MetaTrader 4: The Hidden Reality of Automated Trading

MetaTrader 4: The Workshop Where Money-Losing Robots Are Built

A user opens a thread to share MetaTrader 4 and MQL4 resources. What starts as a link exchange turns into a reckoning with the automated trading industry: hundreds of Expert Advisors tested, worthless backtests, and a handful of ideas—slippage as an indicator, net volume, neural networks—that point to where the real game lies.

The first conclusion of the thread is a warning: according to a participant who claims to have tested hundreds of EAs on demo and micro accounts, the vast majority, including the most famous commercial ones, are clear losers. Backtesting, he adds, is of little use: it can rule out a system if its behavior is clearly deficient, but you never know how it will perform in the future. The uncomfortable takeaway is that learning and experimenting are the only valid paths.

Slippage Is Not Just a Pip Thief

The thread dedicates much of its development to an idea rarely found in manuals: slippage—the difference between the price you request and the price you get—is not just a cost. According to arguments made in the thread, it is the only variable that anticipates price evolution and shows the strength of the current trend. Low volume with high slippage would possibly indicate weak hands; low or zero slippage indicates strong hands.

The proposed calculation is artisanal: take the price difference between ticks, divide by that tick's volume, and obtain slippage in points. To average it over a time range, the system clock is used. The author acknowledges that the only way to know it exactly would be to execute an order, with its commissions and spread, and that the uncertainty principle applies here too: to know the system, you must interact with it.

There is a surprising nuance: the smaller the lot, the higher the slippage, according to the experience shared. Micro lots are penalized because they must be bundled to enter the market, consume latency, and are assigned higher transaction costs. The optimal size, it is claimed, is one lot. Although no one would enter the market with an untested strategy using micro lots.

Dow Theory, Rewritten with Volume and Transaction Costs

The thread attempts to translate the classic trend phases—accumulation, public participation, excess, distribution, panic—into two measurable variables: slippage and net volume. The proposal is that during accumulation, slippage decreases while volume increases; during public participation, slippage rises and volume spikes; during excess, both fall. The discussion shifts to the time factor: a sixty-second window attacks minute trends; a fifteen-minute window attacks hourly trends; a battery of ticks covering full hours attacks daily trends.

Net volume is not obtained with a single function. You must invoke tick-by-tick volume reading and calculate the difference between the previous and current values. If the difference is negative, the previous invocation corresponded to the prior candle. It is watchmaker work, not click-and-done.

The Terminal's Dark Side: Plugins That Hunt Stops

One of the thread's most uncomfortable contributions is the reminder that MetaTrader 4 is designed to sell to brokers, not traders. The Virtual Dealer plugin is mentioned, which allows brokers to create spikes at any moment or hunt stops at will. It is not claimed that all do this, but it is emphasized that it exists and is there. Another participant's response is pragmatic: that is why they use virtual stops executed from code rather than the stop loss field of orders.

The platform's true power, it is concluded, is its developer community. Knowing pros and cons allows for improved operations. Libraries for listing all server symbols and dumping them to a file are also cited, along with pages antiestéticaturing Ehlers indicators or koncorde to detect strong hand movements.

Neural Networks and Tick Analysis: The Next Level

The thread does not stick to classic indicators. Articles on price prediction with neural networks in MQL4 are linked, and the question is raised whether anyone has experience with genetic algorithms to optimize them. The conclusion drawn is that Japanese candles are a simplification of volume analysis, useful but limited, and that what is interesting is tick analysis.

The closing note is a disorienting fact: 90% of volume comes from strong hands, as stated in the thread, and the rest can be considered weak hands. To coordinate movements and avoid chaos, these strong hands would use common pivot levels, wait for the pot to fill with retail traders, and exit in the opposite direction breaking stops. The footprint to detect them would be in the elapsed time, in milliseconds, since the last tick: more time means more pips per contract. Specialists exist with times under a millisecond. Meanwhile, the retail trader keeps looking at a moving average.