Andrews Pitchfork (Median Line)

.

Jharteam, those are amazingly crisp, clean charts. Awesome charting! Your use of the Fib channels is wonerful as well. In fact, you may have "found" something that Tim Morge has alluded to in terms of "phantom medianlines" in one of his "webinars."

Those appear faily clearly in your use of Fib channels. Cool stuff!

Clay

Hi Clay:

Wow, I didn't expect there is reply in this forum. But thanks for the comment. You did mention 'phantom median lines' before and I'm very much interested in knowing the progress of Tim's findings. I will continue to post charts here, whenever I have the time; disregarding whether there are any responses or not. I find this methodologies just too amazing to be ignored. Please take care.

Have a good one
Jeff

Jharteam,

Could you please show in detail, what fractals you are using to construct this modified schiff?

And what channel are you using to base a .618 measure and why just this set and not carried through the remaining measured warning lines, as the others look to be equal of .50 divisions?

When do you use a .618 and not some other fib measure?

regards,

Thoughts?

That's an example of a sliding parallel, drawn on a channel or APO (Alternate Point of Origin) fork.


Question, how would you trade it?


edit: Hopefully my last couple of posts are not taken out of context or the wrong sentiment attached, an Andrews pitchfork or channel, schiff or what ever indicator - is just a tool to be used to help build some framework to the chart, assisting decision making. The tool itself can show some very nice balance/relationships between moves, but while we can move the fork all over the chart finding these nice pics, it's the ability to trade with the tool that matters.


The last several pics on this thread are nice but how are ya gonna trade it in real time, before price makes that next move, where ya gonna enter and where ya gonna jump off or reverse when it dosent go as planned?

Here's a quick "down and dirty" look at your charts Kapital. Hope it helps.

Clay

That's the whole point of showing people how I go about trading the pitchfork. And they are not pointed out in details simply because I believe people who frequent this forum has already the working knowledge in trading the fork. I will answer the other questions you have posted on the other thread shortly. My charts are ALWAYS nicely presented; as I believe there's no point of showing people or saying anything at all; if nothing is being presented in a clear concise manner.

BTW, I'm a full time trader without a job; so I'm quite down to earth and straight to the point about things especially in terms of trading. I only show things to people that work for me when those methodologies are tested with my monies. Then again, at the on start, people trashed Catalin about his Modified Schiff; but his Schiff is the MOST PRACTICAL when it comes to trading with real monies.

Thanks Clay for the very thorough explanation, this is exactly how I looked at this fork in terms of how to go about capitalizing it. Yes, I do think at the latter part of this fork, the price actions had become very corrective and should refrain from short entries in the 240-min time frame. However, I am quite sure I would go for more short entries in the lower time frame; such as the 30-min time frame. To refine entry points, please take a look at the 5-min and 15-min time frames as well; they yield very low risk entry point. Please see attached 15-min & 30-min time frame's patterns for short setup.

At this juncture, it's really a traders' market for the moment; due to the possibility of a rather large corrective in the making. The current pattern is also a potential declining wedge; meaning that the recent DAILY sell off of Gbp/Jpy could pause for a corrective rebound. That said, we still have to keep in mind the current fork is a downward sloping one; and, if one is looking for a short-term trade, I would wait for a corrective rebound for a short setup. Please take note the Daily Stochastic is still very much on a bearish posture and a POTENTIAL slingshot is in the making; any trades should still be gearing towards the shorts side until the current downward sloping fork is violated decisively. Hope this helps.

Have a good one
Jeff

The "Shakeouts" Pattern is really nothing more than an Expanding formation; the real purpose of this formation is quite obvious: To run stops. It can transpire as a topping, bottoming and continuation pattern. Trading BREAKOUT of this formation is quite impractical; as far as risk/reward issues are concern. The "Shakeouts" pattern is really just another term for irregular corrective pattern in Elliott Wave. Utilizing "Shakeouts" setups; stop losses placement become reasonable; and when proved wrong; traders can get stopped out at par in MOST instances. Probability of success to trade this rather 'fearful' formation is very high. Take care.

Have a good one
Jeff

Auxesis:

Please see attached for your question on my modified Schiff fork. And this is Catalin's version of Modified Schiff Fork. Apologize for omitting the construction of the channels. In fact, there are TWO (the blue and the red channels) channels with 0.618 intervals anchored on the Median Line of this modified Schiff fork.

The first chart (Blue & Red channels) are based on a 0.618 intervals and the second chart (Green & Red channels) are based on a 0.50 intervals. Please check it out. Personally, I don't find much differences between the two intervals. The real reason I use 0.618 intervals is because I don't want the chart to clutter up. As far as I'm concern, all fibonacci intervals will work; be it 0.236, 0.382, 0.786, 4.236...etc. It's a matter of just how much potential 'energy points' we want to see in the time frame of interest. Take care.

Have a good one
Jeff

Spicin' things up a bit...

Aaargh! I'm so angry I might have to take a nap.

Alpari have been tinkering with MT4 and now their charts are getting the 'drift' when you change time frames. So now we get situations like this on AUDUSD where on the 30m we have a fork structure and it's parallel that we have dropped out of but may provide some resistance in future so I'm keeping it on for the moment. But then switch down to the 15m and it has shifted. This renders the charting completely useless.

Anyway, that is the problem, so what is the solution? Another charting package of course. But which one? I appeal to other forkers out there to report which charting packages, free or paid for, that they find are accurate and do not suffer from this sort of nonsense.

Also, the holidays have created gaps, and where there are not gaps we have tight ranges where we can only assume that volume must have been negligible. Morge and other traders who trade futures markets have access to accurate volume statistics allowing them to use volume based tick charts to overcome this, but FX does not have this luxury. Are other fork traders discounting the significance of their forks that stretch this period? Personally I like to keep an eye on the bigger picture so have forks drawn on 4H and daily but most of my trading is done on 15m charts so the problem is not huge, but irritating nevertheless.

eSignal has pretty good PF tools. You can color lines individually. You can add fibs to them, along with extensions. The one problem I have is that once you draw one, you can't adjust it as you can with MT4. You have to delete and redraw. It's a minor annoyance though that you'll get used to if you stop using them on MT4. eSignal has various packages, so you should find one that suits you. I've been using them for almost 6 years now and they're stable and consistent.

Hi Birdt:

Is it feasible for you to use the Demo version of MT4; the 'original' MT4 from the MetaTrader website? It doesn't have any problem for the lines to 'stay put' during time frames' switchings. I understand when you are talking about the 'drifting' problem. These companies, they white label the original MT4 and the data guys keep screwing around with timestamps (Of course, they don't give a damn and have no knowledge on chartings when your warning or forks' lines have shifted). As to gap occurrences, I usually ignore this 'hiccup' on the part of MT4. At least to me, it didn't affect my anchoring and the eventual outcomes of the pitchfork. BTW, I notice Tim uses Elliott Wave software to draw his forks and ESignals have really nifty forks (I'm not a current user though) Hope this helps.

Oh, hot damn! Thanks for pointing out the problems using an Aud/Usd chart! Can't believe she has SO MANY harmonics patterns to point for a bottoming!!.....And I MISSED it all. See, told you guys to do some postings already! Take care.

P/S: Aud/Usd has the tendency to demonstrate very nice harmonic patterns.

Have a good one
Jeff

Thanks for the responses guys. All the best with your trading.

Lift-off!!!!

A run to S3 before the close of the day perhaps?

Kcapital, I like you pivot point indicator; could you share this one with me please? I hate mine as it has long lines across the charts.

Have a good one
Jeff

Hi birdt, jharteam,

I tried the solution jharteam suggested which was to download MT4 direct from MetaQuotes. The "drift" problem still occurs, but perhaps the problem may be less on that platform because they use a different datasource. I don't know if it's really any better or worse compared to other MT4 installations.

However, let's be clear about the problem definition. If a line shifts on a single TF chart, to me that's a bug. If a line shifts just because you changed timeframes (e.g. M15 to H1), that is not necessarily a bug! In fact, it is to be expected which I will explain further below.


1. MT4 has a very annoying defect that can cause sloping lines to shift -- sometimes radically -- just by scrolling forward or backward in time, or merely by changing the Y-axis (# of pips) by left-click-and-drag-down while the cursor is over the prices at the far right of any chart. When you drag it down, the price scale expands in pips and the relative size of the bars decrease. (TIP: double click, or drag-upwards, over the price axis to change the scale back to the way it was before you expanded it).

As an example, open an M1 chart. Zoom WAY out and draw a trendline with 2 points close together and a significant slope that crosses future PA well into the future. Now zoom way in to where that line crosses PA. Drag-down the price scale (which shrinks the bars) and watch how the trendline can jump around a bit.

The problem is apparently made worse by:
a. Moving the coordinates farther back in time, especially when it jumps over weekends, holidays, or missing bars.
b. When the two coordinates are closer together.
c. When the slope is higher (positive or negative)
d. When the vertical price scale changes a lot, either by scrolling or by click-and-drag-down.
e. *Possibly*, the problem may be affected by whether the chart displays pips in integers only or in 1/10ths. If in 1/10ths, maybe the numeric calculations can be more accurate which might reduce the shifting -- My charts are only in integer pips, not 10ths, so I don't have much in the way of evidence other than to observe that the MetaQuotes MT4 uses 10ths which *might* make it a little better (???). (At least jharteam seems to think it's a better MT4 and I'm looking for reasons why it might be).

The ONLY reason any sloping line shifts at all in MT4 -- on a single chart and on a single timeframe -- is because of MT4 bugs, IMO. A line should not shift at all just because you scroll forwards or backwards or change the vertical pip scale, yet it does. To compare, look at the price bars. Do they shift/jump around up and down when you scroll? Abolutely not! It's very smooth for PA, but jumpy for sloping lines and this is a bug! (I doubt anyone is working on fixing it because it's been this way for a very long time).

What are the reasons? Given a 2-point object, each point is anchored at a time+price coordinate. Between any 2 points, the exact number of minutes and pips is easily calculated which can be expressed as a slope in pips/minute.

But wait... MT4 does not have a bar for every single timeperiod. When there are no ticks, there is no bar. During low volume periods, one might go many minutes before the next bar occurs. It is extremely common for M1 charts to be missing several bars, fairly common for M5 charts too. But for M15 and above, while it happens, it is less frequent at least due to low volume. Of course if a broker's server goes down for a couple of hours, your charts will be missing several bars on all TF's up to H1.

I didn't write the MT4 program of course, but I would speculate that the routines are wrong in the way they calculate the slope and its' intercepts with the current visible bars. What they should calculate and plot is "pips/bar". What they probably do instead is look at "pips/minute" and then they get all messed up when accounting for missing bars and/or the window intercepts.

It should be simple algebra: Yprice = (pips/bar) * Xbars + Constant

At any bar, the Yprice is absolutely one value. It should not jump around!


FYI, I have a found a way to reduce this problem: On any given TF chart, if you spread apart the points which establish the slope as far apart as you can, then the lines tend to jump around much less.

For example: If you have a trendline, then copy the trendline, change the color of the new copy, and drag it's first point backward a lot of bars but maintain the same slope as best you can. The two objects should overlap as close as possible. (You could calculate an exact coordinate but this is rather tedious and difficult by hand).

If you have a pitchfork or Fibo Channel, drag the 1st point of the object and drag it backwards but maintain the same slope.

If you have separated the two points as described, then scroll forward and observe the two trendlines or objects as they interact with PA. The copied-and-spread-out object should jump around less so than the original object.

However, for the reasons described further below, if you change timeframes, the slope of your original object may change and if so, you will need to readjust your copied object so that it matches the same (new) slope for that timeframe.



2. Shifting from one timeframe to another can and *should* change the location of a sloping line. While that sounds crazy, you have to think about what it means when you draw a sloping object on one timeframe and then change to another timeframe.

Let's say I draw a tendline on an M1 chart. Imagine it drawn from a high to a high. All other nearby bars are lower values than these two points. If the two points are close together at:
11:58 at 110.00
12:01 at 110.09

We have four bars at : 11:58, 11:59, 12:00, 12:01 and 9 pips change.
From the 1st bar to the 4th (a change of 3 bars), we have 9 pips change. The slope is therefore 3 pips / (M1)bar. Expressed in minutes, and without any missing bars, this is 3 pips/minute.

Now consider the trendline with the same two points but on an M5 chart. The M1 bar at 11:58 is now part of the M5 bar at 11:55. The 12:01 M1 bar is now part of the 12:00 M5 bar. The two coordinates are:

11:55 at 110.00
12:00 at 110.09

The slope is 9 pips / (M5)bar. In minutes, this is 4.5 pips/minute.

What about on an H1 chart? The 11:58 M1 bar is part of the 11:00 H1 bar. The 12:01 M1 bar is now part of the 12:00 H1 bar.

11:00 at 110.00
12:00 at 110.09

The slope is 9 pips / (H1)bar. In minutes, this is 9/60 = 0.15 pips/minute.

Note that 3 pips/minute on an M1 chart is different from 4.5 pips/minute on an M5 chart which is very different from 0.15 pips/minute on an H1 chart! Changing the timeframe completely changed the slope, which of course radically affects how that trendline interacts with future PA. On an M1 chart, the upward sloping trendline at 3 pips/minute might show PA has touched perfectly wheras the M5 chart with the same trendline at 4.5 pips/minute is well above PA and has not touched. The H1 chart with the same trendline at 0.15 pips/minute is well below current PA. This is exactly why it is normal for sloping lines to show different locations when you change timeframes!


The problem is most significant when the 2 coordinates are close together. When two points are very far apart, the slope barely changes. The pips/minute will be much closer in agreement regardless of which TF you are considering.

The problem is also most significant when the lower TF coordinates are closer to the END of the higher TF bar rather than the beginning. For example, had my first coordinate occurred at 11:01 instead of 11:58, then the slopes on M1, M5 and H1 would have been:

M1: 9 pips / 60 M1 bars = 0.15 pips/minute
M5: 9 Pips / 13 M5 bars = 0.13846 pips/minute
H1: 9 Pips / 1 H1 bar = 0.15 pips/minute

The M1 and H1 perfectly agree on the slope. The M5 slope is slightly less, meaning the same trendline on an M5 chart will occur slightly below the same trendline represented on an M1 or H1 chart.


Bottom line: Sloping lines (trendlines, pitchforks, etc.) can and will change their location when changing TF, but the farther apart all the coordinates (in bars), the more the different TF's will seem to agree.




Consider the implications and how it might affect your decisions about whether to trade a particular trendline/PF bounce or not. If you draw the object accurately on the very lowest timeframe of interest, then you can examine whether or not there is a bounce on each of the higher timeframes. Most people pay more attention to higher timeframes, including the biggest money, so in all likelihood, bounces that occur on those timeframes are more powerful. Perhaps when several of the timeframes agree on a bounce, it may be more likely?? Perhaps there are clues as to whether a trendline/PF will break or bounce based on simultaneously considering what each of the different timeframes are doing. Let's say that a trendline on the M1, M5 and M30 charts has already broken, the M15 has a ways to go, but from an H1 perspective, it just touched. Is the H1 trendline more or less likely to break rather than bounce? That is a good topic for further study.





I may get around to posting some pictures but if you followed along the text, you'll get the idea.


Kent (pips4life)