Your Profitable EA Is Charging You to Lose Money
You backtested your MT5 EA over 24 months. 10% annual return. Clean equity curve. You deploy it live and watch in confusion as the account bleeds $50, $80, $120 per week for no reason.
Here's what happened: your backtest didn't include swap charges.
Swap points are the interest you pay for holding a currency position overnight. On forex pairs, they're brutal. On leveraged accounts, they're catastrophic. A $3,000 account holding a 10-lot position overnight on EUR/USD pays $8–15 in swaps per night. Over a year, that's $2,000–$5,000 in pure cost. On a $3,000 account, you're bleeding out before you even win a single trade.
Most traders don't calculate this. Their backtestors don't include it. Their brokers don't warn them. So they deploy an EA that the math says is profitable, and reality proves is a wealth transfer to their broker.
What Are Swap Points—And Why They Destroy Your Edge
Swap points are overnight financing charges. When you hold a currency pair position, you're borrowing one currency and lending the other. The interest difference gets charged to you every night at 5pm EST (New York close).
On EUR/USD, if you're long 1 lot (100K of EUR), you pay the difference between EUR interest rates and USD interest rates. On certain pairs—like GBP/USD or NZD/USD—swaps run 50–200 pips per year.
Here's the actual damage:
- 1 lot held 252 trading days on GBP/USD: -200 pips annually = -$200 on a micro lot, -$2,000 on a standard lot
- Same 1 lot on NZD/USD with 100 pip annual swap: -$1,000 on a standard lot
- Your $3,000 account holding 10 micro lots overnight: -$1,000–$3,000 annually in swaps before a single losing trade
That's 30–50% of your edge deleted by overnight financing alone.
Investopedia's guide to forex swaps explains the mechanics. OANDA's interest rate differential documentation shows real-time swap rates by pair and broker.
The Math: How Your Backtest Becomes Fiction
You ran 24 months of backtest data. Your EA made 120 trades. Win rate 58%. Average winner +45 pips. Average loser -30 pips. Net gain: +2,100 pips = 10.5% annual return.
Now add swaps:
- Assume your EA holds positions an average 4 hours per trade (conservative)
- 30% of trades hold overnight = 36 trades per year with overnight exposure
- Average overnight cost on your favorite pair: 2 pips per night = -$20 per trade on 1 standard lot
- 36 trades × $20 = -$720 annually = -0.7% of account
You dropped from 10.5% profit to 9.8%. Still acceptable.
Now the real scenario:
- Your EA scalps in trending markets but holds overnight when positions go slightly underwater (hoping recovery)
- That's 60% of trades holding overnight minimum = 72 trade-nights per year
- Average swap cost: 3 pips per night = -$30 per trade
- 72 trades × $30 = -$2,160 annually = -2.16% of account
You're at 8.3% annual return. Respectable, but the $300/month dream from a $3,000 account is gone.
Now add 5:1 leverage. Swaps multiply by your leverage ratio.
-$2,160 becomes -$10,800 in annual swaps.
Your 10.5% becomes a -0.3% loss. Your EA is bankrupt.
Why Backtesting Never Warns You
Most backtesting platforms have swap modeling, but it's optional and disabled by default. Traders don't enable it because:
- They don't know it exists. Most autodidacts learning MQL5 never touch backtest settings.
- It ruins the backtest numbers. Enabling swaps means admitting the EA doesn't actually work.
- Brokers hide it. Swap rates are in the trading agreement fine print. Nobody reads it.
So traders deploy what they think is a 10% EA. What they've actually deployed is a 10% EA with a -5% tax they forgot to model.
The fix: run backtests WITH swap modeling enabled. Use your broker's real swap rates. Model worst-case scenarios where overnight holds extend longer than expected.
If your EA survives that, it might actually work. If it doesn't—which it usually doesn't—you need a different strategy.
Leverage Makes Swaps Catastrophic
A trader with a $10,000 account using 1:1 leverage pays swaps on 1 unit of risk. It hurts, but it's manageable.
A trader with the same $10,000 account using 10:1 leverage pays swaps on 10 units of risk. Now swaps become a profit killer.
A trader using 30:1 leverage? The swaps alone cost more than the account makes in a year.
This is why retail accounts with high leverage die silently. Price action looks fine. Win rate is 55%. But the compounding swaps plus losses equal account death within 12 months.
If you use leverage, accept this: swaps cost you real money. On a 10:1 leveraged account holding overnight, budget for -3–5% annual drag from swaps alone.
The Swap Rate Trap: Brokers Win Either Way
Swap rates vary wildly by broker. An EA that's break-even at Broker A becomes unprofitable at Broker B because of different overnight costs.
The trap: brokers with lower spreads often have higher swaps. Brokers with high spreads offer lower swaps. You can't escape the cost—it just moves to a different line item.
Some pairs are worse than others:
- EUR/USD: low swaps, tight spreads (safe)
- GBP/USD: moderate swaps, volatile spreads (risky)
- Emerging market pairs (USD/TRY): astronomically high swaps (death wish for bots)
- Crypto pairs on leverage: swaps reach 50–100 pips annually (avoid)
A working bot on EUR/USD might be completely unprofitable on GBP/USD at the same broker, just because of swap rate differences.
The answer: custom MT5 Expert Advisors designed specifically for your broker's swap structure. A professional EA builder models swaps during development and optimizes pair selection and holding duration to minimize overnight costs. Most builders don't. That's why most EAs fail.
How to Actually Profit With 24-Hour Trading Bots
If you want to run an overnight-holding bot, engineer around swap costs, don't ignore them:
- Choose low-swap pairs. EUR/USD, USD/CHF, AUD/USD. Avoid GBP pairs and emerging markets.
- Minimize overnight exposure. Design your EA to close 70%+ of positions before 5pm EST. Hold only high-conviction setups overnight.
- Model swaps in backtests. Enable swap modeling in MetaTrader. Use your broker's actual rates. Include worst-case scenarios.
- Budget for swap drag. If backtests show 10% return, assume 6% after real swap costs, slippage, and margin calls.
- Avoid high leverage. 1:1 or 2:1 leverage caps swap costs. 10:1 leverage multiplies them by 10.
The traders who profit with 24/7 bots all do the same thing: ruthlessly engineer for low overnight costs. They don't deploy generic EAs. They deploy custom systems optimized for their broker's fee structure and specific pairs.
That's exactly what we build at Alorny. Not a template. Not a one-size-fits-all EA. A custom MT5 system designed for your strategy, your broker, your pairs, and—critically—your swap costs. Starting from $100 for simple strategies. You get a full backtest report that includes realistic swap modeling so you know exactly what you're getting before deployment.
Key Takeaways
- Swap points are interest charges on overnight positions. On small leveraged accounts, they cost $3–8K annually. That's 50–100% of your stated edge.
- Backtests are fiction if they don't include swap modeling. Most traders deploy EAs without ever enabling this setting.
- Leverage multiplies swap costs. 10:1 leverage means 10x the overnight drain. It's the hidden killer of retail accounts.
- Not all pairs are equal. EUR/USD swaps won't kill you. GBP/USD and emerging market pairs will.
- The fix is custom engineering, not generic templates. A profitable 24-hour bot needs to be built for your exact broker, pairs, and cost structure.
Pull your backtest report. Check if swap modeling was enabled. If it wasn't, your EA isn't actually profitable. Run it again with swaps on. If the numbers still hold, deploy. If they collapse—which they usually do—you know exactly why live trading differs from backtests.