Pair Divergence Method | Page 40 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

What To Pair Divergence With? 14 replies

Divergence and hidden Divergence 122 replies

bo7a method... method for GBP/JPY 205 replies

Divergence method 11 replies

journal for pair divergence method 33 replies

Attachments: Pair Divergence Method
Exit Attachments

Pair Divergence Method

  • #781
  • Nov 23, 2009 9:16am Nov 23, 2009 9:16am
Quoting toddanderson
Could someone please add a moving average to the price differance indicator

I think it would be helpfull to see a bigging and end of a trend

i think you could drag a moving average holdinh CTRL to the Pricediff subwindow, no?
  • #782
  • Nov 23, 2009 9:39am Nov 23, 2009 9:39am
Quoting giraia_br
market is sideways.This is a trade you would take (if this bar close down)?

EU 15 min
No, because both the E/U and U/C are flat on the left when you look at the pip values. There is no 'peak'.
Hindsight is 20/20
  • #783
  • Nov 23, 2009 10:03am Nov 23, 2009 10:03am
Yes you can thank you


Quoting giraia_br
i think you could drag a moving average holdinh CTRL to the Pricediff subwindow, no?
  • #784
  • Edited Nov 24, 2009 8:21am Nov 23, 2009 12:07pm | Edited Nov 24, 2009 8:21am
Euro buy on the 1h.

edit: stopped out
Attached Image (click to enlarge)
Click to Enlarge Name: eurusd_buy_200911231800.gif Size: 16 KB
Hindsight is 20/20
  • #785
  • Nov 25, 2009 8:16am Nov 25, 2009 8:16am
I finished my March testing last night, and it was a real mess. I ended up with 52 trades: 21 wins, 23 losses, with the remainder break even. Average loss was 85% of the average win. This resulted in an expectancy of 0.033.

Bill's results were
Quote
March
30 W
26 L

Interesting how he had more trades in March than I did, but I had more trades in April than he did. I'm not too worried about it, and I will double check my results eventually.
Hindsight is 20/20
  • #786
  • Nov 29, 2009 12:36pm Nov 29, 2009 12:36pm
bookmarked
  • #787
  • Edited 8:55am Nov 30, 2009 8:44am | Edited 8:55am
I think there were a few trades over the US Thanksgiving holiday, but I didn't track them or include them in the back/forward test because I won't be trading that holiday in the future when I go live.

A triplet of trades this morning, two Euro sells were open when a Euro buy came up (I would have just closed the open trades, NOT stop and reverse), would have been an 8 pip loss (24% of a full loss) on the first, and 15 pip loss (34% of a full loss) on the second.

The 0900 trade is on the 30m because the peak that appears on the 15 is actually flat.
Attached Image(s) (click to enlarge)
Click to Enlarge Name: eurusd_buy_200911300745.gif Size: 19 KB
Click to Enlarge Name: eurusd_sell_200911300900.gif Size: 18 KB
Hindsight is 20/20
  • #788
  • Nov 30, 2009 8:47am Nov 30, 2009 8:47am
Oops! Another one in there too, U/C sell for a loss.
Attached Image (click to enlarge)
Click to Enlarge Name: usdchf_sell_200911300945.gif Size: 18 KB
Hindsight is 20/20
  • #789
  • Nov 30, 2009 2:55pm Nov 30, 2009 2:55pm
That wraps up a quiet November for my stats, 6 wins 7 losses, with average loss being 70% of the average win. Expectancy was 0.087, and trading 3%/trade would have net you 2.9% return on the month.

It is worthy to note that this is a circumstance where trading less % risk per trade would have left you with a greater return at the end of the month. Kelly Criterion in action!
Hindsight is 20/20
  • #790
  • Nov 30, 2009 8:42pm Nov 30, 2009 8:42pm
Hey guys. I have to appologize for being gone so long, but I got busy and then was kind of burnt out on metatrader programming.

Anyway, here are the new files. As a quick refresher, the Overlay is an indicator and the "Correlating Pair Divergence..." is an EA. I've modified the Overlay completely so it now bases everything on the current difference between prices (either by value in account currency or by pips). What this means is the Overlay should always be close to the chart line towards the right side of the chart, but may stray as you move further left. Also both will redraw at the start of each candle. This will keep everything matching up between the two. The Overlay also no longer prints the current candle, but that doesn't matter for this strategy.

Oh, I also added correlation percentage feature to the EA. It will display the current correlation between the two pairs (in the bottom left corner) for two customizable periods (# of candles) as well as the difference between the two percentages. This means you won't have to visit any websites to find out the current correlation. BTW: this uses pearson's correlation algorithm.

I'll be paying attention now, so please let me know if you see anything funny.

Happy Trading!
Attached File(s)
File Type: txt Correlating Pair Divergence - Signal Only README.txt   3 KB | 608 downloads
File Type: ex4 Correlating Pair Divergence - Signal Only.ex4   24 KB | 730 downloads
File Type: ex4 Overlay.ex4   6 KB | 574 downloads
  • #791
  • Nov 30, 2009 9:05pm Nov 30, 2009 9:05pm
I forgot to explain one of the options on the Overlay indicator.

UseValue (true/false)

This defines whether the indicator will draw itself using the relative value of the chart symbol vs. the sub-symbol or, if set to false, will just use pips.

What this means is if you set this to true and the sub-symbol moves, say, 10 pips in typical price between two candles, but the value of those 10 pips (in your account currency) equates to only 9 pips on your primary chart pair, then the line will only move 9 pips from one candle to the next. If you set this to false then 1 pip movement on the sub-symbol will correspond to 1 pip on the primary symbol.
  • #792
  • Dec 1, 2009 9:39am Dec 1, 2009 9:39am
I haven't really used your stuff in the past Lucid, but I'll give it a try soon.

Two E/U buys resulted in wins this morning.
Attached Image(s) (click to enlarge)
Click to Enlarge Name: eurusd_buy_200912010700.gif Size: 19 KB
Click to Enlarge Name: eurusd_buy_200912011245.gif Size: 17 KB
Hindsight is 20/20
  • #793
  • Dec 1, 2009 1:18pm Dec 1, 2009 1:18pm
Quoting cosgrove
I haven't really used your stuff in the past Lucid, but I'll give it a try soon.

Two E/U buys resulted in wins this morning.

Congratulations on your wins. I'm not sure if either of those buys really adhere to the method as originally laid out by Bill though. The top chart looks like divergence, but there is a lot of noise in between the lows. I think Bill would usually trade these, but mention they are outside of the system. The bottom chart fits pretty well with the rules, except the trend is moving up, which means you would be looking for higher highs.

Anyway, my point is that the signal I produced would not have called either of those buys, though it would have called the sell in the first chart.

Keep making that money!
  • #794
  • Dec 1, 2009 2:51pm Dec 1, 2009 2:51pm
Quoting LucidObscuri
Hey guys. I have to appologize for being gone so long, but I got busy and then was kind of burnt out on metatrader programming.

Anyway, here are the new files. As a quick refresher, the Overlay is an indicator and the "Correlating Pair Divergence..." is an EA. I've modified the Overlay completely so it now bases everything on the current difference between prices (either by value in account currency or by pips). What this means is the Overlay should always be close to the chart line towards the right side of the chart, but may stray as you move...

Thank you, Lucid!!!
It's appreciated
  • #795
  • Dec 1, 2009 3:25pm Dec 1, 2009 3:25pm
You're welcome.

Quick question Bill: Do you pay attention to trend? I see quite a few trades in this thread where a higher high is compared to a lower high while the trend is moving down. When I set the TrendDistanceCandles property in my indicator to 0 I see a lot of signals, but the quality of the signals is poor. I seem to get better results with this value around 30 or so. I'd love to get your opinion on this.

Thanks.
  • #796
  • Dec 1, 2009 3:31pm Dec 1, 2009 3:31pm
Quoting LucidObscuri
You're welcome.

Quick question Bill: Do you pay attention to trend? I see quite a few trades in this thread where a higher high is compared to a lower high while the trend is moving down. When I set the TrendDistanceCandles property in my indicator to 0 I see a lot of signals, but the quality of the signals is poor. I seem to get better results with this value around 30 or so. I'd love to get your opinion on this.

Thanks.
I have never paid attention to trend, but I am prepared to if it is profitable.

The problem I've had with trends is an objective way of determining them.
  • #797
  • Dec 1, 2009 4:09pm Dec 1, 2009 4:09pm
Quoting billbss
I have never paid attention to trend, but I am prepared to if it is profitable.

The problem I've had with trends is an objective way of determining them.

I know exactly what you mean.

I guess the question I have is really: When looking at a higher high, if the price was even higher just 4 or 5 candles ago, is it really a higher high? It would be higher than the previous peak, but if the peak before that was higher than both then can the first peak even be called a "high" for the next peak to be a valid "higher high"

There's an example attached. When looking at the second peak (where the red line starts) can that really be called a "high"? And if it's not a valid "high" then can the third peak (where the red line ends) be called a "higher high".

I don't mean to get too specific here, but I see this a lot and it seems to give more false signals than anything else. This is why the signal specifies a tred distance of 30 candles. This means that when looking for a higher high, none of the last 30 candles can have been higher than that.

Thanks for you insight on this.
Attached Image
  • #798
  • Dec 1, 2009 4:16pm Dec 1, 2009 4:16pm
Quoting LucidObscuri
I know exactly what you mean.

I guess the question I have is really: When looking at a higher high, if the price was even higher just 4 or 5 candles ago, is it really a higher high? It would be higher than the previous peak, but if the peak before that was higher than both then can the first peak even be called a "high" for the next peak to be a valid "higher high"

There's an example attached. When looking at the second peak (where the red line starts) can that really be called a "high"? And if it's not a valid "high" then can the third...

I see what you mean and it's an excellent observation on your part.
I have never thought about it that way.

The main thing is- if it works, it works.

Let me look at it- I think you may have found something good
  • #799
  • Edited 10:23pm Dec 1, 2009 10:07pm | Edited 10:23pm
Quoting LucidObscuri
Congratulations on your wins. I'm not sure if either of those buys really adhere to the method as originally laid out by Bill though. The top chart looks like divergence, but there is a lot of noise in between the lows. I think Bill would usually trade these, but mention they are outside of the system.
Yes, these are the "wiggles" that we were talking about earlier in the thread I think.

Quoting LucidObscuri
The bottom chart fits pretty well with the rules, except the trend is moving up, which means you would be looking for higher highs.
I think I share Bill's sentiments on this. And as he said, I would love to hear anyone's take on determining trend.

Thanks for the words of encouragement though! I do admit that I'm not trading with real money yet; that will start next year.

As for the trades, I have adapted the system a little bit to suit my personality and how I hope to be trading it in the future, and I've tried to make that clear recently (not trying to dig on you for missing it, just reiterating it). My results working like this so far are littered throughout the thread, and I hope to post more soon.

Quoting LucidObscuri
I guess the question I have is really: When looking at a higher high, if the price was even higher just 4 or 5 candles ago, is it really a higher high? It would be higher than the previous peak, but if the peak before that was higher than both then can the first peak even be called a "high" for the next peak to be a valid "higher high"
[I know I wasn't asked, but my two cents follow!] IMO both are "valid" (because they're really just the same trade) and it doesn't really matter. Personally I draw the trade on the smaller pattern out of habit. Just don't ever intersect the chart line with your trend lines.
Hindsight is 20/20
  • #800
  • Dec 1, 2009 11:21pm Dec 1, 2009 11:21pm
Yeah, I've still got a few pages to read since I left. I'll be sure to check out the variations which have been working for you. In the end, if makes money then it's exactly what we're after.

I have to agree with both of you on trend. If there were a way to know then we'd all be rich.

Since I'm a programmer I'm big on backtesting, which I know is not quite the same as forward, but It allows me to quickly reject strategies that I come up with or come across which don't hold muster. Also, it's easy to visually backtest a poor system and "see" it as being successful. Unfortunatelly I can't backtest this one since Metatrader only simulates one pair at a time. I'll start forward testing probably tomorrow.

I plan on playing around with some settings in my EA which will either consider trend or ignore it and see how that goes.

BTW, I know you're not (yet?) using the signal I created, but the "SpikeMinPips" value, if increased will cause it to find divergence despite "noise" between peaks or valleys. The "trend" mechanism can be turned off by setting the "TrendDistanceCandles" to 0.

Thanks for your 2 cents. Always appreciated.
0 traders viewing now