slope of an ema efs

[gspiker]

I have a few formula's that calculate slope using the simple rise over run. And it works fine except that since rise isn't normalized between symbols it is hard to program in paramters for slope. Is there a good/common method for doing this?

Garth

[ckryza]

Hi Garth:

Not sure I understand correctly but.... one method would be to convert the raw slope to a percent change (basically just a plain ole' ROC).

Chris

[gspiker]

Chris,

DOH! So simple.

Thanks!

Garth

[mbuta]

Slope of an MA "compiled"

Hi guys, great subject.
I tried to "compile" the information here to get the angle of an MA, but flanked (got all "black")?!?

Can someone help us put it all together, from diferential MA (Ma - Ma(-1)), to "percentage MA", to nAngle?

Thank you.
Mihai

[ckryza]

Mihai:

The %Slope is simply the current value divided by the prior value. So if you were computing a 5-bar MA slope and the current value was $50 and the value 5-bars ago was $40 then the %slope would be 50/40 or 1.25 which equals a 25% change which equals a 5% slope (e.g., per bar).

The angle of this slope could then be computed as:

nAngle = Math.atan( nSlope ) / (Math.PI/180);

nAngle = Math.atan( 0.05 ) / (Math.PI/180);

nAngle = ~ 2.86-degrees per bar

I hope I have the math right

Chris

[mbuta]

Thank you Chris.

Unless I am missing something obvious, I think, the angle calculation should not be dependent of the number of bars.
The angle is the raport between two axis (x,y) but here the two axis are of diferent kind, therefore cannot be simply divided.
This was my confusion all along, how to make them "of the same kind".

If the slope stays the same, over 2 or 100 bars, the calculated angle should be the same.

Any idea? Thanks,

Mihai

[ckryza]

Mihai:

I agree that if the slope is the same (over time) then the angle will be the same. But, on a chart, that is rarely the case in that the 2-bar slope will probably differ from the 3-bar slope, etc. etc.

I don't see how you can separate the # of bars from the calculation of the slope and/or the angle because the slope will, by definition, be dependant upon the period that you are analyzing. If all you are concerned with is the slope/angle of the line drawn from point A to point B then, yes, I would agree that you can safely ignore the number of bars. You would simply compute the %-change from point A to point B and then convert that to an angle using the same calculation. All I was doing in the previous message was taking it one step further and converting that to a degree-change-per-bar which is typically what I use.

Chris

[mbuta]

Thanks Chris,

I just realized that NOT the "angle" of the slope is important, because it will change with many variables (like how many bars displayed, or how vertically large is your chart).

Therefore, the %change is the relevant number, because this remains constant independent of the display parameters.

Thanks again. See if you have any idea in the other thread I posted this morning.

Mihai

[ACM]

Mihai
If the angle is derived from the slope (ie rise over run) it will be independent of the height of a chart or the number of bars displayed even though it may "appear" to be different (see attached image).
Alex

[mbuta]

Re: Reply to post 'slope of an ema efs'

Chris,
Thank you. It seems simple.
However, why the angle calculation is dependent of the number of bars? The
angle may be the same accross all 5 bars, so the result should be the same,
regardless of the number of barss involved.
Mihai

----- Original Message -----
From: <[email protected]>
To: <[email protected]>
Sent: Saturday, April 05, 2003 8:39 AM
Subject: Reply to post 'slope of an ema efs'


> Hello mbuta,
>
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>

[wavetrader]

slope of ema for Bonds

Okay, I need a little bit of help here. I can't seem to figure out how to calculate the slope of ema for Bonds (which is quoted in 1/32nds).

On a 1min chart on Bonds, when the difference in MA from the previous bar is equal or greater than 1 tick (0.03125), visually the slope of the ema is already steep ~45 degrees already, so I'm thinking the formula for ema slope has to be tweaked a little bit for Bonds. Whenever I use the regualr formula here, the maximum number of degrees I can get it roughly 3 degrees, which is not quite right.

Can somebody please help me code the ema slope formula for Bonds, since I am no programmer.

Much appreciated,
wavetrader

[wombat953]

Has anyone here actually used the slope (of a MA, say) in an EFS to identify and/or confirm trends? And tested it for effectiveness?
Thanks
George

[Natalie]

Re: Reply to post 'slope of an ema efs'

;
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

[wombat953]

Natalie
It looks as if your last post did not make it through. Can you pls. repost?
George

[Natalie]

It is something I have been looking at at various times, but never calculated it into a .efs. Certainly on the list of things to do, not because of the slope itself, but because of attraction and proximity research, in which angle of slope plays a part.

Kind regards

Natalie

The only time I ever try to respond via e-mail instead of online it doesn't get through LOLOL Oh well....

Kind regards

Natalie