Is this a Good Result for BackTest and Forward Test ? - page 2

New comment
 
Andrew Pun Magar #:
The OP is using 5min timeframe, so the algo bot is prolly a scalper.... and scalping EAs must use VPS to stay accurate or else it will lose it's touch Anyways we must understand what actually the OP is trying to do. For me personally, 30ms latency is too slow. 

I understand what you're saying, and that's good for you.

Personally, I scalp XAUUSD without issue at about a 37ms latency. The way that I see it... One M5 bar contains 5 minutes, each minute contains 60 seconds, and each second contains 1000 milliseconds. That equates to 300,000 milliseconds per M5 bar. Now, factor in code execution time which is generally several hundred milliseconds. Proportionally, a latency of 30 to 37ms is a mere drop in the bucket. Additionally, XAUUSD ticks are so wild that I find that I'm just as likely to get a better price of execution as a worse price of execution at any give time. In my case, I have no need for the added variable of VPS service.

At the end of the day, a trader's simulation must match what that trader intends to do live to the fullest extent possible. That is all.
 
Ryan L Johnson #:

Ideally, yes. As a potential caveat, a lengthy test based on real ticks can be burdensome on hardware. If you have a beastly pc, network farm, or remote agents, the burden is eased. Some EA's, especially EA's that trade on bar close, can be reliably tested on OHLC prices wherein stops are approximated in the Tester.

Thank you for the explanation. That makes sense. "Every tick based on real ticks" optimization takes around 12 hours to complete. I'll also test my EA using 1 Minute OHLC and compare the results with Every tick based on real ticks.

My EA uses Moving Average and MACD and places trades at the start of a new candle. In my opinion, would 1 Minute OHLC be accurate enough for this type of EA, or would still recommend Every tick based on real ticks ?

 
Ryan L Johnson #:

Personally, I'm not a big fan of optimizing simple EA's. In your case, you have a moving average and a MACD which is relatively simple. I tend to question whether I can actually randomly select the right sample period and forward period for such an EA. I like to refine my entire EA to be profitable in the long-term without optimizing.

A more advanced machine learning EA, for example, is a horse of different color. Therein, the EA could potentially statistically select the right sample period and forward period to apply. In this way, the EA could control its own rolling optimizations. I recall having seen an Article regarding that on this website. It was called a self-optimizing EA.

Thank you for your advice. I understand your point. I'll learn more about self-optimizing EAs in the future.
 
Andrew Pun Magar #:
One thing out of many things, you are testing at 30ms latency whereas many EAs must be ran on VPS.... and VPS provides less than 1ms latency environment... so you must test at a realistic environment. 

Sir, I am using the MQL5 VPS, and the latency to my Exness broker is 30 ms. This is the actual real-time latency reported by the MQL5 VPS.

 
Vikram J U #:

Thank you for the explanation. That makes sense. "Every tick based on real ticks" optimization takes around 12 hours to complete. I'll also test my EA using 1 Minute OHLC and compare the results with Every tick based on real ticks.

My EA uses Moving Average and MACD and places trades at the start of a new candle. In my opinion, would 1 Minute OHLC be accurate enough for this type of EA, or would still recommend Every tick based on real ticks ?

The start of a new candle is literally an open price. Therefore, you might get very similar OHLC results to Open prices results. The other pieces of the puzzle are your exits. Are they dynamically based on OHLC prices? If not, real ticks may be necessary to get more precise results.
 
Andrew Pun Magar #:
The OP is using 5min timeframe, so the algo bot is prolly a scalper.... and scalping EAs must use VPS to stay accurate or else it will lose it's touch Anyways we must understand what actually the OP is trying to do. For me personally, 30ms latency is too slow. 

Thank you for your feedback. I don't think 30 ms latency will have a significant impact on my EA, but I'll test it further. Tomorrow, I'll set up a demo account and analyse the performance of my scalping EA under real market conditions.

 
Ryan L Johnson #:
The start of a new candle is literally an open price. Therefore, you might get very similar OHLC results to Open prices results. The other pieces of the puzzle are your exits. Are they dynamically based on OHLC prices? If not, real ticks may be necessary to get more precise results.

Thank you for the explanation. I understand your point. I'll use "Every tick based on real ticks" for my optimization tomorrow.

 
Vikram J U #:

Sir, I am using the MQL5 VPS, and the latency to my Exness broker is 30 ms. This is the actual real-time latency reported by the MQL5 VPS.

I think you are confused.... MQL5 VPS has latency of below 1ms... check in the VPS section, not the bottom right latency that you see on MT5... which is the latency from broker
 

Question about Self-Optimizing EAs :-

What are the recommended settings for a self-optimizing EA regarding :

  • Number of Maximum Parameter Combinations per Cycle

  • Number of Months for Optimization

  • Number of Months for Forward

  • Number of Re-optimization frequency (Example Every 1 - 2 Month)

Does anyone here have experience with self-optimizing EAs and could share their recommendations?

Thank you!

 
Vikram J U #:
Does anyone here have experience with self-optimizing EAs and could share their recommendations?

I don't use it, but here are some leads for more information:

Forum on trading, automated trading systems and testing trading strategies

Self EA optimization

Fernando Carreiro, 2022.10.10 13:47

Yes, there are many ways. Do a search and you will find several. Here are a few ...



123
New comment