EVENLINETRADING
FR

Last updated: October 10, 2026

Lesson — Going further

EA risks and scams: what can go wrong, and how to spot it

Automation attracts sellers who promise passive income. Here are the real risks and the warning signs to learn before you run, or buy, anything.

"Make money while you sleep" is one of the most powerful sales lines in trading, which is exactly why it is so widely abused. Some bots are honestly built and honestly sold. Many are neither. This lesson covers the risks you can run into even with an honest bot, and the tricks to watch for with a dishonest seller.

Risk 1: the martingale and grid trap

A martingale bot increases its position size after each loss, hoping the next win recovers everything. A grid bot keeps adding trades as price moves against it. Both can post a long, smooth run of small wins, which makes them look superb in screenshots. Then one long move against them arrives.

The martingale curve: many small wins, then one cliff

Illustrative, not real data: a martingale or grid bot doubles down after losses, so it wins small amounts again and again. The risk is hidden until a long run against it arrives, and the size of its positions grows exactly when the account is weakest.

A simple illustration of the arithmetic

A bot starts at 0.01 lot and doubles after every loss. After 10 losses in a row, the next trade would be 0.01 × 210 = 10.24 lots, over a thousand times the starting size.

Even if each trade is a 50/50 bet, a run of 10 losses happens roughly once in every 1,000 sequences. With enough trading, it arrives. The account can reach a margin call well before the doubling "recovers" anything.

For a full lesson on these techniques, see martingale, grid and averaging down. The point is not that every bot using averaging is a scam. It is that the risk is large, delayed and easy to hide, so a good-looking recent record says very little about it. See risk of ruin for the math behind repeated losses.

Risk 2: the overfitted backtest

A bot has settings, such as indicator lengths or levels. If you keep adjusting them until the historical results look great, you may simply have made the bot memorize the past. That is overfitting or curve-fitting. The perfect-looking curve then falls apart on new data.

Backtest versus live: the gap that overfitting creates

Illustrative, not real data: a bot tuned until it looks perfect on old data has learned the past, not the market. The smooth backtest curve and the live curve part ways once the data is new.

Related traps include tests that use unrealistic fixed spreads, ignore commissions and slippage, or use poor-quality price data. The backtesting basics and backtesting checklist lessons show how to test more honestly.

Risk 3: marketing that misleads

Watch for these claims and tactics:

Risk 4: technical and operational problems

A quick red-flag checklist

The takeaway: the biggest risks in automation are the ones that stay quiet for a long time: martingale blow-ups and curve-fitted tests. If a result looks too smooth, treat that as a reason for more questions, not less. Be careful with anything sold on the promise of passive income, and check your local rules if someone is offering to manage trading for you.

Keep going

See the rest of the free lessons.

Back to all lessons

This lesson is educational content only and not financial advice. Trading forex and CFDs involves significant risk of loss and is not suitable for every investor.