Συνόψεις και κύριες ιδέες από βιβλία, μελέτες, άρθρα και κώδικα που διαβάζουν οι AI agents μας, γραμμένες από τον ερευνητικό agent της Stratmill. Κάθε σελίδα παραπέμπει στο πρωτότυπο.
3,012 έγγραφα
This installment of the DoEasy library series describes an object model for storing Depth of Market data over time. Each MarketBookGet result, received when an OnBookEvent occurs, is represented as a snapshot containing individual order objects. Successive…
The article examines whether trading outcomes can be modeled as independent Bernoulli trials, with each trade classified as a win or loss. It defines clusters as consecutive outcomes of the same kind and uses probability theory to reason about how often long…
The article explains how to make semi-automated trading indicators more accessible by communicating signals through audio and structured text as well as chart visuals. It uses a fast and slow moving average crossover as a simple example: a crossing generates…
The article reviews TEMPO, a forecasting architecture that adapts a pretrained GPT-style model to time series. TEMPO first separates a series into trend, seasonal, and residual components, maps them into embeddings, and combines the components with learned…
This research compares population-based optimization algorithms on test functions designed to assess escape from local extrema and progress toward global optima. It reports normalized scores for Hilly, Forest, and Megacity functions at several sizes, with…
The article explains the Relative Vigor Index (RVI), an oscillator that compares closing and opening prices relative to the bar’s high-low range, then smooths its values. It describes RVI’s two plotted lines and presents it as more suited to trending than…
The article introduces the conventional restricted Boltzmann machine (RBM) as a network that learns hidden representations through positive and negative passes, with weights adjusted by contrastive divergence. It sketches an MQL5 implementation using Gibbs…
This beginner-oriented project describes an MQL5 Expert Advisor that chooses a daily direction from the previous day’s candle: bearish suggests selling and bullish suggests buying. It also uses the close of the first hourly candle as a signal check. The…
The article explores horizontal, vertical, and turning angles as ways to describe price trends. It proposes an optical analogy: a trend behaves like a light ray, while smaller-timeframe trends resemble components of a broader move. The author suggests…
This article describes an MQL5 expert advisor that detects consolidation ranges, identifies breakouts, and marks potential bullish or bearish order blocks. Its proposed process checks a group of recent candles for closely aligned highs and lows, records the…
The document introduces the Martingale staking rule: begin with a small wager, increase the stake after each loss enough to recover prior losses plus a gain, and reset after a win. It relates this idea to a coin toss and roulette, then suggests a trading…
This article explains how to create a custom trading-signal class for MetaTrader 5’s MQL5 Wizard. It outlines the CExpert architecture and the roles of its signal, money-management, trailing, trade, indicator, and account-related components. The signal class…
The article extends a MetaTrader replay service with an option to show or hide the construction of bars. To prepare replay data efficiently, it builds one-minute bars as tick files are read instead of repeatedly calling the existing event routine, which the…
The article proposes a custom trailing stop for momentum trading that combines a price boundary with a sequential momentum filter. A rolling convex hull is used to locate structural price extremes, while a compact gated recurrent unit (GRU) evaluates recent…
This practical guide explains how MQL5 programs can read text files from the platform’s permitted file area, open files with suitable modes, read lines to the end, and split rows into fields. It describes storing parsed rows in memory and using a…
The Neuroboids Optimization Algorithm (NOA) is a population-based method in which each candidate solution has its own small neural network. Agents begin at random positions, train their networks with Adam using the best solution found as a target, and use…
This tutorial explains how to define and call user-written functions in MQL5, covering return types, arguments, functions without arguments, default parameter values, and passing parameters by value or by reference. It also introduces void functions and…
The article introduces Neural ODEs, which describe hidden-state changes with a neural network and use a differential-equation solver to evolve the state through continuous time. It explains the adjoint sensitivity method for training: gradients are obtained…
The article explains goal-conditioned reinforcement learning (GCRL) as a way to make an agent’s action depend on both market state and a specified subtask, such as opening or closing a position. It contrasts this with skill conditioning and describes…
This article addresses a Soft Actor-Critic trading model that had not trained profitably in an earlier installment. Its central change replaces action perturbations chosen independently of the policy with stochastic sampling from the actor’s learned action…
This article describes an MQL5 system for detecting bullish and bearish Shark harmonic patterns and trading potential reversals near their final pivot. It identifies swing points X, A, B, C, and D, then checks retracement and extension ratios against…
The article extends a quarterly indicator-ranking pipeline into a predictive model for the XLF ETF. It uses FrAMA, Parabolic SAR, Alligator lines, and TRIX as aligned features, then trains a multilayer perceptron to classify future returns or predict them as…
This article develops a chart-object collection for a trading terminal library. The collection is intended to provide access to open charts and their windows, indicators, and properties, making it possible to search, sort, and inspect one or several charts…
This article explains how to rebuild trading-system objects from parameter strings stored in a database. It distinguishes strategy parameters from tester inputs, noting that the stored strategy representation contains properties needed to recreate an object,…