Your EA Isn't Broken. It's Working. (Why Losing Months Are the Point)

Originally published at doittrading.com

Three losing weeks in. Your finger is hovering over the off switch. The EA you were so sure about is bleeding, and a voice says: it's broken, kill it before it gets worse.

Stop. Before you touch anything — your EA losing money right now is not proof that it's broken. More often, it's proof that it's working exactly as designed. The problem isn't the bot. It's that you were sold a straight line, and real edges don't move in straight lines.

I walked through the whole logic on video — it's on the DoItTrading YouTube channel. Written version below.

You weren't sold a strategy — you were sold a straight line

Every ad you've ever seen showed a smooth, up-and-to-the-right equity curve. No dips. No dead months. That image is what you bought, even if it's not what you paid for. So when reality shows up — chop, drawdown, a week of red — it doesn't match the picture in your head, and your brain screams "malfunction." But the smooth curve was the lie. The bleeding is the truth. You're seeing a real system for the first time, without the marketing filter.

Is your EA losing money — or just working?

You don't guess — you check it against its own history.

Signs the loss is normal (leave it alone):

• The drawdown depth and duration are within what its backtest and live track already showed.
• It's following its rules — entries, exits, and position sizes look like they always have.
• It's quiet because setups don't qualify, not because it's erroring out.
• Other pairs or strategies are behaving normally — it's a strategy-specific cycle, not a system-wide failure.

Signs it might actually be broken (investigate):

• Drawdown has blown well past its historical worst, with no change in market regime to explain it.
• Risk per trade or total exposure has crept past the limits you set — a configuration or logic problem, not a market one.
• It's behaving erratically: missed exits, doubled orders, trades that ignore its own rules (often a data-feed, VPS, or broker-condition issue).

A bot losing money inside its known envelope is working. A bot breaking its own rules or risk limits is the real malfunction. One asks for patience; the other asks for a look under the hood.

Losing months are the cost of the edge, not a malfunction

Every bot, every strategy, every fund on earth has losing weeks and losing months. That's the unavoidable cost of having an edge at all — a statistical lean over hundreds of trades, not a guarantee on any single one. Sometimes "working" even looks like doing nothing: a bot that goes quiet is often correctly refusing setups that don't meet its rules. And during the drawdowns, the smartest move is usually to do nothing — exactly what humans are worst at and automation is built for.

The loop that actually empties your account

The real account-killer isn't any single strategy. It's the loop. You buy a bot. It works for a bit. Then it bleeds — and it will. You decide it's broken, switch it off, and hunt for the next one that's "actually" profitable. Repeat.

Every cycle, you eat the losing stretch and skip the recovery — because you always quit at the bottom and re-enter something new at its top. You're not unlucky. You've built a machine for buying high and selling low and calling it "finding the right system." The strategies aren't what's draining you. The switching is.

There is no best EA. Stop hunting the holy grail.

Before you go looking for THE BEST EA IN THE WORLD that prints money every day — stop. It doesn't exist. A robot that wins in every market condition, on every pair, forever, is the straight-line fantasy in a different costume. Most "low drawdown" EAs you'll see advertised are backtest fits, hidden martingale, or systems that haven't lived through a real volatility expansion yet. Chasing the one that never loses is how you stay stuck in the loop forever.

The fix banks use: a portfolio, not a hero

The fix is the boring one nobody puts on a thumbnail: risk management and diversification. Don't put all your eggs in one basket — one bot, one pair, one idea. Do it the way banks and institutions actually do it, with a portfolio: several strategies, tested over years, that work together. The point isn't that each one is perfect — it's that they don't all bleed at the same time. When one strategy is in its losing stretch, another is in its good one. The winners carry the losers, the swings flatten, and the account survives long enough for the edge to show up.

Survival is the whole game. A single "great" EA that blows up in its first bad quarter makes you zero. A portfolio of decent strategies that survives every quarter compounds for years.

Build it yourself — or use mine

You can build this yourself: start slow on demo, add tested strategies one at a time across different pairs and timeframes, and watch how they offset each other. What matters as you add each one isn't "is this amazing alone" — it's "does this bleed at a different time than what I already have." Uncorrelated strategies are what flatten the swings.

Or use the one I built for exactly this:

DoIt MultiStrategy Pro — eight tested strategies running as a single portfolio, one per pair and timeframe, with portfolio-wide risk control so a single losing stretch doesn't sink the account. It's not a money printer and I won't pretend it is; it's built to survive the bad months that kill single-EA accounts. The live track (drawdown and quiet stretches included, because those are the point) is linked on the product page.

→ See DoIt MultiStrategy Pro →
Portfolio system · one strategy per pair/timeframe · portfolio-wide risk cap

The mindset is the same either way: stop grading individual bots by their worst week. Grade the portfolio by whether it survives the year.

FAQ

Is my EA broken if it's losing money?

Usually not. Every strategy with a real edge has losing weeks and months — that's the cost of the edge, not a malfunction. A bot is "broken" if its logic fails or its risk blows past its defined limits, not simply because it's in a normal drawdown. Judge it over hundreds of trades, not one bad stretch.

Should I turn off my EA during a drawdown?

Almost never mid-drawdown. Switching off at the bottom locks in the loss and skips the recovery — the exact buy-high-sell-low loop that empties accounts. Sitting still through a normal drawdown is one of automation's biggest advantages over manual trading.

Why doesn't the "best EA" exist?

Because no single strategy wins in every market condition, on every pair, forever. Markets change regimes; a bot tuned for one struggles in another. A "perfect" backtest is usually overfit to the past. The realistic goal is a portfolio that survives all conditions well enough to compound.

How many EAs make a good portfolio?

Enough that they don't all bleed at the same time — typically several strategies across different pairs and timeframes, so their losing stretches don't overlap. The diversity matters more than the exact number: uncorrelated strategies are what flatten the swings and keep the account alive.

Trading involves real risk to capital. A normal drawdown is not a guarantee of recovery, and portfolio diversification reduces but does not remove risk — past performance does not guarantee future results, for any EA, including the ones I sell.

Want the honest data every week? Real EA performance — including the losing weeks and the quiet months — AI model comparisons, and portfolio thinking instead of holy-grail hunting: subscribe to the newsletter.