plateman's CCI (aka SuperTrend) auto-trading robot by Steve Hopwood | Page 2 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Back and forward testing EAs aka Trading Robots 6 replies

Supertrend and MTF supertrend IDEA 11 replies

Need code CCI channel, CCI with SAR if you can free coding 0 replies

automatic forex trading software battle aka comparison 0 replies

Issues about Market Depth, AKA LEVEL II 1 reply

Attachments: plateman's CCI (aka SuperTrend) auto-trading robot by Steve Hopwood
Exit Attachments

plateman's CCI (aka SuperTrend) auto-trading robot by Steve Hopwood

  • #21
  • Dec 1, 2010 1:20pm Dec 1, 2010 1:20pm
I'm not flaming or insulting anyone, just take it easy, but the system/methode here is CCI, nothing else. Let's work around it, let's not start adding filters, indicators, ideas and other stuff. You guys already know that the stuff we are aware of does not work, otherwise we wouldn't be here discussing this EA.

To me, the indicator is good, the EA is good, but not enough, we should work on exit strategies and money management.

I'll throw in an idea, what do you think is the best setting as stops/targets? fixed pips? atr? support/resistance? let's experiment this kind of things.

edit:rereading everything from the beginning, no need of stops and targets, my apologize. lol
then i'll think about it, however we have to concentrate on one aspect of the method at a time, or so i think.
  • #22
  • Dec 1, 2010 1:42pm Dec 1, 2010 1:42pm
Quoting KingHigh
Steve,

A single bar move of the CCI to below 0 and back to above 0 will not display using a line indicator. If you change the DRAW_LINE to DRAW_ARROW, you will see a RED arrow at the trade location.

BUT -- the EA should have closed the sell and opened a buy on the next bar. This is the mystery.

No guarantees on my assessment but this is what I have seen many times with Indies.

With Regards,

KingHigh
KH, yes I agree that the short should have been closed on the next candle, in fact it carried on throughout the day until I closed it manually late this afternoon when it was over 70 pips down. Of course that is why we demo test first and if this happens regularly I'm sure Steve will do a fix.
  • #23
  • Dec 1, 2010 1:47pm Dec 1, 2010 1:47pm
Steve,
One more comment....may I make a suggesting regarding stops....can the initial stop be "X", and then have it move according to supertrend. I believe if we use a 3-bar flat line, this creates the new support level. The stop is the prior level. A pic might show this better.

Attached Image


Cheers
CanuckCT
  • #24
  • Dec 2, 2010 3:56am Dec 2, 2010 3:56am
EA is performing pretty well except for one flaw which can be seen illustrated on the two charts below. When the indicator line changes colour permanently the bot works fine but when the colour changes and then changes back within one candle the EA is left pointing the wrong way. I'm wondering if this could be fixed by allowing a trade at any time during a candle, maybe a risk of a few whipsaw trades but at least it wouldn't be left on the wrong side of a big move.

I don't know whether this would be possible in coding, what do you think Steve would this work any better?

As I type trading 0.4 lots on 6 charts, banked £445 in last 24 hrs.
Attached Images
  • #25
  • Edited 9:30am Dec 2, 2010 4:33am | Edited 9:30am
Quoting plateman
EA is performing pretty well except for one flaw which can be seen illustrated on the two charts below. When the indicator line changes colour permanently the bot works fine but when the colour changes and then changes back within one candle the EA is left pointing the wrong way. I'm wondering if this could be fixed by allowing a trade at any time during a candle, maybe a risk of a few whipsaw trades but at least it wouldn't be left on the wrong side of a big move.

I don't know whether this would be possible in coding, what do you think Steve would...
You are not listening to me Mike, so I am going to shout it this time.

THE ROBOT DOES NOT CONSULT THE SUPERTREND INDICATOR

Anybody here happen to have the code to make the above flash on and off, whilst emitting a banshee scream to attract attention?

The bot already takes trades at any point that it detects a trigger. That is probably why you get rogue trades. More likely is that the bot needs to look further back in time, at:

 

  1. CCI now
  2. CCI at the close of the previous candle
  3. CCI at the close of 2 candles ago

If 1 & 2 agree and are in the opposite direction to 3, then take the trade. Same criteria in reverse for exit.

The attached works this way. See if it is more profitable.

Attached File(s)
File Type: mq4 plateman's SuperTrend auto trading robot by Steve Hopwood.mq4   44 KB | 1,660 downloads
  • #26
  • Dec 2, 2010 4:46am Dec 2, 2010 4:46am
Steve, :>))))))), sorry about that, just thinking aloud really, and forgetting how it actually works.

Thanks for the mod, will set it up and test.
  • #27
  • Dec 2, 2010 9:13am Dec 2, 2010 9:13am
Steve, on each chart the values displayed on the screen for "CCI 1 candle ago" and "CCI 2 candles ago" are exactly the same, is this OK?
  • #28
  • Dec 2, 2010 9:30am Dec 2, 2010 9:30am
Hi all.

Might be useful filtering the entries by an additional CCI(50) drawn in its default window, specifically:
- wait for SuperTrend (ST) to go green then enter long when CCI goes above 100, exit when it pops back below 100;
- wait for SuperTrend (ST) to go red then enter short when CCI goes below -100, exit when it pops back above -100.

Looks quite ok on today's M1 EURUSD.
  • #29
  • Dec 2, 2010 9:32am Dec 2, 2010 9:32am
Quoting plateman
Steve, on each chart the values displayed on the screen for "CCI 1 candle ago" and "CCI 2 candles ago" are exactly the same, is this OK?
That is one of my copy/paste bloops. I have replaced the version a couple of posts ago with the fix; it is only the display that is affected.

  • #30
  • Edited 10:44am Dec 2, 2010 10:27am | Edited 10:44am
Settings with results...CAVEAT! Mt4 test parameters N/A
Attached Image(s) (click to enlarge)
Click to Enlarge Name: bb.gif Size: 14 KB
Click to Enlarge Name: TesterGraph.gif Size: 23 KB
Vincit qui se vincit.
  • #31
  • Dec 2, 2010 11:12am Dec 2, 2010 11:12am
trying to find settings to improve BOT's W/L pct. Perhaps entering on close will eliminate some of the whipsaw.
Vincit qui se vincit.
  • #32
  • Dec 2, 2010 11:14am Dec 2, 2010 11:14am
Thanks for running the test Pippo, nice results, I can see that you like the D1 timeframe! In round figures about 17 trades p.a. giving a profit of 8k p.a., definitely not to be sniffed at.
  • #33
  • Dec 2, 2010 11:26am Dec 2, 2010 11:26am
Quoting plateman
Thanks for running the test Pippo, nice results, I can see that you like the D1 timeframe! In round figures about 17 trades p.a. giving a profit of 8k p.a., definitely not to be sniffed at.
Also have it running on 24 pairs at H4.
Vincit qui se vincit.
  • #34
  • Dec 3, 2010 2:50am Dec 3, 2010 2:50am
Quoting SteveHopwood
...

  1. CCI at the close of the previous candle
  2. CCI at the close of 2 candles ago

If 1 & 2 agree and are in the opposite direction to 3, then take the trade. Same criteria in reverse for exit.

The attached works this way. See if it is more profitable.

The CloseTrade test for shorts is borked, has the same criteria as for longs and doesn't test for the 2nd shift.
Can't attach a corrected version, being a junior member and all
  • #35
  • Dec 3, 2010 3:31am Dec 3, 2010 3:31am
Hi Rofleus, perhaps you could post your suggested revised code.
  • #36
  • Dec 3, 2010 3:42am Dec 3, 2010 3:42am
Quoting Rofleus
The CloseTrade test for shorts is borked, has the same criteria as for longs and doesn't test for the 2nd shift.
Can't attach a corrected version, being a junior member and all
Great bit of proofreading. Cheers.

Fix attached.

Attached File(s)
File Type: mq4 plateman's SuperTrend auto trading robot by Steve Hopwood.mq4   45 KB | 1,691 downloads
  • #37
  • Dec 3, 2010 4:15am Dec 3, 2010 4:15am
Paulus used to use this strategy very successfully manually. His sl was always 2 flat levels behind current.
  • #38
  • Edited 4:31am Dec 3, 2010 4:29am | Edited 4:31am
which robot is good it is depend on person like what kind of timeframe he or she want to use and what risk level he or she like. Every robot is good if they are properly set. it is mathamatics calculation. So when the robot do its job, it is well to check it up.
  • #39
  • Dec 3, 2010 5:12am Dec 3, 2010 5:12am
Tested on EURUSD H1, 0.1 size, from 1st Jan to 1st Dec 2010:

-entry/exit at +1/-1 CCI yielded a $2000 profit for a $710 max DD
-entry at +100/-100, exit at +1/-1 CCI yielded a $2400 profit for a $811 max DD

Any attempt at using TP, SL, trailing, fails miserably.
Lower timeframes pretty much suck (also with TP SL or trailing).

On the other hand, I've been using the strategy manually with success on M1. The difference is being careful around pivots and sweet spots, trading only towards H1 or D1 trends and also taking profit sooner than the EA's closing criteria (i.e. when approaching pivot or when I feel it fell enough, don't ask )
I'm sure someone could code all these filters (except my guts ofc).
I'll keep using it manually and keep you posted guys (yesterday took 7 trades, all positive, 31 pips, today took 5, 1 closed at -5, the rest +29, had to help my wife with more important stuff like vacuuming )
Attached File(s)
File Type: zip testing.zip   30 KB | 939 downloads
  • #40
  • Dec 3, 2010 10:25am Dec 3, 2010 10:25am
Quoting Rofleus
Tested on EURUSD H1, 0.1 size, from 1st Jan to 1st Dec 2010:

-entry/exit at +1/-1 CCI yielded a $2000 profit for a $710 max DD
-entry at +100/-100, exit at +1/-1 CCI yielded a $2400 profit for a $811 max DD

Any attempt at using TP, SL, trailing, fails miserably.
Lower timeframes pretty much suck (also with TP SL or trailing).

On the other hand, I've been using the strategy manually with success on M1. The difference is being careful around pivots and sweet spots, trading only towards H1 or D1 trends and also taking profit sooner than the EA's...
did you enter trades round the clock?
Vincit qui se vincit.
0 traders viewing now