18 years Amazing back test result | Page 3 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

AMAZING Strategy!!! 3,266 replies

Back Test vs Forward Test 6 replies

Unknown Indicator and amazing result 28 replies

Why EA test is forward test better than back test?? 1 reply

Firebird V1.0 an amazing EA ! 33 replies

Attachments: 18 years Amazing back test result
Exit Attachments

18 years Amazing back test result

  • #41
  • Jul 5, 2018 12:43am Jul 5, 2018 12:43am
First trade of the week finally triggered.

This is my 3rd strategy which doesn't required a perfect setup as you can see in weekly, it is not really a clear upward trend yet but the EA triggers a buy.

Let's see.
  • #42
  • Jul 5, 2018 2:39am Jul 5, 2018 2:39am
Quoting DonaldDuckie
First trade of the week finally triggered. This is my 3rd strategy which doesn't required a perfect setup as you can see in weekly, it is not really a clear upward trend yet but the EA triggers a buy. Let's see.
Price came so close to TP, 2 pips away and u-turned. Since this is supposed to be 100% automatic EA, I will not manually close any trades. Let's see if it can continue with the uptrend.
  • #43
  • Jul 5, 2018 5:08am Jul 5, 2018 5:08am
Quoting DonaldDuckie
{quote} I will try it out soon in a couple of months time. I think it will work fine. Current demo account has higher spread as compared to the live which I will be using. Example for EURUSD, the spread on demo is around 2 pips which is crazy.
in your backtest, how many lot sizes did you trade after the 18 years? it looked in your last backtest you implemented here ,you increased lotsize a lot with time, so on the end it can be a big size and so spread of 2 pips can happen very easy.
  • #44
  • Jul 5, 2018 5:20am Jul 5, 2018 5:20am
Quoting dukas_trader
{quote} in your backtest, how many lot sizes did you trade after the 18 years? it looked in your last backtest you implemented here ,you increased lotsize a lot with time, so on the end it can be a big size and so spread of 2 pips can happen very easy.
Lot size is automatically calculated based on stop loss.
In backtest, I'm using 3% per trade meaning every trade that I lost will be 3% of my balance.

In demo, I have halve it for most strategies, other than those with very low DD. Max will remain as 3%
  • #45
  • Jul 5, 2018 5:25am Jul 5, 2018 5:25am
Quoting DonaldDuckie
{quote} Lot size is automatically calculated based on stop loss. In backtest, I'm using 3% per trade meaning every trade that I lost will be 3% of my balance. In demo, I have halve it for most strategies, other than those with very low DD. Max will remain as 3%
yes, but how big was your lotsize on the end in your backtest?
  • #46
  • Jul 5, 2018 5:28am Jul 5, 2018 5:28am
Quoting dukas_trader
{quote} yes, but how big was your lotsize on the end in your backtest?
Good question. I didn't really noticed.

I do understand that some brokers will have a max cap for lot and that if the size is too big, it's difficult to fill the price.
That will be a good problem to have down the road and I hope to reach there.

I will take a look when I backtest next time.
  • #47
  • Jul 5, 2018 5:38am Jul 5, 2018 5:38am
Quoting DonaldDuckie
{quote} Good question. I didn't really noticed. I do understand that some brokers will have a max cap for lot and that if the size is too big, it's difficult to fill the price. That will be a good problem to have down the road and I hope to reach there. I will take a look when I backtest next time.
in backtest you need to go throuh any trade and watch at important points:
- miss quotes (wrong quotes from the broker, when you have traades at crazy prices so and fast win or loss)
- weekends (after weekend you often have big spread and with only bid chart you have fast win or loss that not there in real live bid/ask prices
- news (same like weekend with big spreads and so fast out of a trade, but isnet in real)
- rollover time (same like news)
  • #48
  • Jul 5, 2018 6:24am Jul 5, 2018 6:24am
Quoting dukas_trader
{quote} in backtest you need to go throuh any trade and watch at important points: - miss quotes (wrong quotes from the broker, when you have traades at crazy prices so and fast win or loss) - weekends (after weekend you often have big spread and with only bid chart you have fast win or loss that not there in real live bid/ask prices - news (same like weekend with big spreads and so fast out of a trade, but isnet in real) - rollover time (same like news)
As i'm pretty new to mt4, only for around 2 months. i'm using the below simple code to prevent getting into trades when spread is high
(Ask-Bid)<0.00025

I do understand that once i'm in the trade, there is a chance that I get stop out due to sudden spread increase which I do not have any solutions.
  • #49
  • Jul 5, 2018 7:14am Jul 5, 2018 7:14am
Still short of less than 5 pips for TP.
2 more mins to Nonfarm but well, i'm not going to intervene my trade.
  • #50
  • Jul 5, 2018 7:17am Jul 5, 2018 7:17am
Quoting DonaldDuckie
Still short of less than 5 pips for TP. 2 more mins to Nonfarm but well, i'm not going to intervene my trade.
......short of 0.1 pips to TP.....
I'm going crazy soon.
  • #51
  • Jul 5, 2018 7:24am Jul 5, 2018 7:24am
Finally TP after 7 hours.
Entry was good which is programmed in the EA.
Exit point was fixed number of pips which is determined by 18 years backtest.
Almost a perfect trade if TP was 2 pip lower

Attached Image (click to enlarge)
Click to Enlarge Name: EURUSD Trade.PNG Size: 76 KB
  • #52
  • Jul 5, 2018 7:49am Jul 5, 2018 7:49am
Quoting DonaldDuckie
{quote} As i'm pretty new to mt4, only for around 2 months. i'm using the below simple code to prevent getting into trades when spread is high (Ask-Bid)<0.00025 I do understand that once i'm in the trade, there is a chance that I get stop out due to sudden spread increase which I do not have any solutions.
this will not work in backtest, it will do any trade, also all fake trades!
  • #53
  • Jul 5, 2018 8:04am Jul 5, 2018 8:04am
Quoting dukas_trader
{quote} this will not work in backtest, it will do any trade, also all fake trades!
Is there any suggestion how should I code instead?
  • #54
  • Jul 5, 2018 8:11am Jul 5, 2018 8:11am
Quoting DonaldDuckie
{quote} Is there any suggestion how should I code instead?
that is one of the bascis to learn first, what a backtest can do and what not in metatrader!
you cant do differnet (backtester always takes fixed spread what you put in from beginning of the test and not the real that was at the moment of the trade, here are so many points you need to look over by hand later:
- 1 min bid charts gives you always wrong entries and exits because you dont have tickchart and ask
. weekend, rollover, news, flash crash, give you small losses what would be much higher in real and big wins what would not be there in real......
- wrong quotes give wrong wins or losses
- amount available in market will not be checked
- and so on. and so on.....

thatswhy a backtest in mt4 is very easy to manipulate, easy you can have billions wins in some years from a very small account, when you simple use this mistakes or dont check that you trade with this many mistakes inside backtester. thats why the chance your backtest is wrong is 99,99% at the moment.

best sugesstion: learn what advantages and disadvantages the metatrader backtester has, programm like you traded manuell and control each trade with your manuell trades, and you will see very fast how different all is. anyone can has extrem good backtest, very very easy. control each trade in visuell chart, only then its fair backtest (and of course know to what you look for to find the problems).

that you dont find so many backtests online with extrem good result is, because anyone has this backtests, and all know this are fake tests because of the problems inside the metatrader backtester. you can program many things so that many disadvantages are away, but not all are possible to kill, rest you need to do by hand by controlling each trade in visuell chart and best by broker with ask and bid chart same time for problematic trades.......
doing correct backtest manuell and doing correct backtest of program is also one topic of learning the basics first. and important in your case, because when done wrong you cheat yourself with manuell and automatic backtests.
  • #55
  • Jul 5, 2018 8:49am Jul 5, 2018 8:49am
Quoting dukas_trader
{quote} that is one of the bascis to learn first, what a backtest can do and what not in metatrader! you cant do differnet (backtester always takes fixed spread what you put in from beginning of the test and not the real that was at the moment of the trade, here are so many points you need to look over by hand later: - 1 min bid charts gives you always wrong entries and exits because you dont have tickchart and ask . weekend, rollover, news, flash crash, give you small losses what would be much higher in real and big wins what would not be there...
I understand most of your points and trying to minimise risk and errors

  1. I choose to use fixed TP and SL because I think this is the safest way as it eliminate the risk of network and computer issue or less affected by black swan events.
  2. I choose to use at least a 5 mins chart as i'm aware that 1 min chart is not accurate during backtest.
  3. I don't expect real account to perform as good as backtest but at least I need to make sure that it's decent. In fact from back test, i'm looking for 2 things which is how to filter trend and the value of TP and SL which is most profitable. In mathematical sense, i'm trying to find the center point of a normal distribution, make that my center so that effectively increasing the CPK value.

The important thing is not about how much I can win per pair but rather how much i can win combined together.
As long as each pair is profitable, when adding together, it will be huge.

Example if I have 3 pairs which give me 300% win in 10 years.
Total win is not 300%+300%+300% (900%).
It is actually closer to 300% x 300% x 300% (2700%) provided all 3 do not drawdown at the same time and crashes the account.

This is the reason why i'm quite particular in DD and always trying to add more pairs to my trade.

I'm also looking very closely at my demo account trades to make sure that everything is on track.

  • #56
  • Jul 5, 2018 10:03am Jul 5, 2018 10:03am
Quoting DonaldDuckie
{quote} I understand most of your points and trying to minimise risk and errors I choose to use fixed TP and SL because I think this is the safest way as it eliminate the risk of network and computer issue or less affected by black swan events. I choose to use at least a 5 mins chart as i'm aware that 1 min chart is not accurate during backtest. I don't expect real account to perform as good as backtest but at least I need to make sure that it's decent. In fact from back test, i'm looking for 2 things which is how to filter trend and the value of...
1) fix Sl and TP dont help anything. only for connection problems, the backtest will be wrong same.
2) 5 min is same problem then 1 min, 5 min is simple calculated from 1 min bars. 1 min is even the best.
3) real account can be extrem in minus and same time backtest can win millions, that happens very often,. because of the big problems in backteste, same happens when trading in demoaccount, there can also be millions and live account later losses. because demoaccount has many many advantages against real account.

and your calculation is wrong:
when you trade 3 pairs, you can only trade 1/3 of trading amount in each pair. so having traded one pair with 3 times the amount you would have same 2700%. so it helps nothing for getting more win, only when pairs are total unconnected, so dont use gbpusd and eurusd, because in both usd in connected... and so on, then you have a diversification in risk a little bit, but same time increase chance of black swan event because you have more currencies inside who can be affected..... depends all how you prepare against all as one structure...
  • #57
  • Jul 5, 2018 10:25am Jul 5, 2018 10:25am
The backtest results are NOT representative.

Look at your modeling quality.

You NEED to use tickdata for your backtests (with at least 90% modeling quality).

You should also test with variable spreads and slippage.

So do NOT trade that.

The results will NOT be the same.

Garbage in, garbage out.

Stay safe,
FF
www.sme-fx.com
  • #58
  • Jul 5, 2018 10:29am Jul 5, 2018 10:29am
Quoting FractalFreak
The backtest results are NOT representative. Look at your modeling quality. You NEED to use tickdata for your backtests (with at least 90% modeling quality). You should also test with variable spreads and slippage. So do NOT trade that. The results will NOT be the same. Garbage in, garbage out. Stay safe, FF
thats also not correct. you need to learn what backtester can do and what he cant, advantages and disadvantages, your post is total incomplete (and yes, same your manipulation thread, exact same problem,and by the way just for you: faking and cheating is not trading!).

there are trading styles who can be backtested with metatrader backtester without any problem, other cant. depends on programming skills and trading system.
and variable spread is useless, you have to use the correct spread that was at this time, this can be done only by manuell improving or programming a own backtester or ea programm that rebuilds the result correct.
  • #59
  • Jul 5, 2018 11:24am Jul 5, 2018 11:24am
Once again, you are WRONG wannabe-fund-manager.

Again you try to spread MIS-INFORMATION!

Which brokers pay you for that?

You were exposed enough.

WITHOUT proper TICK DATA backtesting capabilities, the newer MT4s will SIMULATE tick movements with a GENERIC algo.

So therefore, the output is completely MISLEADING SINCE THE ACTUAL INNER CANDLE MOVEMENT IS SIMULATED WITH AN ALGO.

DO not try to talk this down.

The backtests are INCORRECT and NOT representative in ANY way.

Every professional trader who has invested thousands of hours into TICK DATA backtests with all kinds of parameters such as variable spreads, slippage, commissions etc. KNOWS that.
www.sme-fx.com
  • #60
  • Edited 11:38am Jul 5, 2018 11:25am | Edited 11:38am
YES, you need SOFTWARE if you want to conduct TICK DATA tests in Mt4.

Do not listen to these ppl.

They have NO clue and try to MISGUIDE you!


Stay safe,
FF
www.sme-fx.com
0 traders viewing now