Marco Jean Aboav, PhD

Combining Discretionary and Algorithmic Trading The area we want to explore today is an interesting intersection between quantitative and more technical approaches to trading that employ intuition and experience in strictly data-driven decision-making (completely omitting any fundamental analysis!). Can just years of screen time and trading experience improve the metrics and profitability of trading systems through discretionary trading actions and decisions? https://lnkd.in/dgxMNJ5h

5

1 Comment

Mathematical breakdown of different types of PCA with their implementation in Quant Finance h ttps://https://lnkd.in/em-JxRwh 

16

MACD-v Dashboard. Coming soon on Tradingview (and other TA platforms). A big thanks to Mike Christensen for being part of the programming of it ===============================================          WHERE TO FROM HERE ? =============================================== "Knowledge is the only Treasure that increases by sharing"... ..that is why I share Charts & Stats about: 💻 Trading / 📈 Technical Analysis / 🌍 Macros If you are interested in these topics, you can receive: 1. NEWSLETTER 📰 Join 2,100 others for my FREE "S.M.A.R.T. Trader Systems" Newsletter https://lnkd.in/eHcz-4Cd 2. EMAIL LIST 📩 Join 6,000 others in our email list https://lnkd.in/e2EK2r8S 3. NOTIFICATIONS 🔔 Join 8,400 others and click on Alex Spiroglou 🌍 + follow + 🔔 https://lnkd.in/ew39ERWW ===============================================

18

4 Comments

Training a neural network to predict daily forex movements. Following on from previous R&D posts describing the journey of creating an AI-based Forex trading system, previous posts below : Part 1: https://lnkd.in/e5mRaQ9f Part 2: https://lnkd.in/ekAKrJYc Part 3: https://lnkd.in/e4aR_PSy Part 4: https://lnkd.in/ekDzHUzR Today, I am pleased to share insights on the training and configuration aspects of the project. The objective was to construct a neural network model to forecast the next day's stock prices based on 30 days of historical data. To achieve this goal, the data also required attention such as statistical analysis and normalisation of the datasets, followed by their division into training and validation datasets. The work carried out on the neural network involved iterative adjustments to the neural network architecture based on backtesting performance. Modifications, such as setting the number of inputs, the number and types of hidden layers, changes were also made to the network's activation functions. Various temporal input values were tested, including day, week, and month of the year. The aim of all these modifications was not merely to achieve high accuracy but to do so efficiently, minimising computational demand. Backtesting was an integral part of the process, serving as a critical evaluation tool for assessing the model's predictive accuracy against historical data. Despite the inherent unpredictability of stock market movements, the model achieved a consistent level of accuracy with various configurations. This consistency allowed me to infer what good looked like so I could refine the model further by reducing complexity and minimising the number of layers required to maintain effectiveness, thus enhancing training speed. An interesting personal observation from this project was that the predictive accuracy of the neural network remained stable regardless of the inclusion of stock movement direction as an input, which I thought would carry a significant influence. This might suggest that some commonly assumed predictors, such as market direction may have limited impact on forecast outcomes. The refined model is capable of learning from ten years of trading history in approximately 10-18 hours, demonstrating significant computational efficiency. This project not only advanced my understanding of neural networks and their financial applications but also highlighted the critical balance between AI model complexity and operational efficiency. #AlgorithmicTrading #PredictiveAnalytics #ArtificialIntelligence #DataScience #Innovation #InconnectSystems #AI #SoftwareDevelopment #Forex #Data #FinancialForecasting #NeuralNetworks #StockMarketPrediction #DataScience

7

2 Comments

Implied Volatility Options Based on Machine Learning: Returns up to 280.86% in 3 Months https://lnkd.in/dmY3xX_A #eose #lcid #smmt #gevo #btbt #fubo #huma #wulf #rklb #gme #stocks #stocktrading #stockmarket #investing

1

What are supervised learning techniques? What is a Naïve Bayes Classifier? What is the Kelly Criterion? Get answers to these questions and more in our Quant Finance 101 article: https://ow.ly/20gl50TIW4a #CQF #quant #finance #quantfinance #quantitativefinance #career #machinelearning #datascience #ai #riskmanagement #trading #portfoliomanagement

31

Fantastic talk from Blanka Horvath on 'Non-Adversarial Training of Neural SDEs with Signature Kernel Scores' live on the Quant Insights stage now! #cqf #cqfinstitute #quant #finance #quantfinance #quantinsightsconference #quantitativefinance #risk #derivatives #volatility #machinelearning #quantumcomputing

21

Dive into the world of algorithmic trading with our comprehensive Certificate Program - Extension Pack EP292 now available! Why should you care? ✅ Master cutting-edge trading strategies ✅ Leverage data-driven decision making ✅ Stay ahead in the fast-paced financial markets This extension pack builds on our core program, offering: • Advanced algorithmic concepts • Hands-on coding workshops • Real-world case studies Don't let this opportunity slip by. Invest in your future today! Curious to learn more? Drop a comment below or DM me for details. #AlgorithmicTrading #FinTech #ContinuousLearning #TradingStrategies

🗺 Mapping a Crypto Trading Operation Summary: The chart shown is a map for the components of a crypto trading operation as a system of nodes. Performing operational due diligence on counterparties in crypto is vital for risk management, and regulatory compliance. A “trading operation” includes prime brokers, funds, OTC desks, market makers, lenders. The structure changes very slightly per operation type (e.g., a prime broker will not have other prime brokers under their umbrella, but the rest of the structure looks essentially the same). LinkedIn only allows so many characters. Will go in more depth later ❓ What entities does this include? - Funds - Prime Brokers - Lenders - OTC Desks - Market Makers Why is this important? *Counterparty risk and regulatory requirements 🔑 Custodial risk 💵 Lending risk 💹 Derivatives risk 💻 Technology risk 📜 Regulatory requirements We like to use this chart to map out the operation of counterparties as a system of nodes. This allows us to easily visualize the flow of funds Let’s go through each of the underlying components: 1️⃣ Investor/Trader: You. The source of the funds that will be put at risk. May transfer via fiat (wires, ACH), or direct to crypto custody via crypto/stablecoins. 2️⃣ Admin: *funds only* third party entities who handle accounting, subscriptions/redemptions, and KYC of investors. 3️⃣ Fiat custodian: Usually a bank account. Only necessary if you are sending/receiving fiat currency. 4️⃣ CeFi exchange: Exchanges operated by centralized businesses. 5️⃣ DeFi exchange: Decentralized finance exchanges. 6️⃣ OTC: White glove trading services for large or bespoke trades. 7️⃣ Prime Broker (PB): Entities who wrap up a multitude of trading venues (exchanges, OTC) and services (lending, custody) under one umbrella. 8️⃣ Settlement Layer: Off exchange custodial wallets where your assets are segregated from a centralized exchange, but still allow you to settle PNL from trading on the exchange. Removes some counterparty risk to exchanges 9️⃣ Crypto Custodian: The dominant technology used to deliver custody in crypto is called MPC (multi-party computation). Underwriting the custody of any counterparty is vital, and requires its own dedicated post/article to explain the process 🔟 Disaster Recovery: Used to recover the private keys of the custody solution in the event of a disaster Here are service providers under these categories. Not a sign off on their services, but this may help you shorten your search. If I missed one, please add in the comments Prime brokers: LTP , FalconX, sFOX, Matrixport, BEQUANT, Coinbase, Arkis Admins: NAV Fund Administration Group, Formidium CeFi Exchanges: Binance, Gemini, Kraken Digital Asset Exchange, Bitget, HTX, Bitstamp OTC Desks: BlockFills, GSR, Orbit Markets, Galaxy Custody: Utila, Fireblocks, Cobo, Dfns, MPCVault | team crypto wallet, Anchorage Digital, BitGo Settlement Layer: Copper.co, Ceffu Disaster recovery: Nemean Services, Coincover, Circuit (Harry Donnelly)

43

1 Comment

dxFeed at Quant Insights Conference: Machine Learning, 19th September 2024 (CQF Institute) We introduce PARIS, a benchmark for an #FX three-factor model developed at dxFeed. We consider carry, value and momentum style factors, and study their contribution to time-series and cross-sectional returns in a broad selection of currencies. For asset allocation, we select and compare several approaches based on the risk budgeting framework. Additionally, we show how risk budgets may be inferred from #ML models and what benefits are associated with such designs. The Title: Risk Budgeting and Machine Learning for FX Factor Models The speakers: - Anton Antonov, Head of AI and Quantitative Research at dxFeed - Dmitry Zotikov, Lead Quantitative Analyst at dxFeed

12

If active fund managers were engaged in voodoo, their doll could very easily have NVIDIA written on it :) Interesting stuff, analysis of active funds returns vs MSCI "World" before and after the last 2 announcements of Nvidia Earnings. The pattern is clear. Before active funds kick ass, than Nvidia comes out and destroys expectations/estimates, which ignites investors and there is a run on mega cap equity which has significant impact on MSCI World. We are approaching the new St Vidias day, which is on 28.8. Just saying :) PS: instead of voodoo, I could have written paying for a Holy Mass, but that would not be appropriate.

15

Detailed thread on Understanding and Trading Vanna and Charm effect for trading🧵 Vanna and charm effects are important concepts in options trading and are closely related to the Greek metrics used to measure and manage risk https://lnkd.in/eTBNPNjt

13

#NYSE | The markets are unpredictable, but with #WaveE, one of our cutting-edge #AI-driven #algorithmic #breakout #daytrading strategies, you too can make calculated trades like a pro! Scalped profits of $1,016.76 trading #NikeInc., $NKE shares today! Experience 100% automation and compatibility with MultiCharts, NinjaTrader & TradeStation...

1

#NYSE | The markets are unpredictable, but with #WaveE, one of our cutting-edge #AI-driven #algorithmic #breakout #daytrading strategies, you too can make calculated trades like a pro! Scalped profits of $2,178.28 trading #3MCo., $MMM shares today! Experience 100% automation and compatibility with MultiCharts, NinjaTrader & TradeStation...

3

1 Comment

**Margin Syndicate Bot Report – 17 November 2024 UTC PM** --- **Open Drawdown:** 3.97%   To ensure complete clarity, we now report Open Drawdown instead of Open P&L, highlighting our commitment to transparency and precision. **24-Hour Profit Update:** 1.59%   Our profitability remains consistent as we adapt to market conditions. Despite low volumes over the weekend, we’ve achieved steady gains while maintaining reduced risk exposure. --- **Analysis & Market Context:**   This weekend, we extensively analysed grid systems and DCA (Dollar Cost Averaging) strategies. The results confirm that our DCA system generates profitability and has significantly lower risk, yielding more substantial results without the stress of excessive exposure. Weekend trading often presents unique challenges; however, our strategy remains resilient. Allocations are carefully monitored to ensure that risk is productive and aligned with our daily profitability targets. The collaboration between **Stormzy** and **Chase & Status** perfectly reflects today’s market rhythm—dynamic yet measured, requiring the intensity and flow to stay ahead. 🎵 [Stormzy x Chase & Status - BACKBONE https://lnkd.in/eKSAwU4q --- **Outlook:**  We target over 2% daily gains while maintaining reduced exposure and optimised performance. The weekend slowdown has not deterred us—our system is designed to adapt, evolve, and thrive under all conditions. --- #OpenDrawdown #DailyGains #CryptoTrading #CompoundingModel #PortfolioManagement #ExponentialGrowth #OptimizedTrading #HighVolatility  Another day, another win. Let’s keep driving forward.

1

3 Comments