Pair Divergence Method

from my experience it's impossible to use divergence to trade on it's own.

Bill

I know you originally didn't want indicators.
I've been playing around with a MTF RSI and SMA 200 on the chart

Waiting for RSI to be OverSold <30 or OverBought >70 in 3 time frames
15tf 60tf 4hrtf and sma 200 above or below on the 4hr time frame
to confirm my buy or sell

Been pretty successful in demo on the E/U U/C but only this month I have no long term data as of yet

I think this disproves that- at least so far.

My schedule was a little messed up last week, so I didn't get a chance to post much. Catching up with 3 wins attached, in no particular order. I saw a few "almosts" that were equal peaks which I did not include; you might see a few on these charts.

Sell E/U



On a long, buy the pair that did not make the new low.
On the short, sell the pair that did not make the new high.

Is the pattern you sold in the pink circle I added? Is the red line E/U?

No I sold E/U where I have the yellow line

I beleive where you have the cricle would have been a sell for U/C.... Yes?


If I'm wrong I'll scream LOL

Both pairs are making new highs where you have the yellow line though, and you don't even have a second peak formed yet...that's why I am confused as to what you are trying to point out.

The pink circle looks like it would have been a U/C sell, yes. But remember, when you switch to the U/C chart, it would actually be a buy (because of the negative correlation).

You are helping me a ton but I need to go back and re- re re read the thread..

For what it's worth I've been over 90% in winning trades.
But I still have trouble grasping the rules

Thanks for all your help

If you're hitting 90%, write down your version of the rules and we'll all use them.

The problem is I might be reading the divergence 90% wrong and I'm really only 10% right

LOL

Below is a link to what I'm doing

http://www.forexfactory.com/showthre...93#post3399193

I'd love to hear:

1. What are your average win and loss sizes?
2. What is your average stop loss size in pips?
3. How many trades have you taken this month?
4. Why are you using 15/60/240 and not 30/60/240 or 15/30/60?

1 I'm using .01 sometimes .25 but my goal is win percentage
If I were live I'd go .1 lots or account balance /25000

2 SL is 100 TP I use e-trailing ea set to 100TS with a trail step50 (5 digits broker) sometimes I'll adjust the TS to ATR*2

I also adjust the SL I look at the RSI and the x-meter S&R Fibs

3. 28 trades I attached a graph...disregard the big gain I placed an order and my son exited the program I turned it on the next day and I was up over 2000 pips. I reset my demo account cause it was screwing up my analysis.

4 I'm using 3 x my lowest TF mine is TF15. TF 5 may give headfakes. 15 and 30 are pretty much the same

so 15 H1 and H4.

When I place an order

First On a long, buy the pair that did not make the new low.
On the short, sell the pair that did not make the new high

To confirm:
I look at the RSI(2) of 15 H1 H4 for OB or OS should all be the same

Then I check the xmeter for the strength of the currency

Then I check if it is over or under the SMA 200
on the highest TF used. Sometimes if it looks good I'll go for it and disregard the SMA


I know what I'm doing is against what Bill wanted in his first post. NO INDIES I don't want any EAs or indicators or filters. This one will stand on its own
but he also said

This is one of the rare mechanical systems that is profitable.
It may be the only one we'll ever see.

It is also one I would not trade. I would not trade a system that shows a negative month.

This method is a framework for a very profitable system. Anytime the basic method exceeds the expectation of random entries, it can be improved upon.

I'm open to suggestions, particularly from Cosgrove and Artov and Lucid, who have worked with it.

These are not my own ideas but bits and pieces from some of the other threads I follow.

I buy no means recommend this it works for me right now in demo I have not backtested it. I don't think I can with these indies. But it looks promising and is technically sound. This whole thing could crap out down the road who knows. Keep what you like and pitch the rest

I'm still struggling with the On a long, buy the pair that did not make the new low. On the short, sell the pair that did not make the new high.
Thanks to Cosgroves help I might get it figured out

Hope I answered your questions

Thanks for your long and detailed answer! I must point out that you cannot ignore average win and loss size when determining the strength of a system. If they are roughly the same (as bill originally suggested) then that's fine and you can concentrate on win/loss %. I just wanted to clarify.

The idea of "congruence" -- waiting for more than one indicator to say the same thing -- is a nice idea with this system, but I'm afraid that you may filter out lots of trades. Even with a 90% win rate you can only exploit leverage so much, and your total return may be lower. Something to think about...

This is true. Over time if I do maintain a high win percentage I could increase my lot size to 5.0 LOL

Sorry, I meant relative lot size. For example, if you win 90% of the time, but your losers are 10x as big as your winners, you will not be profitable in the long run (unless it is by pure luck).

That said, if you truly had a 90% winning system, it looks like you could generate a pretty nice equity curve even if your average win was 1/5 of your average loss.

fun tool: http://www.hquotes.com/tradehard/simulator.html

Catching up from yesterday, as well as one early this morning. The U/C sell had a pretty small SL (2x ATR = 14p).

Kind of a wacky trade, as it showed up as a big pattern on the 60m. If trades show up on the 30m or 60m, I want to see if they also show up on the 15m. If they do, I will take the trade there, as it will have a smaller ATR which should result in less time in the market before a SL or TP gets hit. Less time in the market equates to less time my money is at risk, and it also lowers the probability of having two trades running at once. Of course with a smaller SL, you run a higher risk in general of the SL being hit, but off the top of my head I want to say that most 15m trades backed by higher time frame trades are winners (generally a common theme among traders: find a signal on one time frame, and go one time frame lower to get a good entry point).

That being said, I found that the pattern on the 60m did indeed show up on the 15m, but it was a very large pattern. No big deal! If you didn't catch this trade, there was a shorter pattern on the 15m right afterwards that also resulted in a win. TP would not have been hit in time to take that trade (unless you wanted to double up on a position, which I don't really want to).

January 2010 is on quite the win streak. Here's a snippet of trades since the 13th:

Break Even
Win
Win
Win
Win
BE
Loss
Win
Win
Win
Win
Win
Win

If this is how January ends, average win=1, average loss=0.67. 16 wins, 9 losses. Expectancy is a whopping 0.4, and % return trading 3%/trade would be the second highest yet in my testing so far (July 08-Present).

edit: since I would call it quits at 1pm EST and go to bed if I were trading live, and there were no more trades today, those are indeed the final January stats.

Sell and cover this morning for a 1/4 win.

Large overnight pattern for a loss on the U/C, and an E/U buy in progress, close to TP.