Reality Check

 

Hello all, it's been a while but I'm still plugging away.

I would like to verify something that might seem incredibly basic, but just for the sake of argument, humor me.

IF you have backtested a strategy properly including sample size, excluding time periods you can't actually trade in real life, etc etc. Then any failure to make profit on the strategy is entirely in the realm of user error. Correct?

In other words, so long as I stick to the rules and do not deviate from them, the results should match the backtests more or less.

Thanks!

 
Tristen Shaw:

Hello all, it's been a while but I'm still plugging away.

I would like to verify something that might seem incredibly basic, but just for the sake of argument, humor me.

IF you have backtested a strategy properly including sample size, excluding time periods you can't actually trade in real life, etc etc. Then any failure to make profit on the strategy is entirely in the realm of user error. Correct?

In other words, so long as I stick to the rules and do not deviate from them, the results should match the backtests more or less.

Thanks!

If by backtest you mean use the strategy tester, that’s not necessarily the case.

- Be sure to have good historical data files.
- If you are using an Expert Advisor, for which you don’t have access to the source code, try to find out if there is any history reading code that might cheat the backtest. It is VERY easy to fake backtests results and show total lies, even on the MQL5 market.
- Be sure you take into consideration spreads, swaps and commission.
- MetaTrader also has a tick data generation algorithm, which might differ from real results.
- Check for your leverage as well. 

Otherwise, if by backtest you mean use replay tools such as the one you might find on other platforms, I don’t believe there will be any issues.

And always remember: backtest do not guarantee future results. Believe me, I had way different results Pre and Post COVID, even with backtests.

Though yes, it usually more or less gives you a good idea.
 
Zaky Hamdoun #:
If by backtest you mean use the strategy tester, that’s not necessarily the case.

- Be sure to have good historical data files.
- If you are using an Expert Advisor, for which you don’t have access to the source code, try to find out if there is any history reading code that might cheat the backtest. It is VERY easy to fake backtests results and show total lies, even on the MQL5 market.
- Be sure you take into consideration spreads, swaps and commission.
- MetaTrader also has a tick data generation algorithm, which might differ from real results.
- Check for your leverage as well. 

Otherwise, if by backtest you mean use replay tools such as the one you might find on other platforms, I don’t believe there will be any issues.

And always remember: backtest do not guarantee future results. Believe me, I had way different results Pre and Post COVID, even with backtests.

Though yes, it usually more or less gives you a good idea.

I do not mean strategy tester. I mean going back through MT4 data and manually marking wins losses and BE based on an objective and systematic entry and exit system.

So, in general, if I were to do this for say 100 - 500 entries and the winrate and profit says it's good, then in general if I'm not successful at making a profit it's MY fault, not the strategy. Somewhere along the lines I either didn't stick to my rules or missed my wins, didn't take fees into account, etc etc etc.

Basically I'm trying to establish that it really is that simple. That if the strategy works in backtest, it should work in real life so long as you don't screw it up.

 
Tristen Shaw #:

I do not mean strategy tester. I mean going back through MT4 data and manually marking wins losses and BE based on an objective and systematic entry and exit system.

So, in general, if I were to do this for say 100 - 500 entries and the winrate and profit says it's good, then in general if I'm not successful at making a profit it's MY fault, not the strategy. Somewhere along the lines I either didn't stick to my rules or missed my wins, didn't take fees into account, etc etc etc.

Basically I'm trying to establish that it really is that simple. That if the strategy works in backtest, it should work in real life so long as you don't screw it up.

Yes. Then it that case you are right.

Just be sure your MT4 has enough historical data, and you aren’t trading on void. 
 
Zaky Hamdoun #:
Yes. Then it that case you are right.

Just be sure your MT4 has enough historical data, and you aren’t trading on void. 

Trading on void?

 
Tristen Shaw #:

Trading on void?

An expression, as in « trading on not accurate data ».

Sorry for that, it’s a French expression I translated thinking it would exist in English as well 😅
 
Zaky Hamdoun #:
An expression, as in « trading on not accurate data ».

Sorry for that, it’s a French expression I translated thinking it would exist in English as well 😅

Ah I gotcha, no worries lol.

I'm backtesting on the MT4 charts from an FTMO demo. I would hope that it's accurate, and I do try to take spread into account. Usually swap isn't an issue since most of my trades are very short and I close them before the end of the trading day.

 
Tristen Shaw #:

Ah I gotcha, no worries lol.

I'm backtesting on the MT4 charts from an FTMO demo. I would hope that it's accurate, and I do try to take spread into account. Usually swap isn't an issue since most of my trades are very short and I close them before the end of the trading day.

That’s not necessarily good, MT4 historical data is usually bad unless you download new data.

If I were you I would backtest my manual trading strategies on other platform which permit Trades replay. Unfortunately, I am not sure I can say its name on the Forum.
 
Zaky Hamdoun #:
That’s not necessarily good, MT4 historical data is usually bad unless you download new data.

If I were you I would backtest my manual trading strategies on other platform which permit Trades replay. Unfortunately, I am not sure I can say its name on the Forum.

I'm confused, how can their data be bad? Isn't it the same data that you would use if you were doing an FTMO challenge?

Alternatively, I do use the Soft4FX simulator which uses Dukascopy data. Unless you set it to use TrueFX.

 
Tristen Shaw #:

I'm confused, how can their data be bad? Isn't it the same data that you would use if you were doing an FTMO challenge?

Alternatively, I do use the Soft4FX simulator which uses Dukascopy data. Unless you set it to use TrueFX.

Let’s say you got your account on the 01/12/2023 (DD/MM), and try to go back to 01/01/2020, you won’t necessarily have good data.
 
Zaky Hamdoun #:
Let’s say you got your account on the 01/12/2023 (DD/MM), and try to go back to 01/01/2020, you won’t necessarily have good data.

Do I need 3 years worth of data? I trade on the 5M and I can get 100+ trades in less than a month or two. Within a year of data I can have hundreds of entry results. And what account are we talking about? FTMO? What about the Dukascopy data? From what I saw online it seems like most people consider it to be good data.