Diego Arribas Lopez
Diego Arribas Lopez
4.6 (14)
  • 정보
5+ 년도
경험
71
제품
257
데모 버전
0
작업
0
거래 신호
0
구독자
You've been burned before. That's why you're reading this.

You bought an EA with a perfect backtest. It collapsed in two weeks live. You tried another one — same story. Maybe you've done this five times already. You're not stupid. You were just shopping in a market full of smoke.

Most "AI-powered" bots on MQL5 are a martingale with a ChatGPT logo slapped on top. Perfect equity curves that only exist inside the strategy tester. That's not trading. That's marketing.

I'm Diego, and I build the opposite of that.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

WHO I AM

I design Expert Advisors and AI-driven trading systems at DoIt Trading. My focus: strategies that survive the market, survive drawdowns, and survive YOU — because most EAs don't die from bad logic. They die because the trader panics and pulls the plug.

Everything I sell runs on my own money first. No exceptions.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

TRADING PHILOSOPHY

Portfolio over single EA. One bot is not a trading plan — it's a psychological trap disguised as a strategy. Multiple strategies supporting each other is how you actually stay in the game.

Controlled risk, not zero risk. Every system has losing periods. The difference is whether drawdown is managed with clear stop-losses and defined limits, or whether it's a grid praying for a reversal.

Recovery is more dangerous than drawdown. The urge to "get back to break-even" destroys more accounts than the drawdown itself. My systems are built knowing that.

Proof over promises. I don't post screenshots — I post verified accounts:


. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

WHAT I BUILD

AI-driven EAs, portfolio trading systems, and decision-support tools. From single-pair specialists to multi-strategy portfolios designed for funded accounts.

Full product line + 70+ educational articles:
https://doittrading.com

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

BROKERS I USE WITH REAL MONEY

Don't run your EA on a random broker. These are the ones I trust with my own capital:

IC Markets — Raw spreads, institutional execution
https://doittrading.com/go/ic-markets/

Pepperstone — Solid regulation, global coverage
https://doittrading.com/go/pepperstone/

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

PROP FIRMS / FUNDED PROGRAMS

Axi Select — Scale capital without paying challenge fees. You trade, they add capital. The only funded program where the business model isn't "collect fees from failing traders"
https://doittrading.com/go/axi-select/

FTMO — If you want traditional funded challenges, this is the one with actual payouts
https://doittrading.com/go/ftmo/

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

FREE RESOURCES

Free USDJPY EA — Start building your portfolio at zero cost
https://doittrading.com/free-usdjpy-mt5-portfolio-module/

Weekly Newsletter — What's working, what's not, no hype
https://doittrading.com/newsletter/
Diego Arribas Lopez
Diego Arribas Lopez
MultiStrategy Pro through a 1,000-scenario stress test.

I stress-tested a popular gold EA for a video, and it wouldn't be fair to skip my own. So MultiStrategy Pro went through the same engine: seven strategies on gold, USDJPY and GBPJPY, the same period, a $4,000 reference account and $20 risked per trade, half a percent of the account.

The method: the history rebuilt 1,000 times in 10-day blocks with all strategies in sync, 10% of the trades skipped at random, and every result scaled between 0.85x and 1.15x.

The result: the historical drawdown was 7.46%. The 95th percentile came out at 10.41%, about 1.4 times the history. The gold EA I tested came out at about 1.7 times.

To be fair to the comparison: four of the seven strategies still trade gold, so the same questions apply here. Plan for the stressed number, not the curve.

Before anything else, test the demo in the Strategy Tester with your own account size.

DoIt MultiStrategy Pro: https://www.mql5.com/en/market/product/165426
Diego Arribas Lopez
Every few weeks a new model becomes "the best ever", and this week it's Google's turn with Gemini 4 Argon. If you run a Gemini trading bot , or any AI EA, you're probably asking the practical question: do I switch now, wait, or is my whole setup already outdated...
Diego Arribas Lopez
Diego Arribas Lopez
Start with one strategy, on demo, with only your evenings.

I have a day job and I'm not leaving it. Everything I make from trading is extra, for my family. I took years to trade real money at all: guru rules that didn't work, EAs that lost in their first two weeks, and finally the lesson that stuck. Not all your eggs in one basket, and never load all the baskets at once.

That is why this module exists. It is one XAUUSD H1 strategy from my own portfolio, DoIt MultiStrategy Pro, with the same strategy ID. No licence to activate.

If you only have evenings:

- Run it on demo for a few weeks and watch its bad days, not only the good ones.
- At 0.01 lots its average stop is around $18 and its worst historical losing streak around $138. Use $1,000 or more, or a cent account.
- Once a week, write down what it did. That record is the asset.

When it has earned your trust, the other six strategies are in the full portfolio.

DoIt MultiStrategy Pro: https://www.mql5.com/en/market/product/165426
Diego Arribas Lopez
Diego Arribas Lopez
Claude Opus 5.5 in Alpha Pulse: setup in ten minutes.

1. Create an Anthropic API key at console.anthropic.com and add prepaid credits under Billing.
2. In MetaTrader, allow WebRequest for https://api.anthropic.com.
3. Paste the key in the Claude API Key field and select Claude Opus 5.5 in AI Model Selection.
4. Set Reasoning Effort. Opus 5.5 always thinks; Disabled runs it at its default depth.
5. Set Max Daily API Cost. When the day's budget is spent, the EA stops asking. It never switches to a cheaper model behind your back.
6. Run it on demo. The Strategy Tester can't call the API, so this is a forward test.

Write the configuration down before you start, and don't change it mid-test.

Alpha Pulse AI for MT5: https://www.mql5.com/en/market/product/149360
Alpha Pulse AI for MT4: https://www.mql5.com/en/market/product/149361
Diego Arribas Lopez
Diego Arribas Lopez
Gemini 4 Argon: use it in Alpha Pulse the day it opens.

Google announced Gemini 4 Argon on September 30. For now it is in limited release, and Google says paid API customers come next. There is no API model ID published yet.

You don't need a new version to try it. When the ID appears in Google's API documentation:

1. Set AI Model Selection to Custom Model.
2. Paste the model ID into the Custom Model field. Alpha Pulse detects Google from "gemini" in the name.
3. Use your Google API key and keep generativelanguage.googleapis.com in your WebRequest list.
4. Run it on demo, next to your current model.

What I will record in the first week: failed requests, response time, responses the EA could not read, agreement with my current model, and API cost per day. Not profit. One week says nothing about an edge.

Gemini 3.1 Pro has been the steadiest model I have run live. If Argon keeps that and adds the reasoning the benchmarks claim, it is worth testing. A benchmark is not a trading result.

Alpha Pulse AI for MT5: https://www.mql5.com/en/market/product/149360
Alpha Pulse AI for MT4: https://www.mql5.com/en/market/product/149361
Diego Arribas Lopez
Gold EA drawdown belongs near the beginning of a buying decision. If you are looking at Gold Guardian because you want a specialised way to trade gold, you also need to understand what owning that strategy could feel like when the account is losing. A return figure can create interest...
Diego Arribas Lopez
Diego Arribas Lopez
v3.04: GPT-6.1 Sol, Claude Opus 5.5, GPT-6 Sol and Luna, Xiaomi MiMo, and Jev as an experiment. Free update for every owner.

Anthropic released Claude Opus 5.5 on Tuesday, September 22, and OpenAI added GPT-6 Sol, GPT-6 Luna and then GPT-6.1 Sol. All of them are in v3.04. Same prompt, same risk settings, same API key: you pick the new preset and that is it.

What is in v3.04:

- GPT-6.1 Sol is the new OpenAI default. If you use the OpenAI default, it moves to GPT-6.1 Sol automatically, at the same input and output price as GPT-6 Sol. GPT-6 Sol stays in the list.

- Claude Opus 5.5 replaces Claude Opus 5. Anthropic prices it below Opus 5. Its thinking is always on: your Reasoning Effort input sets how deep it thinks, and Disabled runs it at its default depth.

- GPT-6 Sol and GPT-6 Luna replace GPT-5.6 Sol, Terra and Luna. There is no GPT-6 Terra: GPT-6 Sol costs less than GPT-5.6 Terra did, so it covers that slot.

- Xiaomi MiMo V2.6 Pro and MiMo V2.6 Flash. A new provider. You need a MiMo pay-as-you-go API key (it starts with "sk-") and https://api.xiaomimimo.com in your WebRequest list.

- Jev by TypeSafe, experimental, signals only. Jev is not a chat model: it answers typed questions with probabilities. Alpha Pulse asks it whether to go long or short, which stop distance in ATR, which reward to risk and which market regime, then computes stop loss and take profit itself. It runs in signals-only mode: its decisions show on the panel and no trade is opened. It has no trading track record, mine included, so that is where it stays for now. Single chart only. You need a Jev API key from TypeSafe (early access) and https://api.typesafe.ai in your WebRequest list.

- Tested before release: Claude Opus 5.5, GPT-6 Luna and MiMo V2.6 Pro answered live requests on my setup.

- New models that are not in the list yet, like Gemini 4 Argon when Google opens it on the API: choose Custom Model and type the model ID exactly as the provider publishes it. The provider is detected from the name.

- If your settings use Claude Opus 5 or any GPT-5.6 preset: pick the model again after updating and save your set. Those presets are gone from the list. Nothing else was removed.

Update path: download v3.04 from your MQL5 purchases (MT4 and MT5), reload the chart, pick the preset. Nothing else changes.

MT5: https://www.mql5.com/en/market/product/149360
MT4: https://www.mql5.com/en/market/product/149361
Diego Arribas Lopez
Alpha Pulse AI makes most sense when you can finish this sentence: “I want to use it to develop and evaluate this trading process.” That gives the purchase a job beyond owning the latest model name...
Diego Arribas Lopez
Manual backtesting can turn an evening at the charts into useful practice. It can also turn into a collection of winners you recognised after seeing what happened next. The difference is often decided before the first simulated trade...
Diego Arribas Lopez
Diego Arribas Lopez
Gold Guardian: decide with the drawdown in view.

A gold EA can be interesting and still be the wrong fit for your account.

Before choosing Gold Guardian, open the current product page and public trading record. Examine the losing periods, open-equity risk and the size of losses alongside the wins.

Then translate the risk into the money you would actually put in the account. A technical minimum deposit does not tell you whether you can tolerate the strategy’s behaviour.

Gold Guardian is a specialised gold EA. Its position management includes controlled recovery logic, which does not guarantee that losses will be recovered or that future drawdown will stay within the historical record.

Start on demo, understand the settings and decide on your review process before increasing exposure.

The useful purchase is a tool that fits the process you are building. You should be able to explain why it belongs there, including how it can lose.

Gold Guardian for MT5:
https://www.mql5.com/en/market/product/132497

Gold Guardian for MT4:
https://www.mql5.com/en/market/product/132496
Diego Arribas Lopez
EA portfolio risk starts with a deceptively simple question: when the market moves against you, how many of those charts are making the same bet? You may have spent months collecting strategies that look different...
Diego Arribas Lopez
Diego Arribas Lopez
Before buying Alpha Pulse, give the purchase a job.

What process do you want to build with it?

Your own instructions, an included strategy engine, a documented evaluation of Astra, or a closer look at AI decisions inside MetaTrader: choose a concrete first project.

Alpha Pulse provides the model connection and execution workflow. You still define the method, configure risk and evaluate the behaviour. API usage is billed separately.

The published price schedule has its final step on 26 September. Check the live product page for the current price and your platform.

If the tool already fits a project you want to do, that date is relevant. The date alone is not a reason to buy it.

Save the first prompt and settings. Start on demo. Include losing trades, open equity and API costs in the review. That is how you make the purchase part of your development instead of another tool you never properly evaluate.

Alpha Pulse AI for MT5:
https://www.mql5.com/en/market/product/149360

Alpha Pulse AI for MT4:
https://www.mql5.com/en/market/product/149361
Diego Arribas Lopez
A free gold EA for MT5 can give you something useful to practise with. The first goal is to understand what you have installed and what the account can actually lose when it trades. DoIt Gold Portfolio Module gives you one complete XAUUSD H1 strategy...
Diego Arribas Lopez
Diego Arribas Lopez
The trade you skipped may be the one you needed to record.

Historical charts make a strategy look obvious once you can see the next move. Manual backtesting becomes more useful when you keep the decision separate from the outcome.

Choose the period before you start. Work through it in order. Record the entry, invalidation and reason for staying out before advancing the chart.

When a losing setup makes you want to change a rule, save the change as a new version. Keep the old result attached to the rule that actually produced it.

Backtesting Simulator helps you practise this workflow in MetaTrader with replay controls. The discipline is still yours: include the awkward examples, state your cost assumptions and keep uncertain cases visible.

The goal is a method you can explain and repeat. Historical practice does not guarantee live results.

Backtesting Simulator for MT5:
https://www.mql5.com/en/market/product/101345

Backtesting Simulator for MT4:
https://www.mql5.com/en/market/product/101346
Diego Arribas Lopez
Diego Arribas Lopez
v3.03: Claude Opus 5.5, GPT-6 Sol and Luna, and Jev as an experiment. Free update for every owner.

Anthropic released Claude Opus 5.5 on Tuesday, September 22, and OpenAI added GPT-6 Sol and GPT-6 Luna the same week. Both are in v3.03. Same prompt, same risk settings, same API key: you pick the new preset and that is it.

What is in v3.03:

1. Claude Opus 5.5 replaces Claude Opus 5. Anthropic prices it below Opus 5. Its thinking is always on: your Reasoning Effort input sets how deep it thinks, and Disabled runs it at its default depth.

2. GPT-6 Sol and GPT-6 Luna replace GPT-5.6 Sol, Terra and Luna. GPT-6 Sol is the new OpenAI default. There is no GPT-6 Terra: GPT-6 Sol costs less than GPT-5.6 Terra did, so it covers that slot.

3. Jev by TypeSafe, experimental, signals only. Jev is not a chat model: it answers typed questions with probabilities. Alpha Pulse asks it whether to go long or short, which stop distance in ATR, which reward to risk and which market regime, then computes stop loss and take profit itself. It runs in signals-only mode: its decisions show on the panel and no trade is opened. It has no trading track record, mine included, so that is where it stays for now. Single chart only. You need a Jev API key from TypeSafe (early access; signups were full when I released this) and https://api.typesafe.ai in your WebRequest list.

If your settings use Claude Opus 5 or any GPT-5.6 preset: pick the model again after updating and save your set. Those presets are gone from the list.
Update path: download v3.03 from your MQL5 purchases (MT4 and MT5), reload the chart, pick the preset. Nothing else changes.

Price: 897 USD until Friday, September 25. On Saturday, September 26 the price moves to 997 USD, the final step of the ladder published in July.

MT5: https://www.mql5.com/en/market/product/149360
MT4: https://www.mql5.com/en/market/product/149361
Diego Arribas Lopez
An AI trading forward test should leave you with more than a green number. You want to know which configuration ran, what decisions it made and whether you would understand the same behaviour during a losing period. That is the practical next step after choosing Astra in Alpha Pulse AI...
Diego Arribas Lopez
Diego Arribas Lopez
Four charts. How many different bets?

Two gold EAs can use different entries and still lose on the same move. Two currency strategies can share exposure to the dollar.

Before adding another chart, write down what is already open: instrument, direction, actual size and which positions can be active together.

Then look at the account’s equity, including floating losses. A collection of attractive closed-trade reports can miss the uncomfortable part that happens while positions overlap.

This is a useful step in building your own process: learn to explain what each component contributes before deciding you need more components.

If you want one strategy to observe first, the free DoIt Gold Portfolio Module gives you a complete XAUUSD H1 component. Start on a separate demo and check the scope of its account-level protection.

One component is not diversification. It is a practical place to begin learning what you are running.

Module and setup information:
https://doittrading.com/free-gold-ea/?utm_source=mql5&utm_medium=whats_new&utm_campaign=astra-development-2026-09&utm_content=p152

DoIt Gold Portfolio Module MT5:
https://www.mql5.com/en/market/product/194178

DoIt Gold Portfolio Module MT4:
https://www.mql5.com/en/market/product/194179
Diego Arribas Lopez
GPT-6 Astra trading prompts become useful when they describe a decision you can examine. You may recognise your setup instantly on a chart. The difficult part is explaining it clearly enough that you can tell whether an AI interpretation actually matches it. That gap is worth working on...
Diego Arribas Lopez
Diego Arribas Lopez
DoIt Gold Portfolio Module: your first week.

DoIt Gold Portfolio Module gives you a complete XAUUSD H1 strategy to examine on your own demo account. It is available for MT5 and MT4, with no time limit.

For the first week, make the job concrete:

Confirm the correct chart and timeframe.
Understand the selected sizing mode.
Check the actual order volume and stop distance.
Record open equity as well as closed trades.
Keep a note of anything you change.

The software is free. Gold exposure still carries risk, and the broker’s minimum lot size can make a small account unsuitable. Account-level protection can also affect other trades, so start with a separate demo setup.

A quiet week is not a reason to force activity. A profitable week is not proof of a durable edge. The first result you want is a setup you understand.

Get the module and setup information:
https://doittrading.com/free-gold-ea/?utm_source=mql5&utm_medium=whats_new&utm_campaign=astra-development-2026-09&utm_content=p151

DoIt Gold Portfolio Module MT5:
https://www.mql5.com/en/market/product/194178

DoIt Gold Portfolio Module MT4:
https://www.mql5.com/en/market/product/194179
Diego Arribas Lopez
You want your next year in trading to leave you with something you can build on. GPT-6 Astra in MT5 opens an interesting route: put your trading ideas into instructions, observe the decisions they produce, and develop a process you can actually review...