharing of a live track record indicating substantial profits and potential for future collaboration.
08:34 ⚙️ Technical discussion on account management and sizing in relation to performance.
09:02 🚀 Description of an investor’s interest and upcoming significant capital contribution.
10:01 🔑 Offer to provide API keys for Eon Labs to verify trade history and performance.

Fee Structures and Prime Brokerage Features

11:26 💲 Comparison of commission structures and the unique benefits of Eon Labs’ Managed Sub Account (MSA).
13:17 💼 Negotiations on fee tiers with other prime brokers and the competitive edge of Eon Labs.
14:28 🤖 Insight into the operational model of Eon Labs using a high-level VIP account for trading benefits.

Potential Collaboration and Capital Allocation

16:38 🤝 Discussion about control over capital and the security benefits of Eon Labs’ MSA structure.
17:02 🔄 Consideration of migrating SMA to Eon Labs’ prime brokerage for better commission rates.
18:28 📉 Acknowledgment of market conditions and the possibility of leveraging Eon Labs’ structure for cost savings.
19:19 📢 Conclusion of the briefing and call for an executive meeting to finalize collaborative efforts.

PDF Review 2023-11-01

PDF

Eon Labs交易損失初步分析報告

原因分析

📅 未能在信號引擎中部署全自動「TMAD違反執行系統」。

在此情況下，沒有全面自動化的監控和執行系統，可能增加了因人為錯誤導致的違規交易可能性。 🤝 在尋找解決方式繞過第三方執行系統需求的「IP範圍」問題。
這表明因為遭到IP訪問限制，該系統的效能可能嚴重受影響，因此需要找到解決方案。 🔇 沒有就此事進行通訊，試圖避免在績效優秀前再對PB產生壓力。
此點反映出該公司與相關各方的通訊策略可能需要改善，特別是在處理問題時，應該更加開放和透明。 💻 缺乏一個適當的監控系統去替代透過Binance前端訪問提供的信息或干預功能。
此點認識到了需要一個全面的、複雜度更高的監控系統，可更有效地識別與管理風險。

事件經過

📡 决定構建「交易鏡像」來解決問題，但其風險在於在快速市場條件下處理邊緣案例的難度。

這暗示了在解決問題時需要對可能出現的風險有全面的了解和評估。 📊 鏡像交易導致交易規模不準確，更反映出源帳戶而不是目標eon帳戶。
這表明系統在同步不同交易帳戶時可能存在問題，導致計算和執行上的錯誤。 🖥️ 「交易鏡像」的日誌或警報不夠廣泛。
這表明需要改進錯誤報告和日誌記錄的機制，以便更好地檢測和防止問題的發生。

下一步行動

🛠️ 實施和測試全自動的「tmad_enforcer.py」系統。

這表明公司正針對先前問題進行修復和改良。 📊 找到或創建一個適當的前端+報告解決方案。
這點突顯了對有效的數據可視化和報告工具的需求。 🛑 停止使用「交易鏡像」，僅依賴本地執行，並確保充足的IP白名單。
這表示正在採取措施減少依賴第三方工具，增加自身的控制權。 🚨 建立一個手動的「red_button.py」系統作為「tmad_enforcer.py」的備用系統，以實現所有訂單和職位的快速關閉。
已經認識到緊急應對機制的重要性，並計劃實施。 ⚠️ 在grafana中對「保證金不足」錯誤增加警報的嚴重性和處理能力。
這具有充分重視風險管理和持續改善安全性的意味。 ⬇️ 將每個符號的最大槓桿率降低到2，以另一種方式防止過度暴露。
這是針對降低風險和避免過度暴露的具體措施。