Multi Timeframe Data Grabber Pro
- Yardımcı programlar
-
Abdelhakim Kouachi
Most EAs and indicators are built on mathematical illusions: hidden martingale grids destined to purge accounts, over-fitted backtests crafted for sales screenshots, encrypted DLLs hiding vulnerabilities, and developers who vanish the moment live slippage hits. kouaAlgo exists because serious - Sürüm: 1.2
- Etkinleştirmeler: 5
Your model is only as good as your data — and in machine learning, data preparation is 80% of the work. The unglamorous, error-prone 80%: aligning timeframes, engineering features, building labels without leaking the future, flagging news events, measuring true spread. Most quant projects die here, quietly, before the first model ever trains.
Multi-TF Data Grabber PRO does that 80% for you in one click — and exports it as a clean, deterministic, ML-ready CSV. This is the professional pipeline behind the free Lite edition, extended into a full research dataset: 94 columns per file, engineered features, future labels with zero lookahead, news and spread context included.
**What you get — one file per symbol, 94 columns**
- **Up to 10 symbols** in a single run (or AUTO = your chart symbol).
- **7 base timeframes:** M1, M5, M15, M30, H1, H4, D1 — each exported with its higher-timeframe context blocks aligned bar-for-bar.
- **7 built-in indicators:** MA(20), RSI(14), MACD(12,26,9), ATR(14), Bollinger(20, 2.0), Stochastic(14,3,3), ADX(14) — plus **2 custom slots**: point PRO at ANY indicator you own and its buffers are exported alongside everything else.
- **Engineered features:** log-returns (1 and 5 bars), candle body/wick ratios, ATR volatility regime, range/ATR.
- **Time features:** hour, weekday, trading session (asian/london/newyork).
- **High-impact news flag** pulled from the MT5 economic calendar.
- **True spread** measured from real ticks — not the broker's advertised fiction.
- **Future labels (1–50 bars ahead):** next direction, future return, future high/low — supervised-learning targets constructed with zero lookahead. This is the part almost everyone gets wrong; here it's done right and independently verifiable.
- **Train/test split date** and **append mode** for growing datasets without re-exporting history.
- **One-click ML-Ready preset:** a single toggle forces the proven research bundle — no configuration needed.
**Deterministic by design**
Closed bars only. Frozen export range. The same export produces byte-identical numbers every time. Every label in the file can be recomputed from the file's own closes — we ship an independent Python validator so you can prove it to yourself before you trust a single row.
**How to run it**
1. Drag DataGrabber_PRO_v102 onto any chart.
2. Set your symbols (up to 10, comma-separated, or AUTO), base timeframe, and date range — or enable the ML-Ready preset and change nothing.
3. Press OK. One CSV per symbol appears in MQL5 → Files → DataGrabber, named SYMBOL_TIMEFRAME_YYYYMMDD_HHMMSS.csv.
4. Open in Python/pandas, Excel, or any ML stack.
**Patience is part of professionalism**
A full ML-Ready export typically takes **10–13 minutes**. Why so long? Because it isn't skimming — it's computing 94 columns across thousands of bars: 7 indicators on multiple timeframes, engineered features, news flags, true spread from real ticks, and future labels. That is the weight of real research data, and it's exactly why the free alternatives feel instant: they give you almost nothing.
Leave it running until the dashboard shows **DONE** with its visual and audio alert — then your files are waiting in MQL5 → Files → DataGrabber, named SYMBOL_TIMEFRAME_YYYYMMDD_HHMMSS.csv. Thirteen minutes once — versus the months you'd spend discovering your shortcut data was poisoned.
**From CSV to model in one line**
The files load directly, no cleaning required:
`pd.read_csv("XAUUSD_m15_20260102_233400.csv", sep=";")`
94 ready columns, train/test split included, labels verified. Your model starts learning in minutes — not next weekend.
**Who this is for**
Quant researchers, ML practitioners, and serious system traders building predictive models on MT5 data. If you just need clean candles and core indicators, the free Lite edition on CodeBase is enough — PRO is for the moment your research outgrows it.
**Who this is NOT for**
This is a data tool, not a trading system. It places no trades, generates no signals, and makes no promise about your model's profitability. What it promises is clean, honest inputs — because garbage in means garbage out, no matter how good your network is.
**Why PRO — and not the free exporters?**
Free exporters do one thing: dump raw candles into a CSV. That's where their job ends — and where 80% of your ML work begins.
- Raw candles → CSV: both do it. PRO adds 10 symbols and 7 timeframes per run.
- Indicator values: free exporters rarely include them. PRO ships 7 built-in + 2 custom slots for ANY indicator you own.
- Engineered ML features: 94 columns — log-returns, body/wick ratios, volatility regime, session flags. Free exporters: none.
- Future labels 1–50 bars ahead with **zero lookahead** — independently verifiable. Free exporters: none.
- High-impact news flags from the MT5 economic calendar. Free exporters: none.
- True spread measured from real ticks. Free exporters: none.
- Deterministic, byte-identical exports, with an independent Python validator. Free exporters: none.
- One-click ML-Ready preset. Free exporters: none.
**The only data exporter on the Market that delivers all of this precise data from its true sources** — 94 engineered columns, future labels with zero lookahead, news flags and true spread in one deterministic export.
(This claim is grounded in our October 2026 scan of the 10 closest competing exporters, none of which combine these capabilities.)
*For reference: the highest-rated free data exporter in this category has not been updated since 2021.*
**Launch pricing**
$60 for the first 50 buyers — then $79, and finally $100 as each tier sells out. Rent available at $50/month. Early buyers lock the launch price forever.
---
*No trading, no signals, no DLL. Pure data — it cannot touch your positions. Version history: v1.2 — ML-Ready preset, custom indicator slots, news/spread/labels pipeline.*
