EA Health Monitor (MetaTrader 5 and MetaTrader 4)

A free indicator that shows one table row per Expert Advisor on the account, found by magic number, plus a row for manual trades and a total: closed trades, net result, win rate, profit factor, max drawdown of the realized result, today's result, open positions and their floating P/L. It reads the account history and the open positions only; it never trades. Source for both editions, a Python reference of the statistics, and the tests that were used to check that MT5, MT4 and Python agree.

Website: allanninal.dev/mql5 (utilities) · MT4 editions · allanninal.dev · Contact: message me on LinkedIn.

It reports what the account did. It does not judge whether an EA will work. Nothing here is a profit claim (see Honest limits).

Get it on the MQL5 Market

Edition	Platform	Price	Link
EA Health Monitor	MT5	free	https://www.mql5.com/en/market/product/201797
EA Health Monitor MT4	MT4	free	https://www.mql5.com/en/market/product/202428

The Market builds are the same calculation as the source in this repository.

Contents

1. What it is and who it is for
2. Why it was built
3. How it works
4. How it was tested
5. Architecture
6. Inputs, install and compile
7. Run the Python reference and the tests
8. Honest limits
9. Licence

What it is and who it is for

What: a panel of chart objects, refreshed every two seconds. Each row is one magic number (one Expert Advisor), in the order each first traded; magic 0 is "Manual"; the last row, "All", adds them up. Columns: Trades, Net, Win %, PF, Max DD, Today, Open, Floating, Last deal. Hover a cell for gross profit, gross loss, wins, losses and expectancy per trade. A manual close of an EA's trade stays with that EA.

For: traders who run several Expert Advisors on one account and want to see which one is doing what, not just one combined statement. Also for developers: the statistics are a small, tested core you can read.

Not for: judging an EA from a few trades, or anyone who wants alarms, period filters or curves.

Image notes: the screenshots use an example account with made-up data (five magic numbers, names typed in the Names input); they are illustrations, not terminal captures, and not trading results. MT4 versions are in docs/images/mt4/.

Why it was built

• A combined statement hides which EA carries the others. A per-magic table answers it at a glance.
• Weak, scattered demand, stated plainly. Our October 2026 research found a Myfxbook reply listing checks people skip for bots (open-basket floating loss, restart during an open trade), a Programming-forum thread on EA portfolio analysis (235 views) and a handful of equity and drawdown panels on the Market. That is a small signal, not proof of mass demand.
• So it is free and its value is that the numbers are checked. The statistics follow common definitions and are compared with an independent Python version on thousands of deals.

mindmap
  root((EA Health Monitor))
    Problem
      One statement for many EAs
      Manual closes blur whose trade it was
      Partial closes counted as several trades
    Design rules
      A trade belongs to the EA that opened it
      Partial closes are one trade
      Costs included, entry commission too
    Proof
      Python reference
      Synthetic histories compared exactly
      MT4 tester with real orders

How it works

The rules, step by step

For each deal, in time order (MT4: orders turned into deals, see below):

1. Position and owner. An opening (IN) deal creates a position owned by the deal's magic number. Every later deal of that position (partial closes, the final close, even a manual close with magic 0) belongs to the same owner. If the opening deal is not in the history, the first deal seen gives the owner.
2. Realized curve. Each deal adds profit + swap + commission + fee to its owner's curve at its time (the commission charged at opening included).
3. Max drawdown. peak = max(peak, curve), maxDD = max(maxDD, peak - curve); the peak starts at 0.
4. A trade is a position whose closed volume has reached its opened volume. Its net is the sum over all its deals. A reversing deal (IN/OUT) closes the trade and opens the next one with the volume beyond.
5. Statistics per owner. Trades N; wins W (net > 0) and losses L (net < 0); gross profit GP and gross loss GL;
   • win rate = W / N
   • profit factor = GP / GL (shown as "no loss" when GL = 0 and GP > 0)
   • expectancy = (GP - GL) / N.
6. Today is the owner's realized result since server midnight. Open and Floating come from the open positions (floating = profit + swap; MT4 adds commission, which MT4 books at the close).
7. Names. The Names input (magic=name,...), else the comment of the EA's first opening deal plus the magic, else #magic.

MT4 keeps orders, not deals. Each order becomes an opening deal (open time, lots) and, once closed, a closing deal (close time, lots, profit, swap, commission). A partial close leaves the rest open under a new ticket whose comment is from #<ticket>; all parts are one position. The book is rebuilt whenever the orders change.

Per-deal algorithm

flowchart TD
    A[Next deal in time order] --> B{Opening deal?}
    B -- yes --> C[Create or extend the position<br/>owner = deal's magic]
    B -- no --> D{Position known?}
    D -- no --> E[Owner = this deal's magic]
    D -- yes --> F[Owner = the position's magic]
    C --> G[Add profit + swap + commission + fee<br/>to the owner's curve]
    E --> G
    F --> G
    G --> H[Update peak and max drawdown]
    H --> I{Closed volume reached opened volume?}
    I -- no --> Z[Next deal]
    I -- yes --> J[One trade: net = sum of its deals<br/>win, loss, gross profit or loss]
    J --> K[Forget the position]
    K --> Z

A position's life

stateDiagram-v2
    [*] --> Open: opening deal (owner = its magic)
    Open --> Open: partial close (same owner, still one trade)
    Open --> Closed: closed volume reaches opened volume
    Open --> Reversed: IN/OUT deal
    Reversed --> Open: the volume beyond opens the other way
    Reversed --> Closed
    Closed --> [*]: counted as one trade, then forgotten

Refresh flow

sequenceDiagram
    participant T as Terminal
    participant I as EA Health Monitor
    participant B as Book (statistics per magic)
    participant P as Panel
    T->>I: timer every 2 s (and on ticks)
    I->>I: new deals since the last refresh? new day?
    alt a new day, or older deals synced in
        I->>B: start the book again from the whole history
    else only new deals
        I->>B: add them
    end
    I->>B: open positions: count and floating per magic
    B-->>I: rows and total
    I->>P: update only the cells whose text changed

How it was tested

1. The core against Python, on synthetic histories. tests/reference.py follows the rules operation by operation. tests/check_core.py writes seeded synthetic deal lists (hedging-style positions with partial closes, manual closes, reversals, entry commissions, some positions left open). The test script tests/mql5/Scripts/EAHealthDump runs them through the MT5 core in the MetaTrader 5 terminal and writes the statistics, and the checker compares every number (counts exactly, money within 1e-9 relative).

History	Deals	Magics	Trades	Result
seed 1	900	5	346	pass
seed 2 (2,000 positions)	4,496	5	1,817	pass
seed 3 (manual only)	905	1	355	pass
seed 4 (50 positions)	117	5	46	pass
seed 5 (two EAs)	1,799	2	708	pass
seed 6 (30 magics)	6,839	30	2,732	pass
Total	15,056		6,004	ALL PASS

Source: tests/results/core-parity-mt5.txt.

xychart-beta
    title "Deals and trades compared per synthetic history (all 0 differences)"
    x-axis ["seed 1", "seed 2", "seed 3", "seed 4", "seed 5", "seed 6"]
    y-axis "count" 0 --> 7500
    bar [900, 4496, 905, 117, 1799, 6839]
    line [346, 1817, 355, 46, 708, 2732]

Bars: deals. Line: trades.

2. The MT4 edition in the MetaTrader 4 Strategy Tester. tests/mt4/HealthTrader.mq4 trades a seeded random mix (several magic numbers and manual orders, comments, stop loss and take profit, partial closes, full closes, close-by) and, every few hundred bars and at the end, writes the raw orders and its own book. tests/mt4/check_mt4.py turns the raw orders into deals on its own and compares every statistic, open count, floating and row name with the Python reference.

	Result
Runs	9 (2 seeds x 4 symbols + one every-tick run)
Snapshots compared	146
Trade operations made by the test	25,220
Trades in the final snapshots	7,972
Result	ALL PASS

Source: tests/results/parity-mt4.txt.

flowchart LR
    R[reference.py<br/>Python, rules as code] --> C1
    G[check_core.py write<br/>seeded synthetic deals] --> D5[EAHealthDump script<br/>MT5 core, statistics to CSV]
    D5 --> C1[check_core.py check<br/>exact counts, 1e-9 money]
    T4[HealthTrader EA<br/>MT4 tester trades] --> O[raw orders + book per snapshot]
    O --> C2[check_mt4.py<br/>orders to deals on its own]
    R --> C2
    C1 --> V{all pass?}
    C2 --> V
    V -- yes --> M[Market validation]
    V -- no --> F[fix, rerun everything]

3. Market validation. Both Market builds passed the MQL5 Market's automated product validation before publication (the Market's own check, not part of this repository).

What the tests do not cover. The panel text and layout (the MT4 tester gives an indicator run through iCustom no chart, so only the book the panel shows is compared; the indicator is loaded in every run as a smoke test); the live history of a real account (only synthetic histories and tester trades were compared); the MT4 test account is a cent account, so many orders were refused for margin and some runs ended early.

Architecture

flowchart TB
    subgraph MT5["mql5"]
        A5[Indicators/EAHealth/EAHealthMonitor.mq5<br/>inputs, panel rows, timer]
        L5[Include/EAHealth/HealthLive.mqh<br/>refresh, names, formatting]
        C5[Include/EAHealth/HealthCore.mqh<br/>deals to statistics]
        P5[Include/EAHealth/Panel.mqh<br/>grid of chart objects]
        A5 --> L5
        A5 --> P5
        L5 --> C5
    end
    subgraph MT4["mql4/Indicators/EAHealth"]
        A4[EAHealthMonitor.mq4]
        L4[HealthLive.mqh<br/>orders-based refresh]
        C4[HealthCore.mqh<br/>same statistics + orders to deals]
        P4[Panel.mqh<br/>exact copy]
        A4 --> L4
        A4 --> P4
        L4 --> C4
    end
    subgraph TESTS[tests]
        R[reference.py]
        K1[check_core.py]
        D5[mql5/Scripts/EAHealthDump]
        K2[mt4/check_mt4.py]
        D4[mt4/HealthTrader.mq4]
    end
    D5 -. includes .-> C5
    D4 -. includes .-> L4
    K1 --> R
    K2 --> R

classDiagram
    class EAHealthMonitor {
        +InpNames string
        +InpManual bool
        +InpHideIdle bool
        +OnInit()
        +OnTimer()
        +Show()
    }
    class EHLive {
        +EHBook b
        +EHRefresh()
        +EHName()
    }
    class EHBook {
        +EHStat st[]
        +EHPos ps[]
    }
    class EHStat {
        +magic long
        +n, w, l int
        +gp, gl, net double
        +maxdd, today double
        +open int
        +floating double
    }
    class CSDPanel {
        +Setup()
        +Layout()
        +SetCell()
        +Destroy()
    }
    EAHealthMonitor --> EHLive
    EAHealthMonitor --> CSDPanel
    EHLive --> EHBook
    EHBook --> EHStat

The indicator has no buffers and no iCustom interface: it is a panel.

Inputs, install and compile

Input	Default	Meaning
Names: magic=name, comma separated	empty	e.g. 771500=Trend EA,0=Manual
Show manual trades (magic 0)	true	The "Manual" row
Hide EAs with nothing to show	false	Hide rows with no trade and nothing open
Panel corner, distance from the side and the top, font size	top left, 10, 20, 9	Position and size
Colours: panel, cell, text, header, positive, negative	dark theme	Appearance
Debug log for bug reports	false	Writes extra lines to the Experts tab, to attach when you report a problem

Install from source. MT5: copy mql5/Indicators/EAHealth/ into <terminal data folder>/MQL5/Indicators/ and mql5/Include/EAHealth/ into <data folder>/MQL5/Include/, open EAHealthMonitor.mq5 in MetaEditor and press Compile (F7). MT4: copy mql4/Indicators/EAHealth/ (it holds all four files) into <data folder>/MQL4/Indicators/ and compile EAHealthMonitor.mq4. Both compile with 0 errors and 0 warnings (checked with the MetaEditor of a MetaTrader 5 and a MetaTrader 4 build on a Windows machine; the compiled files are not in the repository). The debug log needs the Chart Kit file as well: copy mql5/Include/ChartKit/ (MT5) or mql4/Include/ChartKit/ (MT4) into the Include folder of the data folder.

Run the Python reference and the tests

python3 -m venv .venv && . .venv/bin/activate && pip install -r requirements.txt

import sys; sys.path.insert(0, "tests")
from reference import compute, synth
deals = synth(seed=1)                  # seeded random history, deals in time order
stats = compute(deals, day_start=0)    # per magic: n, w, l, gp, gl, net, curve, peak, maxdd, today, last, deals

To repeat the MT5 core check: python3 tests/check_core.py write <dir>, copy <dir>/deals_*.csv to the terminal's common Files\ehtest\ folder, compile and run tests/mql5/Scripts/EAHealthDump (it writes stats_<n>.csv there), copy those back to <dir> and run python3 tests/check_core.py check <dir>. For MT4: compile tests/mt4/HealthTrader.mq4, run it in the Strategy Tester (open prices) with the inputs in tests/mt4/configs.tsv, then python3 tests/mt4/check_mt4.py <folder with the files>.

Honest limits

• It reports history; it does not predict. A few trades prove little. No profit or accuracy claim is made.
• Max drawdown is of the realized result, from the first deal in the history, not of the account's equity. In the "All" row it is the largest of the rows, not the account's drawdown.
• History must be available. The panel can only count what the terminal's account history holds; if you cut the history range in the terminal, the numbers change.
• A bug found on the way. MQL4's StringSubstr(s, 0, 0) returns the whole string, so a comment that was only a server-close marker such as [sl] was read as a name; fixed in EHCleanComment (MT4 core) before the published test run.
• MT4 books the commission at the close, so an open trade's floating P/L includes it there; MT5 books it at the entry deal.
• Scope of the tests. Synthetic histories and tester trades, not a real account; the panel text is not covered in the MT4 tester.

Licence

Everything is © 2026 Allan Niñal under the PolyForm Noncommercial License 1.0.0: source-available, free for personal and other noncommercial use, not for commercial use. Each source file has an SPDX-License-Identifier line; the text is in LICENSE. The licence applies to this source, not to anything you build or trade with the indicator. Trading involves risk; the software is provided as is.

---

Website: allanninal.dev/mql5 · MT4 · allanninal.dev · Questions: message me on LinkedIn.