stevehopwoodforex.com
https://www.stevehopwoodforex.com/phpBB3/
Print view

Slope LEVEL -- or DIRECTION?
https://www.stevehopwoodforex.com/phpBB3/viewtopic.php?t=724		 Page 3 of 22
Author:  phil_trade [ Thu Jul 26, 2012 9:02 am ]
Post subject:  Re: Slope LEVEL -- or DIRECTION?
Hi all

I analyse data from 2008 to 2012
I want to point out that TMA Slope has different behavior with different pair :

for GBPJPY price always fit Slope direction (and EA is very successful with it), but wih EURCAD (and it's not alone), Slope can decrease from 1.16 to 0.2 and price increase or stay flat ! Same behavior with normal TMA TRUE and TMA TRUE no repaint.

I think we need to add some detection if price fit slope direction.

An idea?

Philippe
Author:  spotdespot [ Thu Jul 26, 2012 10:05 am ]
Post subject:  Re: Slope LEVEL -- or DIRECTION?
Philippe - it seems to me what you have there is just divergence - something I have seen quite a bit with slope. Slope was decreasing whilst price stayed pretty much flat. After that last bar you show for July 2008 price dropped 1200 pips between then and mid October. If you had entered a trade when the slope started descending and had a SL above the last swing high you would have made a lot of pips - a long wait for it but then I wouldn't have the patience for trading Weekly timeframes!
Author:  phil_trade [ Thu Jul 26, 2012 10:14 am ]
Post subject:  Re: Slope LEVEL -- or DIRECTION?
spotdespot wrote:Philippe - it seems to me what you have there is just divergence - something I have seen quite a bit with slope. Slope was decreasing whilst price stayed pretty much flat. After that last bar you show for July 2008 price dropped 1200 pips between then and mid October. If you had entered a trade when the slope started descending and had a SL above the last swing high you would have made a lot of pips - a long wait for it but then I wouldn't have the patience for trading Weekly timeframes!
Yes weekly trades need patience...but offer +7000 pips trade :) I test W slope as I saw it's less choppy than daily. But I'm also writting a daily version.

So you call that a divergence ! but 14 weeks divergence... I call that a weakness :lol:

This strategy test to Sell when the weekly angle of the slope decrease and vice versa for buy. I want to avoid a lot of bad trade when you wait +/-0.4 level and slope switch just after. We lose a lot of price mouvement. It' very good on GBPJPY for now.

the question is : how to detect/avoid this ... divergence/weakness ?
Author:  spotdespot [ Thu Jul 26, 2012 11:06 am ]
Post subject:  Re: Slope LEVEL -- or DIRECTION?
phil_trade wrote:
spotdespot wrote:Philippe - it seems to me what you have there is just divergence - something I have seen quite a bit with slope. Slope was decreasing whilst price stayed pretty much flat. After that last bar you show for July 2008 price dropped 1200 pips between then and mid October. If you had entered a trade when the slope started descending and had a SL above the last swing high you would have made a lot of pips - a long wait for it but then I wouldn't have the patience for trading Weekly timeframes!
Yes weekly trades need patience...but offer +7000 pips trade :) I test W slope as I saw it's less choppy than daily. But I'm also writting a daily version.

So you call that a divergence ! but 14 weeks divergence... I call that a weakness :lol:

This strategy test to Sell when the weekly angle of the slope decrease and vice versa for buy. I want to avoid a lot of bad trade when you wait +/-0.4 level and slope switch just after. We lose a lot of price mouvement. It' very good on GBPJPY for now.

the question is : how to detect/avoid this ... divergence/weakness ?
I am not sure there is a clear way to avoid it as, whether you call it divergence or weakness, it doesn't show a clear turning point just a change in strength of a trend. When I have been looking at daily and lower trades I look at a combination of slope direction on the lower timeframe, slope direction on D1 and the value of the D1 slope.

For example - if I am looking at a trade on H4 (where I take most of my trades). I firstly look at D1 slope value, let's say it is >0.8, I then look at slope direction (angle) on H4, if it is "up" then I check D1 slope direction - if that is up then I am looking for a longer term trade - obviously as everything is in the direction of the trend.

If D1 is >0.8, H4 slope is down and D1 direction is up then I treat that as a retrace (unless D1 slope value and direction changes confirming a trend change) and try to get back in to the direction of the trend when H4 slope direction changes to up (I actually look at H1 slope direction for early warning/use basic price action S/R as well). Alternatively I could take a CT trade but I would actually avoid this unless D1 slope direction is down.

Obviously these are just 2 scenarios but the point is I look at multiple timeframes and slope direction and D1 value to determine what I am seeing (retrace, trend change, ranging period). I am sure this could apply to Weekly too but I don't have enough history on most pairs to look at what Monthly TMA is doing to check.

Sorry - that probably doesn't help you much but hopefully someone will find some value there.

One thing I did notice during those 14 weeks - price did initially drop considerably (well maybe not for weekly trades but 1000 pips or so) then it retraced. If you had taken the trade when the slope direction was down (on weekly) you would never have been in a drawdown or not significant anyway (depending on how exactly you trigger that trade) - just tied up margin. Also I guess a 500 pip trailing stop or something like that would have locked in a small profit and allowed you to reenter later after the retrace?
Author:  garyfritz [ Thu Jul 26, 2012 2:09 pm ]
Post subject:  Re: Slope LEVEL -- or DIRECTION?
I think there are several reasons for what you're seeing:

1. It worked so well on GBPJPY in that timeframe because GJ was trending hard and steadily, for 5 months = 20 weeks. But on the weekly timeframe you were looking at, EURCAD was going more or less sideways. It trended up in late 2007 / early 2008 and the W1 slope went up, but then from April to Nov it didn't have a clear trend. TMA isn't good at going dead sideways when the market's not trending.

2. I think you're looking at too short a period. Your slope indicator is using the W1 timeframe, but it's not going to work well on really short weekly moves. It worked fine on the long clear GBPJPY trend but not on the choppier EURCAD moves. You need to move to a shorter TF to catch those shorter moves.

3. Even so, the W1 trend did a pretty fair job of moving with the larger-scale trends, if you look at a longer period. The orange line on this chart is equivalent to your second (true) histo on EURCAD:
Slope.gif
It may be that W1 slope is not a good tool for catching the moves you want, or maybe you need to combine it with something else. Notice that the D1 slope (blue) catches those smaller wiggles pretty well. Maybe you could trade just when the W1 and D1 slopes agree? Maybe you could use W1 when it's > 0.4 (trending hard) and D1 when the weekly slope says it's ranging? Maybe only use the change in W1 slope when the rate-of-change is high enough (indicating a strong trend)?
Author:  phil_trade [ Thu Jul 26, 2012 2:56 pm ]
Post subject:  Re: Slope LEVEL -- or DIRECTION?

1. It worked so well on GBPJPY in that timeframe because GJ was trending hard and steadily, for 5 months = 20 weeks. But on the weekly timeframe you were looking at, EURCAD was going more or less sideways. It trended up in late 2007 / early 2008 and the W1 slope went up, but then from April to Nov it didn't have a clear trend. TMA isn't good at going dead sideways when the market's not trending.

2. I think you're looking at too short a period. Your slope indicator is using the W1 timeframe, but it's not going to work well on really short weekly moves. It worked fine on the long clear GBPJPY trend but not on the choppier EURCAD moves. You need to move to a shorter TF to catch those shorter moves.

3. Even so, the W1 trend did a pretty fair job of moving with the larger-scale trends, if you look at a longer period. The orange line on this chart is equivalent to your second (true) histo on EURCAD:
Thanks Gary, but my concern is more about divergence we can see on EURCAD. Slope can have big move as price mouvement do not follow and stay flat.
It looks for me as Slope calculation is not good enough and do not reflect what we espected : a pretty good prevision of price movement. Or may be a decreasing slope could say two things :
- Price is decreasing
- Price is no more increasing and stay flat
But how to read well this indication ?
Author:  cmeade2 [ Thu Jul 26, 2012 6:32 pm ]
Post subject:  Re: Slope LEVEL -- or DIRECTION?
It seems to me that the indicator that would be of greatest use for this idea is the difference between the TMA true slope between two bars. This would be a measure of the slope of the slope. Basically if there is a big move in the TMA in a particular direction (up or down) that would be a sign to get in, regardless of the level.
Author:  garyfritz [ Thu Jul 26, 2012 6:55 pm ]
Post subject:  Re: Slope LEVEL -- or DIRECTION?
That's what Phillipe is looking at, cmeade2... :)
Author:  cmeade2 [ Thu Jul 26, 2012 7:42 pm ]
Post subject:  Re: Slope LEVEL -- or DIRECTION?
garyfritz wrote:That's what Phillipe is looking at, cmeade2... :)
you mean he is going to develop it?
Author:  garyfritz [ Thu Jul 26, 2012 7:44 pm ]
Post subject:  Re: Slope LEVEL -- or DIRECTION?
I don't know what anyone is developing. I started this thread to open discussion on it, Phillipe is doing some research, other people may be looking at it.
All times are UTC Page 3 of 22

Powered by phpBB® Forum Software © phpBB Limited