All (110)Personal Introduction (1)AI Summary (8)Project Management (4)Investment Strategy (9)Financial Market Analysis (6)AI Software Engineering (9)AI Research (7)Content Management (9)Cognitive Development (1)Philosophical Reflection (3)Entrepreneurship (1)AI Tools Development (9)Product Development (3)Troubleshooting (2)Quantitative Finance (12)AI Cost Analysis (1)System Architecture (1)Simulation Framework (1)Technical Log (3)AI Social Systems (3)Personal Reflection (1)
FMAB Signal Performs Excellently, Ready for Live Trading Deployment
Quantitative Finance
👤 Cryptocurrency traders, quantitative strategy developers, investment analysts
This article reports the excellent performance of the FMAB signal on ETH, with a baseline return of 100% and an increase to 4600%~120000% after anti-Martingale betting, far surpassing other strategies. The author announces plans for live trading, emphasizing the need to prepare capital for a long-term engineering effort. It also notes that the FMA signal performed poorly, causing drawdowns, and reflects on insufficiently scientific position management. Based on test results from February 11, 2026, the article aims to share signal strategy validation and live trading deployment plans.
- ✨ FMAB signal achieved a baseline return of 100% on ETH, increased to 4600%~120000% after anti-Martingale betting
- ✨ Preparing for live trading, requiring capital preparation for a long-term engineering effort
- ✨ FMA signal performed poorly, causing drawdowns, reflecting on insufficiently scientific position management
📅 2026-02-11 · 213 words · ~1 min read
Performance Analysis of Anti-Martingale Betting Strategy in BTC Trading
Quantitative Finance
👤 Quantitative traders, cryptocurrency investors, strategy developers, financial data analysts
This article analyzes the application effectiveness of the Anti-Martingale betting strategy in BTC trading. By importing BTC data for testing, it was found that benchmark signals performed poorly on 5m data, while the trend-following dual moving average strategy performed excellently on 15m, 30m, and 1h data. The Anti-Martingale betting strategy performed outstandingly on BTC 1h data, achieving a return rate of 9994.17%, far exceeding the benchmark strategy's 18.94%, with a smoother net value curve. The article points out the existence of favorable trend clustering phenomena, and the Anti-Martingale strategy can effectively utilize this characteristic to enhance returns. Meanwhile, through an extreme parameter case (1024x TP), it emphasizes the importance of signal strategy quality. Finally, it proposes future research directions based on the three-body dynamics hypothesis to design gating mechanisms.
- ✨ The Anti-Martingale betting strategy achieved a return rate of 9994.17% on BTC 1h data, significantly outperforming the benchmark strategy.
- ✨ Favorable trend clustering phenomena exist, and the Anti-Martingale strategy can effectively utilize this characteristic to enhance overall returns.
- ✨ Signal strategy quality has a decisive impact on the effectiveness of the Anti-Martingale betting strategy; extreme parameters may lead to net value resetting to zero.
- ✨ Future optimizations could involve designing gating mechanisms based on the three-body dynamics hypothesis to improve the periodic performance of trend-following strategies.
📅 2026-02-10 · 498 words · ~3 min read
Three-Body Dynamics Signal Gating Mechanism and Market State Variable Analysis
Quantitative Finance
👤 Quantitative trading researchers, strategy developers, analysts interested in market dynamics and signal gating mechanisms
This paper first introduces a signal gating mechanism based on three-body dynamics, which determines strategy entry and exit timing by estimating market state variables δ (premium), μ (momentum), and σ (volatility) to maximize strategy returns. The author elaborates on the intuitive understanding of these three state variables: the core of δ is the psychological anchoring effect, which can be analyzed through volume distribution; the core of μ is the speed of price changes, measurable by moving averages of log returns; the core of σ is the magnitude of price changes, measurable by the standard deviation of log returns. The article also discusses criteria for judging the effectiveness of estimation methods, i.e., evaluating based on the quality of gating effects, and notes that advanced signal strategies often already include estimates of these state variables but require systematic understanding. Finally, the author suggests that after decoupling signal gating, these state variables can serve as key factors, while the signal strategy itself may only need the simplest form.
- ✨ Proposes a signal gating mechanism based on three-body dynamics, dynamically adjusting strategy entry by estimating δ, μ, and σ
- ✨ Explains in detail the intuitive understanding and estimation methods of market state variables δ, μ, and σ, emphasizing psychological anchoring, price speed, and magnitude
- ✨ Discusses criteria for judging estimation effectiveness, i.e., the improvement in strategy returns due to gating effects
- ✨ Points out that advanced signal strategies already include estimates of market state variables but require systematic understanding
- ✨ Suggests that after decoupling gating, state variables can serve as key factors, simplifying signal strategy design
📅 2026-02-10 · 1,208 words · ~6 min read
Market State Variable Modeling Scheme for Three-Body Gating
Quantitative Finance
👤 Financial quantitative analysts, market researchers, and technical personnel interested in financial market modeling and gating mechanisms
Building on the three-body dynamics hypothesis and gating mechanism concept, this paper systematically outlines the modeling scheme for market state variables δ (premium), μ (momentum), and σ (volatility). The core innovation lies in the definition of δ: through the volume gravitational field model, nonlinear operations (Gaussian kernel functions and gradient calculations) are introduced to maintain its independence from μ and σ. μ is defined as the exponential moving average of returns to extract trend information; σ is defined as the standard deviation of returns to measure volatility amplitude; δ is based on the distribution of volume along the price axis, calculating the regression force when prices deviate from high-volume concentration areas. The article details the specific steps for calculating these three variables from candlestick sequences, including parameter settings and independence arguments, providing a new modeling framework for financial market analysis.
- ✨ δ (premium) is defined through the volume gravitational field model, introducing nonlinear operations to ensure independence from μ (momentum)
- ✨ μ is defined as the exponential moving average of returns, and σ is defined as the standard deviation of returns
- ✨ Specific steps and parameter recommendations for calculating δ, μ, and σ from candlestick sequences
- ✨ Independence arguments for the three variables (δ, μ, σ) are based on nonlinear operations and different information sources
- ✨ Kernel functions (e.g., Gaussian kernel) model psychological anchoring effects, with bandwidth adaptable to volatility
📅 2026-02-10 · 1,581 words · ~8 min read
Prediction Market Arbitrage Project Launch and Technology Selection
Quantitative Finance
👤 Readers interested in prediction markets, high-frequency trading, Rust programming, or technical project development
This article describes the prediction market arbitrage project launched on February 8, 2026, which falls under the high-frequency trading (HFT) category and has extremely high requirements for execution efficiency. The technology selection decision is to use Rust language to build a low-latency trading execution system to cope with the rapid elimination of arbitrage opportunities. The team's current technology stack is limited, and they plan to advance the project through vibe coding, taking this opportunity to deeply learn the Rust ecosystem and toolchain in preparation for future projects. The article also mentions that the team previously had basic Rust experience with Solana smart contracts but not in-depth, and they look forward to embracing challenges through this project.
- ✨ Prediction market arbitrage project launch, belonging to the high-frequency trading (HFT) category
- ✨ Technology selection adopts Rust language to achieve a low-latency trading execution system
- ✨ Team technology stack is limited, planning to advance the project through vibe coding
- ✨ Take this opportunity to learn the Rust ecosystem and toolchain for future preparation
- ✨ The project has high requirements for execution efficiency to quickly capture arbitrage opportunities
📅 2026-02-08 · 178 words · ~1 min read
Derivation of SDE Equations for Three-Body Dynamics in Capital Markets
Quantitative Finance
👤 Financial modeling researchers, quantitative analysts, economists interested in capital market dynamics
Building on the article 'The Three-Body Dynamics Hypothesis in Capital Markets,' this paper derives a complete system of stochastic differential equations (SDEs) to describe the interactions among momentum capital (M), value capital (V), and liquidity capital (L) in capital markets. The article defines fast variables (such as log premium, momentum, volatility) and slow variables (the volumes of the three types of capital) and extracts 12 formalizable core constraints. Through a detailed analysis of the SDE equations, the article validates these constraints one by one, including positive feedback for M, negative feedback for V, directionless feedback for L, positive and negative feedback loops, payoff matrices, and crowding effects. All constraints are validated, indicating that this SDE system fully implements the qualitative mechanisms from the original article, such as volatility clustering, fat-tailed distributions, and chaotic behavior. The article also conducts phase analysis and statistical property validation, providing a foundation for subsequent numerical simulations, bifurcation analysis, and parameter calibration.
- ✨ Derived a complete SDE system describing the interactions among three types of capital
- ✨ Validated 12 core constraints, including positive/negative feedback and payoff matrices
- ✨ The system explains market characteristics such as volatility clustering and fat-tailed distributions
📅 2026-02-07 · 1,839 words · ~9 min read
AI-Assisted Capital Persistence Experiments and Community-Based Subjective Trading
Quantitative Finance
👤 Quantitative traders, AI application researchers, investment strategy developers, professionals interested in combining subjective and quantitative approaches
Based on experimental experiences from February 3, 2026, this article discusses AI's role in improving efficiency in capital persistence experiments, emphasizing that redesigning experiments can systematically evaluate signal strategies and betting strategies. The author argues that subjective trading should be limited to signal strategy design, avoiding interference with betting strategies to overcome irrational decision-making in human fund management. To address potential issues where humans might peek at betting accounts, a community-based trading model is proposed: multiple subjective traders provide signal strategies, a consolidated betting account is managed by programs, and profits are distributed based on contributions, balancing fairness and efficiency to encourage better signal strategy design and control risks.
- ✨ AI enhances the efficiency of capital persistence experiments, achieving a qualitative leap
- ✨ Redesigning experiments systematically evaluates signal strategies and betting strategies
- ✨ Subjective trading should only apply to signal strategies, avoiding interference with betting strategies
- ✨ Human fund management suffers from irrational decision-making, requiring programmatic execution
- ✨ Community-based trading prevents subjective traders from adjusting strategies based on betting accounts
📅 2026-02-03 · 883 words · ~4 min read
Capital Persistence Battle Experimental Design
Quantitative Finance
👤 Quantitative traders, capital management strategy researchers, investment professionals interested in Anti-Martingale strategies.
This article details the experimental design for the Capital Persistence Battle, with the core concept of using a benchmark account as a reference and a betting account that dynamically adjusts positions using an Anti-Martingale strategy. Key components include: time scale t as discrete market moments; the benchmark account trades with fixed positions to provide a cumulative profit and loss curve; the betting account calculates input cash flow C(t) and benchmark stop-loss amount StopLoss(t) based on benchmark performance, constructs a risk control line RiskLine(t) and risk capital VC(t), and determines position size via the formula Position(t) = floor(VC(t) / StopLoss(t)); it defines handling logic for profit-taking and stop-loss events, as well as trading suspension conditions during observation periods. The overall aim is to maximize the efficiency of risk capital utilization, achieving aggressive yet controlled betting.
- ✨ Time scale t is a discrete market moment used for all time series.
- ✨ The benchmark account trades with fixed positions, providing cumulative profit and loss BasePnL(t) as a reference.
- ✨ The betting account uses an Anti-Martingale strategy, dynamically adjusting positions based on benchmark performance.
- ✨ Input cash flow C(t) and benchmark stop-loss amount StopLoss(t) are calculated from the benchmark account's historical performance.
- ✨ The risk control line RiskLine(t) moves downward over time, ensuring unrealized profit and loss does not fall below this line.
📅 2026-02-02 · 883 words · ~4 min read
Three-Layer Structure and Experimental Design Reconstruction of Backtesting Systems
Quantitative Finance
👤 Quantitative investment researchers, backtesting system developers, investment strategy analysts
This paper proposes reconstructing the backtesting system into a three-layer structure: market sequences as investment objects, signal strategies as responses, and betting strategies as investment subjects. The author suggests splitting each experiment into combinations of these three components and emphasizes the need for a thorough revision of experimental design and evaluation systems. The new evaluation focus shifts from peak expectations to the frequency and distribution characteristics of profit-taking events, particularly the average time interval of profit-taking events given M_T, to provide more meaningful investment guidance.
- ✨ The backtesting system should be reconstructed into a three-layer structure of market sequences, signal strategies, and betting strategies
- ✨ Market sequences are investment objects, which can be generated from synthetic or historical data
- ✨ Signal strategies are responses to market sequences, producing buy/sell signals
- ✨ Betting strategies determine capital allocation and risk management, reflecting investment subject preferences
- ✨ Experimental design requires evaluating each signal strategy under all market sequences and betting strategies
📅 2026-02-01 · 571 words · ~3 min read
Experience with OpenClaw and Opus Models, and Capital Endurance Battle Experiment
Quantitative Finance
👤 Tech enthusiasts interested in AI tool deployment and model comparisons, as well as researchers or investors focused on quantitative trading strategies.
This article documents the author's experience using the OpenClaw AI tool on January 31, 2026, including the process of deploying it on an Alibaba Cloud ECS server and connecting it to a Feishu robot. The author notes that OpenClaw is more suitable for local deployment, as cloud servers are costly and functionality is limited by tool installations like browsers. The article compares the performance of MiniMax M2.1 and Opus models, concluding that Opus is significantly better for programming tasks. The author used OpenCode+Opus to complete the code for the Capital Endurance Battle experiment, open-sourced it on GitHub, and shared preliminary findings: in a GBM high-volatility market model, a mean reversion strategy combined with anti-Martingale money management can achieve exponential capital growth under transaction costs, while a trend-following strategy cannot, highlighting the advantage of high-win-rate strategies. The author states that further validation is needed and invites attention to the open-source project.
- ✨ The OpenClaw AI tool can be deployed on a cloud server and connected to a Feishu robot, but it's more suitable for use on a local idle machine.
- ✨ Cloud servers are expensive, and OpenClaw's functionality is limited by tool installations such as browsers.
- ✨ OpenClaw equipped with MiniMax M2.1 performed poorly, and the author considers this model weak.
- ✨ The Opus model is significantly more effective than MiniMax M2.1 for programming tasks and is praised as the SOTA model for coding.
- ✨ The author used their GitHub Copilot Opus quota to complete the Capital Endurance Battle experiment code and open-sourced it.
📅 2026-01-31 · 385 words · ~2 min read
EA Project Introduction: AI-Driven Priority Fund for Quantitative Trading
Quantitative Finance
👤 Investors interested in blockchain investments, quantitative trading, and stable returns, particularly those seeking low-risk, principal-protected priority fund participation.
EA (Earnby.AI) is a priority fund project deployed on the BSC chain, settled in USDC, offering stable returns to investors through AI-driven quantitative trading strategies. The project uses a priority/subordinated capital structure, where priority capital enjoys principal protection, and subordinated capital is borne by the project's own funds to assume risks. The management team consists of professionals in quantitative trading and blockchain, including 5 co-founders. The project offers floating returns, currently with an annualized yield of 12%, and investors can redeem at any time. Strategies include directional portfolio strategies and delta-neutral strategies, with historical performance showing a cumulative return of 39.22% and an annualized return of approximately 22%. The project has no management fees, flexible lock-up periods, and aims to provide low-risk, sustainable returns for investors.
- ✨ EA is a priority fund project deployed on the BSC chain, settled in USDC
- ✨ Utilizes AI-driven quantitative trading strategies, including directional portfolio and delta-neutral strategies
- ✨ Capital is divided into priority and subordinated tiers, with priority capital enjoying principal protection
- ✨ The management team consists of 5 professionals in quantitative trading and blockchain
- ✨ Offers floating returns, currently with an annualized yield of 12%, and investors can redeem at any time
📅 2025-11-01 · 1,525 words · ~7 min read
Full Spectrum Analysis: The Optimal Method for Information Monetization
Quantitative Finance
👤 Quantitative traders, investment strategy developers, financial engineers, and advanced investors interested in the Kelly Criterion and leverage optimization.
This article proposes Full Spectrum Analysis (FSA), an investment trading strategy framework optimized based on the Kelly Criterion. It first analyzes the limitations of the traditional Kelly formula in investment applications, such as lack of leverage and short-selling considerations, and liquidation timing issues. Then, FSA constructs a systematic trading decision model by defining outcome spaces, calculating optimal leverage and compound returns. The article elaborates on the mathematical principles of FSA, including the calculation of expected returns and compound returns, as well as the algorithm for solving optimal leverage using Newton's iteration method. Additionally, it introduces historical backtesting methods (e.g., Gross Profit Margin GPM calculation), considerations for live trading modules, and measures to address black swan events. The core advantage of FSA lies in its ability to utilize imperfect probability information to maximize long-term returns by optimizing leverage decisions, reducing the high requirements for information quality.
- ✨ Full Spectrum Analysis (FSA) is based on the Kelly Criterion, optimizing investment leverage to maximize compound growth rate
- ✨ Define outcome spaces, probability distributions, and returns to calculate optimal leverage and compound returns
- ✨ Use Newton's iteration method to solve for optimal leverage, handling feasible regions and convergence issues
- ✨ Introduce Gross Profit Margin (GPM) for historical backtesting to evaluate strategy profitability and capacity
- ✨ Incorporate symmetric black swan event probabilities to limit leverage and prevent abuse and extreme risks
📅 2025-08-10 · 2,810 words · ~13 min read