ALGORY_OS // PARITY GUIDEalgory.app
MT5 PARITY TEST
SET THE STRATEGY TESTER UP RIGHT, OR PARITY WILL LIE TO YOU.
A parity test checks that your bot takes the same trades in MetaTrader 5 as it did in Algory. But it only means something if MT5 runs the bot over the exact same conditions Algory used: the same symbol, the same timeframe, and the same date range.
Get any one of those wrong and the test will not match, even though the bot is perfectly fine. This is the number one reason people think a strategy failed parity when it did not. This guide makes sure you set it up right.
THE BOT'S NAME TELLS YOU HOW TO SET IT UP
Everything you need is in the bot's name. It contains the symbol it was bred on, the timeframe it trades, and the date it started trading.
For example, a bot named like GBPUSD_H1_..._20250103 means: symbol GBPUSD, timeframe H1, first trade 3 January 2025. Your job in the Strategy Tester is simply to match those three things exactly. If you also have Algory open, the strategy's first-trade date is in its trade list.
RIGHT SYMBOL
EXACTLY AS IN THE NAME
RIGHT TIMEFRAME
EXACTLY AS IN THE NAME
RIGHT START DATE
THE FIRST-TRADE DATE
1-MIN OHLC
MATCHES ALGORY, RUNS FAST
01OPEN THE STRATEGY TESTER
The parity test runs inside MT5's built-in Strategy Tester.
- In MT5, go to View then Strategy Tester, or press Ctrl+R.
- On the Overview tab, click Single so you run one test (not the optimiser).
- In the Expert field, select your compiled bot.
02MATCH THE SYMBOL AND TIMEFRAME EXACTLY
This is where most parity tests go wrong. The symbol and timeframe are baked into the bot. The tester must match them, not whatever chart you had open.
- Symbol: set it to the exact symbol in the bot's name (for example GBPUSD). Not the last chart you looked at.
- Period: set the timeframe to the exact one in the bot's name (for example H1 or H2).
- Double-check both against the bot's name before you go any further.
HEADS UPWrong timeframe = a meaningless test
Leaving the Period on its default, or on your chart's timeframe, is the most common mistake. A bot bred on H1 tested on H4 trades completely different signals. It will never match Algory, and the bot is not the problem, the setting is.
03SET THE START DATE FROM THE BOT'S NAME
Algory tested the bot over one specific window. MT5 has to use the same window, starting from the strategy's first trade.
- Tick Use date (or set a custom date range).
- From date: set it to the strategy's first-trade date. That date is in the bot's name, and also in Algory's trade list.
- To date: set it to the end of the training window, or just leave it at today.
- Do not leave the From date on the MT5 default.
HEADS UPThe biggest parity-killer of all
If you leave the From date on MT5's default instead of the bot's first-trade date, MT5 runs the bot over a totally different stretch of history than Algory did. The returns will be miles apart and it looks like parity failed, when really you just tested the wrong dates.
04USE 1-MINUTE OHLC MODELLING
The modelling mode has to match how Algory works. Algory does not simulate ticks. It resolves trades on one-minute (M1) OHLC sub-bars, so the Strategy Tester should do the same.
- Set Modelling to 1 minute OHLC. This mirrors exactly what Algory does, gives the truest parity, and finishes in seconds instead of minutes.
- Do not use Every tick or Every tick based on real ticks. Algory never models ticks, so those modes are far slower and can actually diverge more, not less.
- Set the Deposit and Leverage to match what you used in Algory (the default balance is 100,000).
05RUN IT AND READ THE RESULT
Now you can trust what the test tells you.
- Click Start and let it finish.
- Compare the net return to Algory's. They should be in the same ballpark. Small differences from spread and slippage are normal.
- Open the Journal tab and read the Parity Score the bot prints at the end.
NOTEReading the score
Grade A or B is safe to deploy. Grade C can still be traded if the MT5 return is strong and out-of-sample held up. Grade D or F needs a careful look before you trust it.
TIPIf the numbers are way off
Before you blame the bot, re-check three things: the symbol, the timeframe, and the From date. Nine times out of ten one of those is wrong, not the strategy.
QUICK ANSWERS
Where do I find the timeframe and start date?
Both are in the bot's name: the symbol, the timeframe, and the first-trade date. The date is also in Algory's trade list for that strategy.
It matched fine, but I tested on H4 instead of H1.
Then it did not really match. A different timeframe is a different test. Set the Period to exactly the timeframe in the bot's name and run it again.
My MT5 return is way lower than Algory's.
Almost always the From date or the timeframe is wrong, so MT5 tested a different period. Fix the symbol, timeframe and start date first, then re-run.
Should I use real ticks?
No. Algory resolves fills on one-minute OHLC bars, not ticks, so set Modelling to 1 minute OHLC. It matches Algory exactly and runs in seconds. Real ticks is much slower and can actually diverge more.
What parity score is good enough?
A or B is safe. C is tradeable if the return is strong and out-of-sample held. D or F means investigate before trusting it.
STILL NOT MATCHING?
Send your bot name and a screenshot of your Strategy Tester settings in the Algory Discord and we will spot what is off. It is nearly always the symbol, timeframe or start date.
ASK IN DISCORD