Slope LEVEL -- or DIRECTION?

[garyfritz]

I've noticed that the Slope takes a long time to get up to the key levels. You might miss several hundred pips of a move before the D1 slope finally works its way up to 0.4. Furthermore, because we're using specific levels, that means any difference in crims/data/etc can result in slightly different levels for different crims -- not a good situation.

I'm wondering whether Slope level is the right thing to use -- or if we should be looking at Slope direction.

Assuming Neotrader's historic display of the TMA is true -- and he assures me it is, even if it looks "too good" -- then the direction of the slope might be a better indicator than the slope level.

Take a look at this (very busy ) chart:

Slope.gif

I've marked two different sorts of signals on here. The green and pink boxes are areas where the D1 slope is > 0.4 or < -0.4. These are the areas that "traditional NB10.2A" would call out as long/short candidates. You can see that some of them are good, but 1) you miss a lot of the start of a new trend, and 2) the signal often persists long after the trend has turned. Sometimes these two combine to give you a signal that hits just as the market turns, and you get whipsawed.

Now look at the vertical red/green lines. These are a crude eyeballed attempt to specify whether the D1 trend is currently heading up or down. I placed a green line when the D1 slope went up twice in a row, and a red when it went down twice in a row.

It's not always correct, but it gets you in a lot earlier and you generally catch a lot more of the trend. If you just went long at the green lines and short at the red lines, you'd do pretty well.

Or maybe you combine them? Go long when Slope > 0.4 AND Slope is increasing?

I'll throw it out there and let some of you creative types take a whack at it...

[Thunder123]

I have thought about this before and I believe there are 2 ways we can use this (I think #2 is more promising):

1. Combine slope and direction like you said. But the minimum slope level would have to be less than 0.4 (otherwise it basically the same as before, but with even less signals).

2. We take slope of the slope, and plot it in an indicator window. So if this slope (call it Slope2) is increasing at a rate higher than a threshold level (we will have to determine this), then it means slope is increasing rapidly and we should look to enter. I believe this will help us identify the start of a trend. This gives us more information than simply the direction of the slope, it gives us the rate of change (which can be VERY useful).

Something to think about.

I just finished reading Jems Basket trading thread and will try it next week. Hopefully I can contribute more as I learn the system.

[Thunder123]

I don't have much Forex experience but here is something else to think about:

Why exactly do we need identify (and get in) at the start of a trend? If we take a step back and think, we realize that there are always some pairs trending according to TMA slope. At a particular point in time, we only need to trade these pairs. If we assume we can get in at the middle of the trend and get out near the end, then by the time we are done with these pairs, new pairs will be in the middle of the trend (which we can enter). Therefore at any point in time we can enter pairs that are in a trend, and there is no need to pinpoint the start.

Again, I don't have much Forex experience and I'm not sure if we will always have couple pairs with slope >0.4 <-0.4.
Let me know what you guys think.

I do believe we need to fix the difference in slope problem.

[garyfritz]

Missing the start of the trend is not a big problem **IF** the trend continues long enough for you to grab a piece out of the middle. It doesn't always cooperate that well. Some trends run out just about the time the Slope gets up to the key level. An indicator that detected the trend as soon as it started would get you in on more winning setups, and it wouldn't keep telling you to go long after the market has turned down. I don't know if the rate of change of the slope would do that or not, but it shows some promise.

But it may be that "all 3 D/W/M pivots in order" is much simpler and works better. Just tossing out some ideas.

[Thunder123]

Well I guess its worth a try.

Perhaps some kind of weighted average of TMA Slope and rate of change of slope. We will end up with a indicator that looks like TMA Slope but perhaps better.

[harry45]

I've noticed that the Slope takes a long time to get up to the key levels. You might miss several hundred pips of a move before the D1 slope finally works its way up to 0.4. Furthermore, because we're using specific levels, that means any difference in crims/data/etc can result in slightly different levels for different crims -- not a good situation.

I'm wondering whether Slope level is the right thing to use -- or if we should be looking at Slope direction.

Assuming Neotrader's historic display of the TMA is true -- and he assures me it is, even if it looks "too good" -- then the direction of the slope might be a better indicator than the slope level.

Take a look at this (very busy ) chart:

Slope.gif

I've marked two different sorts of signals on here. The green and pink boxes are areas where the D1 slope is > 0.4 or < -0.4. These are the areas that "traditional NB10.2A" would call out as long/short candidates. You can see that some of them are good, but 1) you miss a lot of the start of a new trend, and 2) the signal often persists long after the trend has turned. Sometimes these two combine to give you a signal that hits just as the market turns, and you get whipsawed.

Now look at the vertical red/green lines. These are a crude eyeballed attempt to specify whether the D1 trend is currently heading up or down. I placed a green line when the D1 slope went up twice in a row, and a red when it went down twice in a row.

It's not always correct, but it gets you in a lot earlier and you generally catch a lot more of the trend. If you just went long at the green lines and short at the red lines, you'd do pretty well.

Or maybe you combine them? Go long when Slope > 0.4 AND Slope is increasing?

I'll throw it out there and let some of you creative types take a whack at it...

To extend trade I use MA applied to NEO indi.As you can see if WP and MP are in row trade can be taken (SHORT) in positive Slope and vice versa.Conditions: SHORT-Slope<MA,LONG Slope>MA.Amazing?

[Jemook]

To extend trade I use MA applied to NEO indi.As you can see if WP and MP are in row trade can be taken (SHORT) in positive Slope and vice versa.Conditions: SHORT-Slope<MA,LONG Slope>MA.Amazing?

I can't seem to attach an MA to the NT indicator. When I drag it on top of the NT indicator it just applies to my charts. Any ideas?

[harry45]

To extend trade I use MA applied to NEO indi.As you can see if WP and MP are in row trade can be taken (SHORT) in positive Slope and vice versa.Conditions: SHORT-Slope<MA,LONG Slope>MA.Amazing?

I can't seem to attach an MA to the NT indicator. When I drag it on top of the NT indicator it just applies to my charts. Any ideas?

Jem,
You have to choose MA from Custom indicators drag and drop it into the window where Neo indi is and apply to previous indi data.
See template.

[garyfritz]

Direction: we have the value in same direction twice in a row (if it works well, with your authorization, I’ll add it to the definition’s thread).

Nah, I don't claim that's the right way to do it. It was just a simple test for experimentation purposes. I'm sure there are better measures, like Thunder's MA or slope-of-slope (rate of change) measurement.

And like Thunder123 mentioned, there are plenty of fishes in the sea. So I stay away from using the slope in a ranging market.

In a true ranging market, yes. But a market can change quickly from up-trend to down-trend, and there is a period where you're clearly in a strong down-trend, and the slope hasn't made it to -0.4 yet. This is the area where the slope rate-of-change might be useful.

Or it may be that it the slope rate-of-change looks pretty in a few examples, but overall it's safest to wait until slope gets to +/- 0.4 or whatever. I'm not sure.

But where it is extremely important to quickly identify the direction, is when there is a reversal so we can avoid being whipsawed (we don’t want that canon ball to fall on our head) .

Yes, possibly the slope rate-of-change isn't good for *entering* a trend -- maybe 0.4 is a safer indication -- but maybe it's a good way to quickly *exit* when the trend turns?

[lucklogic]

Garyfitz..............

Or maybe you combine them? Go long when Slope > 0.4 AND Slope is increasing?

I'll throw it out there and let some of you creative types take a whack at it.......

I've being doing this for the last couple of weeks - it definitely seems better to me