Divergence

As this link is over 100 pages long, i will try to summarise his method:

The setup:
My signals all come from the SP market (not emini). I don't know if they'd be exactly the same. I don't see the noisey spikes in the pit prices so that's what I use. I execute all trades in the Emini which is usually within .25 of the price I see on the SP contract. If not, I pass on the trade. My screen uses 5 min. bars using only day session data. My time is set to Chicago time so all times on the charts are Central time. I have three indicators on my screen. The first is the Keltner Channel with 20 period and 2.5 ATR's (average true range) based on the close. The second is the Chaikin Oscillator using Volume, 3 for the Fast and 10 for the Slow. The third is a 14 period ADX. My window is set to 8 bar spacing.
 

Attachments

  • divergence setup.gif
    divergence setup.gif
    13.5 KB · Views: 1,382
The basic method is pretty simple. It gives me plenty of warning of a upcoming trade so I know when to get down to business.
The setup requires three basic things. 1). The ADX indicator must be over 30. 2). The Chaikin Oscillator must be showing divergence from the price. 3). The price must be near or outside the Keltner Channel.

Usually price is not a problem. I just added it as a rule to avoid taking bad risk/reward trades.

Once these 3 conditions are met, the trade is executed after the first full bar that doesn't set a low/high in the direction of the trend. I use the open of the second bar to base my slippage estimates. Sometimes I get better fills and sometimes I'm off by .5. They seem to balance out over time.

The amount of risk per-contract is determined by subtracting the high from the low of the Keltner Channel at or near the time of the trade. Then 1/3 of this number is used as the stop loss amount. For example, if the Keltner Channel is 12 pts. from low to high, then the stop loss would be 4 points. I risk 2% per trade, so if I have 100k then .02*100,000 = $2,000 risk per-trade. The number of contracts would then be 2,000/200 (4*50) or 10 contracts
 
Thanks for drawing this to our attention fazalv. Have you got any other examples of positive signals that you can post? Presumably there are plenty on elitetrader.com ?
 
RogerM,

This journal had the highest win rate (an obscene 70%) out of the recent ones I came across on Elite.

The dude went on to say:

With this method along with some advanced stuff I've added, I've taken a 25k account to over 200k in about a year and a half. I trade with 2% of my capital risked per-trade. I've had more than 10 winners in a row and I've had more than 10 losers in a row. Overall this method has given me a larger profit than loss on each trade and a 70% win rate. I only get 1-2 trades a day on average, so if you're a scalper this is not the journal for you.

Hope this helps!

I will post some actual trades at a more godly hour!
 
Thanks fazalv. I'll trawl thru the thread and try to recreate the charting set-up.
 
Sadly the thread ends rather abruptly.
The member did not trade the method on trending days btw.
 
For anyone else interested there are loads of well annotated set-ups on the elitetrader thread, which is set up as a very clear journal, with clear, simple entry and exit rules.
 
Last edited:
One thing that does need to be clarified is that you have to use a
feed for the actual S&P future and not the Emini S&P because
there is a difference. Those that used the Emini didnt get the
same signals as those with the full S&P data.


Paul
 
Sierracharts doesn't have a built in Chaikin Oscillator and I can't find a .dll to download. Anyone got any ideas?
 
There is a Chaikin Oscillator file available for download from eSignal's website.
 
Trade Example

1. ADX crosses above 30 - 9:15 bar.
2. Divergence starts with 9:20 bar.
Keeps making new lows.
9:40 bar first up bar.
3. Keltner width 10.07 - stop loss is 3.5 on this trade
4. Buy near open of 9:45 bar at 966.75 stop loss 963.25.
5. Exit at 11:20 bar at 972.5

Entry long 966.75
Exit long 972.50

Profit(loss) +5.75
Risk 3.50

Risk/reward this trade about 1.6

Cumm. trades since 8/1.
 

Attachments

  • divergence2.gif
    divergence2.gif
    20.4 KB · Views: 1,174
I have checked this setup over the last week and so far can only
find one trade that would have been taken. Has anyone else
looked at this ?

I will be doing some backtests as well to see if more trades
would have been taken further back.


Paul
 
Hi Imran,

I used the data as was described in the method on ET ? I am not
sure how else it could be done ?


Paul
 
Paul, I mean did you use full contract data? the guy described the method using full futures contract data although he said he hadnt looked at using emini data.

Imran
 
Imran,

Yes I did and attached is a chart of one of the trades that would
be taken on September 30th.

ADX > 30
Divergence shown by yellow lines
First bar to close without further decline in price
So enter on open of next bar at 991.30 as shown by white up
arrow
Close when price touches high of Keltner channel at 997.7

Profit = 6.4 points which is $320 on Emini or $1600 on full S&P

Any comments ?

Cheers

Paul

PS Sorry for image quality TS2000 is poor for this ( I will try and
edit it and post improvement if achieved)
 

Attachments

  • image4.png
    image4.png
    15.5 KB · Views: 894
Last edited:
I looke dat this several weeks or months back wjhen originally posted. Found zero trades on ES in 5 days. Looked at it again on DAX: no trades..

(or ther were if you held overnight and were lucky ..)
 
Trader333,
A better indicator that gives better divergence signals is
described in a book called 'Trading with the Odds' by Cynthia
Kase. She claims that divergence signals corroborated by two of
her indicators gives 100% success rate.

see:
http://www.statware.com


Imran
 
Imran,

'Editor's Note: Enron Capital & Trade Resources recently joined forces with a small independent consulting firm, Kase and Company, to offer customized risk management services'

The above is an extract from http://www.fenews.com/fen5/kase.html

I am intrigued by the Enron connection.
 
Top