Follow the BIG Candle | Page 4 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Follow Last Candle EA 10 replies

Follow the prior Heiken Ashi candle for profits 61 replies

How to make EA that send Open Price of Candle for every new candle 5 replies

Day's first H4 candle correlation to daily candle 14 replies

Correlation between first 4 hour candle and daily candle 1 reply

Attachments: Follow the BIG Candle
Exit Attachments

Follow the BIG Candle

  • #61
  • Mar 2, 2014 3:54am Mar 2, 2014 3:54am
Few comments on the above test results:

  1. I tested the EA initially for GBP/USD (hence the best results), although I didn't over-optimize.
  2. Then I applied the same settings (no change what so ever), to the other pairs, and interestingly all pairs ended with profit. Not amazing profits but at least no loss.
  3. Notice that other pair results can highly improve if we choose the "correct" parameter settings (especially ATR multiples). And this is normal in my opinion, compared to different behavior of each pair.
  4. The EA made 47% ROI ($32,440) over 14 months on total initial investment of $70,000.
  5. The test was done on a short-term time frame (M30), which is some what not reliable to catch big moves.

I encourage people with Good Historical Data to test it on different time frames and longer periods 3 to 5 years of data.

Regards,

No guts, no glory
  • #62
  • Mar 2, 2014 4:35am Mar 2, 2014 4:35am
Hi Abokwaik

Is it possible to get the following data

One a big bar is triggered the average pull back against the trade before it comes into profit

I have attached a screen shot the bars may not be 'system designed big bars' they are used for example only

Each of the hi lighted signals then go into draw down before the significant move happens and Im assuming before the TP is hit

Thanks P
Attached Image (click to enlarge)
Click to Enlarge Name: big bar.png Size: 70 KB
  • #63
  • Mar 2, 2014 5:09am Mar 2, 2014 5:09am
Quoting Paulus
Hi Abokwaik Is it possible to get the following data One a big bar is triggered the average pull back against the trade before it comes into profit I have attached a screen shot the bars may not be 'system designed big bars' they are used for example only Each of the hi lighted signals then go into draw down before the significant move happens and Im assuming before the TP is hit Thanks P {image}

Sure. I ran the EA against your chart. Results are attached along with parameters used. Please note that as you go with higher time frames ATR values will be smaller. So the parameters used here are smaller that those for M30 time frame.
Attached Image(s) (click to enlarge)
Click to Enlarge Name: gbpusdrh1.png Size: 37 KB
Click to Enlarge Name: FBG_GU_H1.PNG Size: 40 KB
Attached File(s)
File Type: rar fbg_03_gu_h1.rar   < 1 KB | 722 downloads
No guts, no glory
  • #64
  • Mar 2, 2014 5:14am Mar 2, 2014 5:14am
Now I dont want to rock the boat here your far down the line on this one with very good results..Im just throwing odd things in that may gro they may not

But

On the same chart I posted that the range of the big candle and divide by 2 so if the bar is 60 pips hi to lo then we have a value of 30 pips thats our stop
From this we can work out 5% of account .... now the TP is the hi of the down big bar the lo of an up big bar

If the TP is hit we get 2 x our risk and our stop loss is 50% of the bib bar candle.

This gives us RR of 2;1 and works on the principle that once the big move is over the price retraces to its previous level. But I think it would have to be a news driven big bar you need a big bar that is injected quickly through massive buying selling but doesnt effect the sentiment of the market to much

Only a thought if 2 much off subject please delete post or ask me to delete no offence taken ;-)
Attached Image (click to enlarge)
Click to Enlarge Name: bb.png Size: 61 KB
  • #65
  • Mar 2, 2014 5:18am Mar 2, 2014 5:18am
@Paulus:

Please note that the last short trade was closed for partial profit $236 in the attached chart. and the remaining position closed by trailing stop on Feb.5th with $570 profit.
No guts, no glory
  • #66
  • Mar 2, 2014 5:31am Mar 2, 2014 5:31am
Sorry I got engaged in a meeting. I will reply as soon as I finish the meeting.
No guts, no glory
  • #67
  • Mar 2, 2014 5:59am Mar 2, 2014 5:59am
Quoting Paulus
Now I dont want to rock the boat here your far down the line on this one with very good results..Im just throwing odd things in that may gro they may not But On the same chart I posted that the range of the big candle and divide by 2 so if the bar is 60 pips hi to lo then we have a value of 30 pips thats our stop From this we can work out 5% of account .... now the TP is the hi of the down big bar the lo of an up big bar If the TP is hit we get 2 x our risk and our stop loss is 50% of the bib bar candle. This gives us RR of 2;1 and works on the...

OK. I didn't understand exactly what do you mean. However I've posted the rest of the chart the shows the last deal. I tried to change the colors to match yours for easy Comparing.

This strategy does not take news or fundamentals into account, although these factors will affect the behavior. It is based mainly on price action.

EDIT: re-posted the chart as it was not complete.
Attached Image (click to enlarge)
Click to Enlarge Name: gbpusdrh1_2.png Size: 41 KB
No guts, no glory
  • #68
  • Mar 2, 2014 6:16am Mar 2, 2014 6:16am
@ Paulus:

Further to my reply above, the EA is using ATR multiples (not fixed pips), and Reward_to_Risk (RR) ration is usually much bigger that 2:1. Basically we are looking to capture evolving trends triggered by Big Candles whenever that is possible.

Since we are using ATR, it means that the big candle does not have to be of a fixed size (i.e. 100 pips) during the EA lifetime. No, it will change according to price momentum. This was clear in my earlier posts # 56 & 57.
No guts, no glory
  • #69
  • Mar 2, 2014 10:06am Mar 2, 2014 10:06am
Hi who can post the strategy as mq4 file because i want test it on jforex and dukascopy has only a funktion to compile mq4 files in java files and not ex4 files

is that possible?
  • #70
  • Mar 2, 2014 12:14pm Mar 2, 2014 12:14pm
i don't know how exactly set the best parameter for BC3, been trying to make some 'extre3m' set, because i want more position in short time and in short TF.. here's..

Attached Image (click to enlarge)
Click to Enlarge Name: BC3.png Size: 41 KB
After Highest Top is start Downward, and Vice Versa
  • #71
  • Edited 1:55pm Mar 2, 2014 1:41pm | Edited 1:55pm
Quoting ZiC
i don't know how exactly set the best parameter for BC3, been trying to make some 'extre3m' set, because i want more position in short time and in short TF.. here's.. {image}

Thank you for your continous contributions. I noticed that Modeling Quality is very low (around 35%), this should at least 80%.

As for parameter settings, it should be something that makes sense. for example its reasonable to use 1.5 or 2 ATR multiples on Daily timeframe, while it is not reasonable to use less than 3 on H4, 4 on H1 and 5 on M30. and so on.

Also SL should be at least equal to ATR multiples (at least to protect the trigger candle). Trailing Stop should be greater or eaual to SL, and so on.

To catch big profits we need to make the TP very high, and exit the trade profitably at the trailing stop instead. The introduction of partail profit taking had made a difference in the results. Partial take profit should be at least as SL (or BreakEven Stop -- introduced in V3). Using larger value for trailing stop will enhance the performance.

Out of my tests, especially on smaller time frames: Opening on the next candle resulted better that Opening on Current Candle.

I my self did not search for the Ultimate Parameters, as I am still working on the Core Idea.

The ideal output that we should be looking for is small losses, huge profits, minimum drawdown, not large number of trades. as the added filter will try to take only the best (A or A+) setups.

Regards,
No guts, no glory
  • #72
  • Mar 3, 2014 12:28am Mar 3, 2014 12:28am
Anyone tested this? http://eareview.net/tick-data-suite
Please excuse the bad English via Google Translate.
  • #73
  • Mar 3, 2014 12:39am Mar 3, 2014 12:39am
Some notes on using the "Candle Effective Size" parameter :

The "Candle Effective Size" parameter should be set along with "Candle Period" parameter set to 1, otherwise make it 0.
Meaning that if we are opening our trades based on the candle Next to the Big Candle, we need the Big Candle to be a Healthy Candle. If the Big Candle is a Bull candle we want to be sure that Bulls are in control, and visa versa.

A Healthy candle should have an effective size of at least 1/2 of the candle size. Or in other terms if we set atr_multiple parameter which identifies the Big Candle to 5, then the Candle Effective Size should be at least set to 3.

Chart Example: The denoted candle on the chart is a Big "Bull" Candle, however the effective size of the candle "Close-Low" is less than 1/3 of the candle size. Before adding this parameter the system would have traded this candle which might resulted a loss or a large Draw Down.

I Hope that the concept is clear.
Attached Image
No guts, no glory
  • #74
  • Mar 3, 2014 4:12am Mar 3, 2014 4:12am
Hello guys.

Im new here and i saw this big candle ea. I might try this on my micro account on live after i have done some testing. Do i keed anything else than ATR and this EA?
  • #75
  • Mar 3, 2014 4:51am Mar 3, 2014 4:51am
Quoting medIce
Hello guys. Im new here and i saw this big candle ea. I might try this on my micro account on live after i have done some testing. Do i keed anything else than ATR and this EA?
No. Even ATR indicator is not needed as it is included in the EA. However you might add it to the chart for visual following.

No live trading please as this is still work in progress.

For example I discovered a bug in position size calculation in case your using risk_perc parameter instead of fixed_lots on 5-digit brokers. This will be fixed in the new release.
No guts, no glory
  • #76
  • Mar 3, 2014 5:01am Mar 3, 2014 5:01am
Quoting abokwaik
{quote} No. Even ATR indicator is not needed as it is included in the EA. However you might add it to the chart for visual following. No live trading please as this is still work in progress. For example I discovered a bug in position size calculation in case your using risk_perc parameter instead of fixed_lots on 5-digit brokers. This will be fixed in the new release.
He can you post your new release as a mq4 file? for me is it then possible to test it on dukascopy history tester
  • #77
  • Mar 3, 2014 5:46am Mar 3, 2014 5:46am
Quoting abokwaik
{quote} No. Even ATR indicator is not needed as it is included in the EA. However you might add it to the chart for visual following. No live trading please as this is still work in progress. For example I discovered a bug in position size calculation in case your using risk_perc parameter instead of fixed_lots on 5-digit brokers. This will be fixed in the new release.
Yeah. I will test it with demo first. And when you think its ready i will use it first only on my micro account.
  • #78
  • Mar 3, 2014 5:53am Mar 3, 2014 5:53am
Quoting klugec
{quote} He can you post your new release as a mq4 file? for me is it then possible to test it on dukascopy history tester
I am speechless.

I might do that depending on member contributions.


Any way the strategy is simple and any coder should be able to replicate it on jforex.
No guts, no glory
  • #79
  • Mar 3, 2014 6:45am Mar 3, 2014 6:45am
(Dynamic ATR multiples) vs (Static ATR multiples)

So far we were using Dynamic ATR multiples. Meaning : suppose we had the following setup:

  1. ATR period = 21 on M30 timeframe
  2. ATR multiples = 5
  3. ATR Stop Loss = 5
  4. ATR Trailing Stop=5
  5. ATR Break Even=3
  6. ATR MA partial Take Proft =5

Now suppose ATR(21) was 10 pips. and a Bull candle appeared of size 60, so we took the Long trade (60>5x10).
Our Stop Loss (5x10)=50 pips, and so TS=50 pips, BE=3x10=30pips, PTP=50pips.
After 6 hours the ATR(21) became 7 pips. So since we are using Dynamic ATR values, the TS will become 5x7=35pips, and so PTP. BE will become 7x3=21.

I don't like this. So may be its better to keep these values as is (Static) when we opened the trade. And they will get re-calculated only for a new trade. I will try to add a parameter for that in the next release and compare the results.

Opinions required.

Reagards,

No guts, no glory
  • #80
  • Mar 4, 2014 1:39am Mar 4, 2014 1:39am
Are ATR values reliable ?

Remember we are using ATR to decide on which is the Big Candle over a period.
I ran some tests on GBP/USD H1 and H4 time frames for the period of Jan 1,2014 until Mar 1,2014.
All parameters were fixed except ATR Period varying from 10 to 100 with step of 10.

On H1 results were as follows:

ATR Period _____ Trades _____ Profit _____ Drawdown%
____________________________________________________
10 _______________ 5 ________ 2,289 _____ 17.81%
20 _______________ 4 ________ 2,416 _____ 34.76%
30 _______________ 4 ________ 2,261 _____ 34.59%
40 _______________ 9 ________ 2,999 _____ 33.14%
50 _______________ 8 ________ -3,594 _____ 57.59%
60 _______________ 8 ________ 2,885 _____ 33.02%
70 _______________ 8 ________ 2,907 _____ 33.04%
80 _______________ 8 ________ -3,544 _____ 56.87%
90 _______________ 5 ________ 2,306 _____ 35.89%
100 ______________ 7 ________ 2,557 _____ 34.00%

On H4 results were as follows:

ATR Period _____ Trades _____ Profit _____ Drawdown%
____________________________________________________
10 _______________ 5 ________ 2,018 _____ 14.06%
20 _______________ 6 ________ 2,603 _____ 20.01%
30 _______________ 4 ________ 2,089 _____ 20.22%
40 _______________ 1 ________ -4,028 _____ 52.31%
50 _______________ 6 ________ 2,678 _____ 20.37%
60 _______________ 6 ________ 2,646 _____ 20.39%
70 _______________ 4 ________ 2,090 _____ 20.31%
80 _______________ 4 ________ 2,072 _____ 20.29%
90 _______________ 6 ________ 2,690 _____ 20.38%
100 ______________ 4 ________ 2,103 _____ 20.33%

As you see all results are consistent except 1 or 2 on each time frame, and that worries me. We need a way to overcome this issue as the selection of ATR period will highly affect the results.

What if we take an average of ATR over different ATR periods ? I'gonna try that.

Any other ideas how can overcome this problem ?

EDIT:re-formatted the data.
No guts, no glory
0 traders viewing now