If you've ever run a backtest in MetaTrader 5 and stared at the Modeling Quality line in the Results tab, you've seen the two options: Every Tick and 1 Minute OHLC. Most traders click one and hope for the best. That's a mistake. Choosing the wrong mode can turn a profitable-looking EA into a live-trading disaster because the backtest results simply aren't reproducible.
This guide walks you through exactly what each mode does, when to use them, and how to interpret the Modeling Quality indicator. You'll leave knowing which setting to pick for your specific EA and why the other one can waste hours of your time.
What You'll Learn
- The technical difference between Every Tick and 1 Minute OHLC modeling
- How MetaTrader 5 generates price ticks during a backtest
- When Each Tick mode is mandatory and when 1 Minute OHLC is acceptable
- How to read the Modeling Quality percentage and what it means for your results
- Practical steps to verify your EA's behavior under both modes
Prerequisites
Before diving into the settings, make sure you have:
- MetaTrader 5 (build 2000 or newer) installed. The same concepts apply to MT4, but the exact dialog names differ slightly.
- An Expert Advisor (EA) you want to test. I'll use a simple moving average crossover EA in the examples, but the advice applies to any EA.
- Historical tick data for the symbols you plan to test. MT5 downloads this automatically when you select a symbol in the Strategy Tester, but older data may be incomplete. You can check by opening the Tools > History Center and looking for tick data for your symbol.
- A demo or live account connected. The Strategy Tester requires an active connection to download data, even for offline testing.
Step-by-Step: Accessing and Selecting Modeling Modes
- Open the Strategy Tester in MT5 (View > Strategy Tester or Ctrl+R).
- In the tester panel, set your Expert Advisor, Symbol (e.g., EURUSD), and Period (e.g., H1).
- Click the Settings tab (the gear icon).
- Under Modeling, you'll see two radio buttons:
- Every tick (based on all ticks)
- 1 minute OHLC (based on 1-minute bar opening, high, low, closing prices)
- Select the mode you want to test. The Modeling Quality indicator at the bottom of the dialog will update to show a percentage. This percentage tells you how much of the available tick data is being used for the backtest.
- Click Start to run the backtest.
Every Tick Mode: The Gold Standard
Every Tick mode uses every recorded tick in the historical data to simulate the market. For each tick, MT5 recalculates your EA's logic—checking conditions, executing orders, and updating indicators. This is the most accurate representation of how your EA would behave in real time.
Here's what happens under the hood: MT5 reads tick data from its history. For each tick (which can be hundreds per minute for major pairs like EURUSD), it calls the OnTick() function in your EA. If your EA uses OrderSend() or PositionOpen(), the order is executed at that exact tick price, subject to slippage and spread settings you configure in the tester.
The key benefit is that your EA sees every price movement—including sudden spikes, gaps, and volatile periods. A scalping EA that relies on fast entries and exits will only behave correctly under Every Tick mode. If you use 1 Minute OHLC, the EA might miss the exact tick where an entry signal triggers, leading to unrealistic fill prices.
However, Every Tick mode has a cost: it's significantly slower. A backtest over 10 years on EURUSD H1 might take 30 minutes instead of 5 minutes. It also requires complete tick data. If your broker's history has gaps (common for exotic pairs or older data), the Modeling Quality percentage drops below 90%, and the results become unreliable.
1 Minute OHLC Mode: The Compromise
1 Minute OHLC mode is a compromise between speed and accuracy. Instead of using every tick, MT5 condenses the tick data into 1-minute bars. For each minute, the tester generates a synthetic sequence of four ticks: open, high, low, and close (in that order). Your EA's OnTick() function is called four times per minute, once for each synthetic tick.
The problem is that this synthetic sequence doesn't reflect the real order of price movements within that minute. In reality, price might have hit the high first, then the low, then closed near the open. In the synthetic model, the order is always open, high, low, close. This can break EAs that rely on the sequence of events—for example, a stop-loss triggered at the low before a reversal to the high.
Here's a concrete example: Suppose your EA places a pending buy stop order at 1.1050. In real tick data, price hits 1.1050, triggers the order, then continues to 1.1060. In 1 Minute OHLC mode, if the high for that minute is 1.1060 and the low is 1.1040, the synthetic ticks are: open at 1.1045, high at 1.1060, low at 1.1040, close at 1.1055. The pending order at 1.1050 might trigger at the high tick (1.1060) or the low tick (1.1040) depending on the order of the synthetic sequence. This is not how real markets work.
That said, 1 Minute OHLC is perfectly adequate for EAs that only trade on bar close—for example, an EA that checks conditions once per bar and enters at the next bar's open. These EAs don't react to intra-bar ticks, so the synthetic model is fine. It's also much faster: you can run a 10-year backtest in a few minutes.
Modeling Quality: What That Percentage Really Means
When you select a modeling mode, MT5 displays a Modeling Quality percentage at the bottom of the Settings tab. This number represents the ratio of ticks actually used in the backtest versus the maximum possible ticks available in the history.
For Every Tick mode, if you have complete tick data, you'll see 99.9% (it's never exactly 100% due to minor data gaps). For 1 Minute OHLC, the percentage is always lower because you're using only four ticks per minute instead of all ticks. Typically, it ranges from 2% to 15% for major pairs, depending on tick density.
Here's the critical rule: Never rely on a backtest with Modeling Quality below 90% for Every Tick mode. If your data is incomplete, the missing ticks can cause your EA to miss entries or exits, making the results meaningless. For 1 Minute OHLC, the percentage is inherently low—that's expected—but you should still check that the 1-minute OHLC data itself is complete. You can verify this by looking at the Bars count in the tester results: if you backtested 10 years on H1, you should see roughly 87,600 bars (10 years × 365 days × 24 hours).
When to Use Each Mode: A Decision Framework
Use this simple decision tree:
- Your EA trades intra-bar (e.g., uses tick-level signals, pending orders, or partial fills) → You must use Every Tick. 1 Minute OHLC will produce invalid results.
- Your EA only trades on bar close (e.g., checks conditions at the close of a 1-hour bar and enters at the next bar's open) → You can use 1 Minute OHLC safely. It's faster and sufficient.
- Your EA uses trailing stops, stop-losses, or take-profits that depend on intra-bar price movement → Use Every Tick. The 1 Minute OHLC model can't simulate partial fills or stop-loss triggers accurately.
- You're doing a quick optimization over many parameters → Start with 1 Minute OHLC to narrow down the parameter space, then validate the best candidates with Every Tick.
Practical Tips from Experience
Over the years, I've seen traders fall into the same traps. Here are the ones to watch for:
- Don't trust a backtest where the Modeling Quality is below 90% for Every Tick mode. I've seen new traders run a backtest on a minor pair like NZDJPY, see 60% quality, and still think the results are valid. They aren't. You're essentially guessing what happened during the missing ticks.
- Always run a final validation on Every Tick mode before live trading. Even if your EA seems to work fine on 1 Minute OHLC during optimization, the final 3-6 month forward test should be on Every Tick. I've caught EAs that had 50% drawdown in Every Tick but looked clean on 1 Minute OHLC.
- Check your tick data quality manually. Go to Tools > History Center, select your symbol, and look at the tick data. If you see large gaps (e.g., no ticks for 2 hours during a high-volatility news event), your backtest will be unreliable. Some brokers clean their data; others don't.
- Use the "Open prices only" mode for initial prototyping. MT4/MT5 also has a third mode called "Open prices only" (or "Every tick" based on open prices in older versions). This is the fastest and least accurate mode. Use it only for quick feasibility checks, never for final results.
- For high-frequency EAs (scalpers), Every Tick is non-negotiable. If your EA enters and exits within seconds, 1 Minute OHLC will produce completely fictional results. I've seen scalping EAs show 80% win rates on 1 Minute OHLC and 30% on Every Tick. The difference is all in the tick sequence.
Common Mistakes and Troubleshooting
Mistake 1: Ignoring the Modeling Quality percentage
You run a backtest, see a good equity curve, and don't bother to check the 99% vs 60% quality. Later, the EA fails live. Fix: Always check the Modeling Quality in the Results tab after the test. If it's below 90% for Every Tick, download fresh data or use a different broker's history.
Mistake 2: Using 1 Minute OHLC for a scalping EA
Your EA is designed to trade on 1-minute bars and uses tight stop-losses. You run the backtest on 1 Minute OHLC and get fantastic results. Live, it loses money. Fix: Scalping EAs must be tested on Every Tick. The 1 Minute OHLC model cannot simulate the micro-price movements that trigger your stops.
Mistake 3: Assuming "Every Tick" is always better
You always select Every Tick because it's more accurate. But your EA only trades on bar close, and the backtest takes 10x longer for no benefit. Fix: Match the mode to your EA's logic. If your EA doesn't react to intra-bar ticks, 1 Minute OHLC is fine and faster.
Mistake 4: Not checking for data gaps before testing
You run a backtest on a symbol you rarely trade, and the Modeling Quality is 20%. You don't realize the data is corrupt. Fix: Before any major backtest, open the History Center and verify that tick data exists for the entire period. If you see gaps, use the Download button to fetch data from your broker, or switch to a different data provider.
Summary and Next Steps
Here's the bottom line: Every Tick mode is the only way to get accurate backtest results for EAs that react to intra-bar price movements—scalpers, pending order systems, and any EA with stop-losses or trailing stops. 1 Minute OHLC is a faster, acceptable alternative for EAs that only trade on bar close. The Modeling Quality percentage is your early warning system: below 90% for Every Tick means your data is incomplete, and the results are suspect.
Your next step should be to run your current EA under both modes and compare the results. If they differ significantly, you know your EA depends on tick-level behavior—and you need to trust the Every Tick results, not the 1 Minute OHLC ones. If they're similar, you can safely use 1 Minute OHLC for faster optimizations.
Finally, consider investing in high-quality tick data from a third-party provider like Dukascopy or Tick Data Suite if you're serious about backtesting. MT5's built-in history is adequate for most purposes, but for professional-grade validation, cleaner data makes a real difference.
Frequently Asked Questions
Can I use 1 Minute OHLC to test a scalping EA?
No. Scalping EAs rely on precise tick-level entries and exits. 1 Minute OHLC generates synthetic ticks in a fixed order (open, high, low, close), which doesn't reflect real market behavior. Your backtest results will be misleading—often showing unrealistic win rates. Always use Every Tick for scalpers.
My Modeling Quality is 99% on Every Tick. Does that guarantee accurate results?
Not entirely. 99% means the tick data is nearly complete, but the data itself could still have errors (e.g., incorrect bid/ask spreads, missing ticks during high volatility). Always forward-test your EA on a demo account for at least a month to confirm the backtest results hold in real market conditions.
Why does 1 Minute OHLC show a lower Modeling Quality percentage than Every Tick?
The percentage is calculated as the ratio of ticks used to the total ticks available. In 1 Minute OHLC mode, MT5 uses only 4 ticks per minute (open, high, low, close) instead of the hundreds of real ticks. So the percentage is naturally low—often 2-15% for major pairs. This is expected and not a problem, as long as the 1-minute OHLC data itself is complete.
Should I ever use "Open prices only" mode?
Only for the fastest possible feasibility check—for example, to see if an EA has any edge at all on a 4-hour chart. Never use it for final validation. It ignores all intra-bar price movement and can make any EA look profitable. Treat it as a rough filter, not a reliable test.
