The difference between 1 minute OHLC and the every tick backtest??

 

Hey, everyone could anyone help me and explain the difference between 1 minute OHLC and the every thick backtest, because I have a great and fantastic results in the 1 OHLC modeling even if is a long term backtest, and in the opposite on the other modeling a weak results and negative results?

How it could affect my profitabily in my EA?

What i can do?

Best Regards, Will

 
Will_FX:

Hey, everyone could anyone help me and explain the difference between 1 minute OHLC and the every thick backtest, because I have a great and fantastic results in the 1 OHLC modeling even if is a long term backtest, and in the opposite on the other modeling a weak results and negative results?

How it could affect my profitabily in my EA?

What i can do?

Best Regards, Will

The difference is huge, on OHLC data, the tester only works with the open, high, low and close of 1 minute data, while on every tick it takes into consideration every single tick of the price.

If your EA is based on small price movements, the trustworthingness of the results could be severely affected.

 
Eleni Anna Branou:

The difference is huge, on OHLC data, the tester only works with the open, high, low and close of 1 minute data, while on every tick it takes into consideration every single tick of the price.

If your EA is based on small price movements, the trustworthingness of the results could be severely affected.

On M1 timeframe there is not much difference ;-)
 
Eleni Anna Branou:

The difference is huge, on OHLC data, the tester only works with the open, high, low and close of 1 minute data, while on every tick it takes into consideration every single tick of the price.

If your EA is based on small price movements, the trustworthingness of the results could be severely affected.

OK,@Eleni Anna Branou thanks for the help!

 
Alain Verleyen:
On M1 timeframe there is not much difference ;-)

I don´t think so ahah!

thanks for your time!

 
How can I use OHCL  i n real-life trading ?
 
Thao Pham #:
How can I use OHCL  i n real-life trading ?

You would have to force your EA to act on new bars exclusively . But then you would have to enforce symmetry between the backtest and live conditions .

Some things would be :

  • Deciding on how to act on your stops after the bar closes and has pierced them , so , virtual stops .
  • Be vigilant about your drawdown by measuring the edge prices in the backtests (which would be nearly impossible to measure with a multi asset system as the highs and lows of each bar of each symbol do not happen at the same time)
  • Force closing trades that gap more than usual across 2 bars (meaning data gaps) . Deciding how to treat those in the backtest evaluation .
 
Eleni Anna Branou #:

The difference is huge, on OHLC data, the tester only works with the open, high, low and close of 1 minute data, while on every tick it takes into consideration every single tick of the price.

If your EA is based on small price movements, the trustworthingness of the results could be severely affected.

Be aware that the "every tick" is not based on real ticks, unless you select the "real ticks" mode on your tester. The modeling for "every tick" is a rigid interpolation for OHLC. Prefer "real ticks" in case of reliable historical data. How data is interpolated: https://www.metatrader5.com/en/terminal/help/algotrading/tick_generation There is another comparison written in the book: https://www.mql5.com/en/book/automation/tester/tester_ticks

MQL5 Book: Trading automation / Testing and optimization of Expert Advisors / Generating ticks in tester
MQL5 Book: Trading automation / Testing and optimization of Expert Advisors / Generating ticks in tester
  • www.mql5.com
The presence of the OnTick handler in the Expert Advisor is not mandatory for it to be tested in the tester. The Expert Advisor can use one or more...
 
Thao Pham #:
How can I use OHCL  i n real-life trading ?

  • Check bar prices and indicators from latest closed bar to the past, never check the current candle values. 
  • Don't use Ask/Bid in your functions (like trailing, breakeven etc.) Do that by candle or don't do that at all (probably the best thing :D)
  • Use SL and TP on your trades and avoid using EA commands to close trades (avoid using functions based on profit amount, % and so on)
Probably missing some further considations, but most important things are them!