Hedging EA: locked positions, swap bleed, and why most hedging forex robots are grids in disguise

What a hedging EA does

Buy and sell on the same symbol at the same time: that is what a hedging EA does. MT4 allows it natively and MT5 allows it on hedging-mode accounts, which is why the category is so common among MetaTrader trading bots and nearly unknown elsewhere. The sales pitch is that the hedge "locks" a losing position so the loss cannot grow, after which the forex robot waits for a favourable moment to unwind one side and let the other recover.

There are honest uses. A pairs or basket EA that goes long one correlated pair and short another is hedging in the proper sense and belongs with mean reversion EA. But the typical retail hedging bot opens the opposite position after a loss of a certain size, then opens more positions in both directions as price moves, and closes clusters when the net is positive. That is a bidirectional grid, and everything in grid trading EA applies with the swap cost doubled.

The statistics you will see

On closed trades, a hedging EA looks like a grid: win rate 75–90%, small uniform wins, and rare large losses when a locked pair is finally unwound on the wrong side. The distinctive number is in Detailed stats and long vs short: both directions show similar trade counts on the same symbol over the same period, often with one direction profitable and the other losing by a similar amount. That near-cancellation is the signature.

The holding-time histogram is bimodal. Profitable legs close in the 1–4 hours range; the locked legs sit for days or weeks in the 1–7 days and over-7-days buckets. Trades per week is high. In Open positions, a viewer sees a buy and a sell on the same symbol, sometimes several of each, with the floating losses and gains roughly offsetting while the swap column grows more negative every day.

The balance curve is smooth. The equity curve is flat where a hedge is locked, because the locked pair neither gains nor loses, then jumps when the EA unwinds. Max DD (equity) in Glance is usually moderate rather than catastrophic, which is the one way a hedging EA looks safer than an open grid. The Z-score is strongly negative because closes happen in clusters.

A worked example: what a lock actually costs

The EA is long 1 lot EURUSD and the position is 50 pips underwater, a $500 floating loss. Instead of closing it, the EA sells 1 lot. The $500 loss is now locked: whatever price does, one leg gains what the other loses. The sales page calls this protection. Arithmetically, it is identical to closing the long at a $500 loss and having no position, with two differences. First, the account is paying swap on both legs; at a typical −$8 per night per lot on the long and −$3 on the short, the lock costs $11 a day, $330 a month, to preserve a loss that has already happened. Second, the margin for 2 lots is tied up.

To "recover", the EA must at some point remove one leg and be directionally exposed again, which is the exact risk the hedge supposedly removed. If it picks the wrong moment the $500 becomes $900. After 60 days of waiting and $660 of swap, the best case is that the trader has lost $500 plus $660 instead of $500. The lock does not reduce risk. It defers and taxes it.

Advantages

• For a genuine pairs or correlation strategy, simultaneous long and short positions are the point, and the net exposure can be far lower than either leg.
• A locked hedge does cap the floating loss at the moment of the lock, which can buy time through a volatile event.
• Equity drawdown on a hedged grid is often shallower than on a one-directional grid of the same size.

Disadvantages and failure modes

Swap bleed is constant and compounding. Margin is doubled for no net exposure, which pushes the account closer to a stop-out when the EA adds directional positions, and a weekend margin increase can force-close a locked pair at a loss the EA never intended to realise.

The structural failure is the unwind. A hedging EA has to un-hedge to make money, and it does so on a rule, typically when price has moved a set distance. That converts a locked loss into an open one at the worst time often enough that, over a year, the closed-trade stats show the grid pattern: many small wins, a few huge losses. The failure looks slower than a martingale's and ends in the same place.

What the AI review and Score tend to flag

The grid check catches most hedging EAs because legs are closed in same-minute clusters; expect strategy type Grid and the grid watch flag. No stop-loss is near-universal in this category since the hedge replaces the stop. The tail flag appears once an unwind has gone wrong. If the EA also increases lot size on the added legs, which many "recovery" hedging bots do, the martingale check fires first and the type becomes Martingale.

The Score often looks acceptable for longer than a plain grid's, because equity drawdown is masked by the lock. Risk control is still penalised once any unwind produces a real drawdown, and consistency is hurt by symbol concentration, since hedging EAs almost always run one pair.

Red flags when buying a hedging EA

• The words "lock", "recovery", "zone recovery" or "never lose a trade" in the description.
• Long and short trade counts on one symbol that are nearly equal with offsetting results in Detailed stats.
• Open positions showing a buy and a sell on the same symbol at the same lot size.
• A large negative swap total relative to net profit.
• A record that only covers a period when the pair ranged. Check the calendar against the pair's trending months.

How to check on an account page

Open positions first: same-symbol buy and sell at once is the confession. Detailed stats and long vs short for the offsetting pattern. By symbol to confirm it is one pair. Trade records with the Swap column exposed, summed against net profit. AI review for the grid score, stop-loss percentage and flags. Glance for the gap between Max DD (equity) and Max DD (balance), remembering that a lock flattens equity rather than recovering it. Activity distributions for the bimodal holding times. Risk Lab's Monte Carlo shuffles closed legs and cannot see the locks, so the chance of ruin ladder understates the risk; treat it as a floor. On deeper plans, Performance depth's deepest drawdowns table will show unusually long durations for modest depths, which is what a locked loss looks like over time. If you want the same high win rate without the structure, compare with an honest mean reversion EA, and read how to vet an EA before paying for anything in this category.