Old Dog with New Tricks | Page 38 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Old Dog's Taming of the Beast? 34 replies

Old Dog Fib Trading - EA Development 18 replies

Old Dog Fib Trading - EURUSD 399 replies

5 Tricks of Trend Trading, by Rob Booker 25 replies

How To Work Around A Market Maker's Tricks 0 replies

Attachments: Old Dog with New Tricks
Exit Attachments

Old Dog with New Tricks

  • #741
  • Aug 10, 2010 3:58pm Aug 10, 2010 3:58pm
Quoting stevegee58
Hey Ron, is the MinRiskReward setting really risk/reward? Or is it reward/risk?

Phrased another way, if you want to risk 1 to make 2, is the MinRiskReward 2 or 0.5?
The formula is reward / risk.

So 2 in your example.
  • #742
  • Aug 10, 2010 6:00pm Aug 10, 2010 6:00pm
Quoting Ronald Raygun
I don't think so... unless the bar updates with refreshrates(). There's only one way to find out.
I've attached it to a 2R Chart, guess I'll find out.

This is what the writing in the top left looks like.

No idea if that means it will trade or not.
Attached Image
  • #743
  • Aug 10, 2010 6:08pm Aug 10, 2010 6:08pm
Well it worked ok on the offline 2R chart

Trading the exact same trade i was considering this morning.
Attached Image (click to enlarge)
Click to Enlarge Name: 0.jpg Size: 238 KB
  • #744
  • Aug 10, 2010 6:37pm Aug 10, 2010 6:37pm
Hit the 10 Pip TP target no problem.
Attached Image (click to enlarge)
Click to Enlarge Name: 0.jpg Size: 232 KB
  • #745
  • Aug 10, 2010 6:54pm Aug 10, 2010 6:54pm
2x10 pip winners.

According to the backtest the win ratio is 54.15%
Attached Image (click to enlarge)
Click to Enlarge Name: 0.jpg Size: 236 KB
  • #746
  • Aug 10, 2010 7:45pm Aug 10, 2010 7:45pm
Seems breakouts work well when a strong trend in place eh...

7 winners 1 loser

I'm guessing once a range begins it will even up somewhat.
Attached Image (click to enlarge)
Click to Enlarge Name: 0.jpg Size: 231 KB
  • #747
  • Aug 10, 2010 8:04pm Aug 10, 2010 8:04pm
Ronald: My post must have got lost a couple of pages back, but basically I think that there may be an issue with the successrate filter. I covered it in my post here.

Basically, the optimized filter has the following, for example:
4: successrate: 0.58
5: successrate: 0.65
6: successrate: 0.63

If you then apply a minimum successrate of 60, while changing nothing else, I get an optimized set full of -1 values. With this example set, hours 5 and 6 should be tradeable, should they not?
  • #748
  • Aug 10, 2010 8:45pm Aug 10, 2010 8:45pm
Quoting Mindsphere
Ronald: My post must have got lost a couple of pages back, but basically I think that there may be an issue with the successrate filter. I covered it in my post here.

Basically, the optimized filter has the following, for example:
4: successrate: 0.58
5: successrate: 0.65
6: successrate: 0.63

If you then apply a minimum successrate of 60, while changing nothing else, I get an optimized set full of -1 values. With this example set, hours 5 and 6 should be tradeable,...
As I mentioned to iulli, I'm not sure.

It seems that the backtesting data changed.

Here's a thought.

 

  1. Take a backtest/optimization.
  2. Go to the experts/files and back up the initial files.
  3. Rerun the optimization.
  4. Compare.

  • #749
  • Aug 10, 2010 8:54pm Aug 10, 2010 8:54pm
I was having the same problem the the success rate. I solved it by entering "1" for a success rating of 100, not "100". So for a success rating of 60, try "0.6".
  • #750
  • Aug 10, 2010 9:04pm Aug 10, 2010 9:04pm
I have a question,

Does this Green and Red Dotted Line indicate the zone by which the breakout strategy trades once price goes past these levels.

If so can you elaborate as to how these levels are determined, intraday or end of day? or both.
Attached Image (click to enlarge)
Click to Enlarge Name: 0.jpg Size: 223 KB
  • #751
  • Aug 10, 2010 9:21pm Aug 10, 2010 9:21pm
Quoting Musashi
I have a question,

Does this Green and Red Dotted Line indicate the zone by which the breakout strategy trades once price goes past these levels.

If so can you elaborate as to how these levels are determined, intraday or end of day? or both.
Green = Long
Red = Short

These are intraday levels
  • #752
  • Aug 10, 2010 9:28pm Aug 10, 2010 9:28pm
Quoting Ronald Raygun
Green = Long
Red = Short

These are intraday levels
So do these level move at all during the day or are they only determine via the current day H/L range.

At which point do they reset to the initial 10 Pip range?
  • #753
  • Aug 10, 2010 9:36pm Aug 10, 2010 9:36pm
Quoting Musashi

At which point do they reset to the initial 10 Pip range?
00:00
  • #754
  • Aug 10, 2010 9:45pm Aug 10, 2010 9:45pm
What do you consider a good success rate using the inbuilt back tester with a R:R of 1:1

Obviously the higher the better, however considering the back tester is not perfect is there a cut of level you would consider to small?

Or anything over 50.1% win ratio is ok for you?
  • #755
  • Aug 11, 2010 2:34am Aug 11, 2010 2:34am
@ronald
Optimizer :
Inserted Code
 if(TimeDayOfYear(Time[SearchShift]) != TimeDayOfYear(Time[SearchShift + 1]))
   {
   DayStartShift = SearchShift;
   }
 //Find the end of the range and establish initial high and low
 if(DayStartShift - SearchShift == InitialRange)
   {
   RangeEndShift = SearchShift;      
   HighShift = iHighest(SymbolUsed, 0, MODE_HIGH, DayStartShift - RangeEndShift, RangeEndShift);
   HighValue = iHigh(SymbolUsed, 0, HighShift);
   LowShift = iLowest(SymbolUsed, 0, MODE_LOW, DayStartShift - RangeEndShift, RangeEndShift);
   LowValue = iLow(SymbolUsed, 0, LowShift);
   }
Master function:
Inserted Code
int DayStart = iBarShift(NULL, 0, StrToTime("00:00"), False);
int RangeStart = DayStart - InitialRange;

//Determine Current Hilo
int HighShift =  iHighest(NULL, 0, MODE_HIGH, [color=Red]RangeStart[/color] - 1, 1);
int LowShift = iLowest(NULL, 0, MODE_LOW, [color=Red]RangeStart[/color] - 1, 1);

double HighPrice = iHigh(NULL, 0, HighShift);
double LowPrice = iLow(NULL, 0, LowShift);
Shouldn't that be DayStart instead of RangeStart? Because the optimizer is start calculating hilo from daystart. And the optimizer... feels weird you know . I'm still comparing generated csv data to understand how it works and how it suppose to works. I'll let you know if I found something. Btw, your MM lines doesn't include SL in it, is it ok? It seems trade doesn't close around percent loss MM value.

@musashi
Use attached file, excel 2007 format. It draws random equity curve expected from certain winrate, RR and risk. But remember, that any equity curve emulation derived from those value will be quite different from reality. In my experience, profit loss distribution / sequence is not a normal distribution one.
Attached File(s)
File Type: zip t-score.zip   363 KB | 417 downloads
  • #756
  • Aug 11, 2010 7:42am Aug 11, 2010 7:42am
Not quite.

Initial range is the number of bars after the day start that the EA waits before calculating the high/low. I generally wait an hour because it only takes an hour for the day's high and low to separate enough to minimize whipsaw.

If I started calculating the high / low from the day's start, then there is a chance the EA will ignore the initial range and trade during that time.
  • #757
  • Aug 11, 2010 11:05am Aug 11, 2010 11:05am
Quoting Rawkstar
I was having the same problem the the success rate. I solved it by entering "1" for a success rating of 100, not "100". So for a success rating of 60, try "0.6".
Hmm..mabe RR can weigh in on this.
Does it matter if it is set at 70 for 70% or should it be 0.7
  • #758
  • Aug 11, 2010 11:46am Aug 11, 2010 11:46am
Quoting unknown2u
Hmm..mabe RR can weigh in on this.
Does it matter if it is set at 70 for 70% or should it be 0.7
In successscore, it should be whole numbers. (0 to infinity)

In risk/reward, it should be decimals (1 being equal)

In win rate, it should be whole numbers. (0 to 100)
  • #759
  • Aug 11, 2010 2:46pm Aug 11, 2010 2:46pm
Quoting Ronald Raygun
In successscore, it should be whole numbers. (0 to infinity))
Hey RR, here's an example of using 90 and 0.9 to get a successscore of 90:

Could there be a bug in the code that multiplies the successcore by 100?
Attached Images
  • #760
  • Aug 11, 2010 2:54pm Aug 11, 2010 2:54pm
Quoting Ronald Raygun
In successscore, it should be whole numbers. (0 to infinity)

In risk/reward, it should be decimals (1 being equal)

In win rate, it should be whole numbers. (0 to 100)
RR

Checking the csv file it would appear that success score of 1 is equal to 100. For instance a win rate of 50% (or 0.5) and risk reward of 2.0 would equal 1.0 in the csv file. It appears on the trading screen as 100 so the ea must multiply this at some point.

Paul
0 traders viewing now