Indicators MT4 MT5 – Welcome to the MT4 & MT5 Indicators Section on Remind.cz

At Remind.cz, indicators are built for decisions, not for hype. Each tool aims to reduce noise, highlight structure, and make your next step more deliberate. Whether you’re new or experienced, the goal is simple: clearer context, disciplined execution, and control under uncertainty.

Below is a short map of the main instruments. They are not black boxes: each one is testable, transparent in behavior, and designed to work with your own logic.

RTS5Pattern Profi v9.01 (EN)

Updated v9.01 with Confidence Engine. A pattern-matching workflow for MetaTrader: click to mark a structure, and the indicator searches history for similar formations, calculates the statistical win-rate, and visualizes the weighted projection. Use it for context and probabilities—not as an autopilot.

Fourier-PatternStat Elite — MT5 Indicator (EN)

Fourier analysis for market cycles. It decomposes price movement into components so recurring rhythm becomes visible. Useful when you want turning points and cadence shown explicitly—not guessed.

RTS AR WD (EN)

A context indicator built on autoregression and participation/breadth logic. It focuses less on isolated signals and more on where pressure builds, where it releases, and where the market has structural support.

RFNN — MT5 (CZ)

A neural indicator trained by backpropagation to capture non-obvious relations. Forecasts are drawn directly on the chart (frames, arrows, paths), so you can compare model output with real price action in context. More details

RemiFour (CZ)

A multi-method view designed for confluence. Instead of betting on one indicator, it helps you anchor decisions where multiple methods agree.

---

NNMQL5 — a compact engine that learns on the chart

NNMQL5 is my lightweight neural core for MetaTrader. It is built to learn from your instrument’s history directly on the chart, in real time. No server dependency, no opaque shortcuts—just deterministic layers, safe defaults, and full control.

• Explicit architecture: dense layers with clear activations (e.g., TANH for normalized inputs; linear output for regression). What you configure is what runs.
• On-chart learning: mini-batch training with shuffling, warm-up passes, and live redraws so you see behavior while it learns.
• Stability guards: learning-rate auto-reduce on spikes, basic explosion checks, and display clamps—so a single outlier doesn’t destroy readability.
• Reproducible workflow: fixed preprocessing over an explicit window and deterministic splitting by bar indices.

NNMQL5 is not a replacement for judgment. It’s a tool to test hypotheses faster and see structure sooner.

NNMQL5 Predictor (MT5) — future as a path, not a single point

NNMQL5_Predictor_future_fix.mq5 learns a mapping from the last Lookback closes to the next steps ahead, then draws the forecast as a continuous path into the empty right side of the chart. The point is not certainty—it’s a visible thesis you can evaluate.

• Multi-step forecast: recursive rollout (+N bars) drawn as clean segments so you can read the trajectory.
• Safe training loop: mini-batches, shuffling, conservative defaults, LR auto-reduce on MSE spikes.
• Transparent preprocessing: mean/variance normalization on the training slice, with display clamps that protect readability.
• Designed for evaluation: historical “Pred +1…+5” lines stay for context; the future path shows the current hypothesis.

Practical defaults: AnchorShift=1 (learn on closed bars), HiddenAct=TANH, start with LR=0.001 and BatchSize=32. Keep FuturePts long enough to see structure, short enough to act.

---

All tools on Remind.cz share the same goal: measurable clarity. They are not commands; they are testable instruments. Visit each product page for documentation, examples, and live demos.

---

Author: Tomáš Bělák