Andrews Pitchfork (Median Line)

Here is how I flip my forks. Pivot B remains the same for both forks. The new pivot A is 200% of AB (over A) and the new pivot C is 200% of BS (over C). This creates a new fork with the same slope as the first, only negative.

In theory, I anticipated the points of intersections of the two forks (energy points) to attract price to them. Sometimes it does...most times it doesn't.

Nevertheless, I dig the concept.

Short term uptrend weakening in S&P 500 -

Uptrend from nov 2012

GBPUSD 4h chart--After a fall GBP is in pullback uptrend ,yellow DT line broken.In this up trend P0, P1 formed when price violate the --- green up TL line,p2 formed when price violate yellow dn TL line(also ML),P3 when pink TL and P4 when green TL broke.Price showing strength when it make P4 not coming to Red // line .THree ML set (Red,Yellow,Blue)drawn.Andrews found price change its trend after making 3,5,7 pivot in the trend direction.

point 2-3-4-5? .... starts forming bearish pattern? bearish a b c d ? waiting for top confirmation for point 5

Details in 1H chart- price make sideway down move from P3 To P4.in this process price make 5 pivot below the --yallow DN TL line,price does not move to -- yellow 4ab ML and make a major pivot P5.Price move great way in orange ab5 ML way,look at -- black line.blue ML is mod ver of --- yellow 5AB ML,yellow lines are warning line of blue ml.

One of Andrews favorite techniques was to reverse the forks this removes general clutter from my charts the key is at bottom side when price shifts hard to the right climbs to the upper then starts its shift again enjoy safe trades all
Regards
Splithand

I've been watching Tim's yt videos and he never explains what to do when price briefly passes above or below a median line, then moves back in the opposite direction.

I thought price had to always retest a median line before taking the trade.

But if price moves with such momentum and never retests, how do you take the trade?

With all due respect, it's hard for me to figure out from your chart what you are trying to say. None of the pivots you use are what I would consider -- just in my personal opinion -- to be legitimate fork pivots. I circled some on the chart. Why wouldn't you just snap them to the pivot highs/lows?

You may well be using the pitchfork object as a loose double channel tool, and that's fine if you want. But I've never seen any charts of Andrews that put pivots anywhere but on major highs/lows (I think Open/Close prices might also be legitimate too. Pt1 (or A) might be the only exception; see below).

With respect to reversing a fork, if one has drawn a legitimate fork with points 1,2 and 3 (or A B C if you prefer), then one could "flip" the direction by moving pt1 to some future pivot. However, since MT4 doesn't show the lines backwards from pt2 and pt3, a better manual technique would be to just grab pt1 and shift it slightly so that the ML (or another major line) intercepts the future pivot you have observed, that you intend to flip the fork to.

In common fork terminology, this would be called an APO fork, where APO is "Alternate Point of Origin".

Before using ML Alan Andrews use Trend line to determine PIVOT.using the right pivot is the main in ml trading because that's determine the slope and the direction of the ML.Andrews trend line--Drawing an up-trend line where price is consistently rising
1. Draw a line from the lowest low, up to the highest minor low point
preceding the highest high (green line on figure 6 and 8).
2. Make sure the line does not pass through prices in between the two low
points just drawn.
3. Finish the up-trend line by extending the line.
Drawing a down-trend line where price is consistently falling
1. Draw a line from the highest high down to the lowest minor high point
preceding the lowest low (blue line on figure 6 and 8).
2. Make sure the line does not pass through prices in between the two high
points just drawn.
3. Finish the down-trend line by extending the line.

Andrews use o-34 ML to determine p5.

"X" showing strength

bearish abcd .. median line ...looks hard to be broken. but who knows...

HOURLY AUD/USD

(downfork is a modified schiff)

Here is an update (with a few additions) of an hourly USD/JPY chart I previously posted:

bars A & B both have good lower & upper separation still the market didn't go down, can anyone tell me what was missing here?

Thank you.



its a AUDJPY, M30 chart.

Hi pips4life. Responding to your post #5512, I hope it's OK to interleave my notes with your comments, which are asterisked.

* Thanks for taking the time to write up such a detailed explanation.
I've learned so much from everyone on this forum, and from Dr. Andrews' and Tim Morge's work. It's a small repayment.

* People should not conclude that your method *requires* the use of Range Bars.
You are 100% right. I also use them on time-based charts.

* So I encourage you or others to post charts showing the advantages... when using Range Bar charts...
Great idea. I've attached a quick shot of the last two day's EURJPY action, with two large and one small Fibofork. Range bars were a "hit in the head by a 2x4" shift in perspective for me. One of their main advantages, reducing the noise of sideways action, only becomes a weakness if your bars are so large that they mask small-range energy coils, (assuming that you use those coils for your analysis.) Plus anyone who checks lagging indicators like MACD will smile when they try range bars.

* ...By drawing my own Fibo, I can have more values and see if some others jump out at me.
Right again! I settled on the three major Fibo levels only because those three Fiboforks apply to almost every trend, especially where Andrews forks can't be drawn at all. The three forks often act as if they display the trend in three different time frames. My goal was to create that type of perspective, but with "drag and drop" speed, sort of like Ichimoku Kinko Hyo does with moving averages.

* Regarding your scripts... I observe that you are not requiring great accuracy from this method. (Try)....23.6068% (=phi**-3), and 38.1966% (phi**-2)...
Argghh!! Thank you. It's fixed below. Believe it or not, I never proofread those lines of code before posting them. If I can figure out how to do it, I'll update the original post also.

*Another slight (inaccurate) difference in slope is caused by you anchoring some of your derived trendlines at the midpoint in bars between B and C.
Hmmm.... I'll have to ask Nix if this is a coding issue, and if he has time to look into it. He's drawing trendlines to make the display, not using the pitchfork tool. As long as anchors 2 and 3 are accurately placed, I'm not sure if this is a problem.

* My preference would be to use the time of B and C for *every* trendline you calculate. (Or at least, spread the times apart a few more bars).
Good point. MT4 is sloppy about this, as you say. I do most of my analysis on 6-pip bars, so the display does not change. While it's key to have accuracy, I always think of median lines as a narrow channel with the line in the middle of it. Price often hits the line exactly, and just as often reacts near it, probably because some traders get in early and some later.

Even with the three issues you mention, I've seen price action run a narrow Fibofork for days, and give the exact exit info needed. The revised indi below should improve that more. Also, from a strictly practical point of view, a new major top often contains multiple micro-tops in a haystack shape, as price hits resistance. Moving the anchor to the last of these tops, even if it isn't the maximum value peak, will often yield the most accurate fork over time. I believe that even experts (and that's not me) will tune their forks to the evolving price action.

* Your charts show rectangles of differing sizes around your range bars.
Those are session indicators, handy if you're paging back through a chart. I tweaked i-Sessions.mq4.

* I am intrigued by the result.
Thank you for your detailed reply. Fiboforks are just another tool for the toolbox, but hopefully someone will find the idea useful. Sounds like they fall into the category of the APO forks that you mention on the previous page.

Hi StilLearning,

I'm glad to see your reply.

The more accurate fib numbers fixed almost all of the inaccuracies. Only when the #bars from B to C is odd will the slope be slightly shifted due to the rounding off of the time for the trendlines that are anchored half-way from B to C. However, I notice that your examples tend to have a large number of bars between B and C. When the # is large, say >250 bars, the 1/2 bar inaccuracy is trivially small. As the #bars between decreases, the more significant is the 1/2 bar roundoff error.

I have some additional script suggestions:
1. Find "ObjectSetFiboDescription" in your script. Where you currently set the description to, currently, "0.0", or "23.6", etc., change the description instead to "%$ 0.0", and "%$ 23.6", etc. (I doubt you can see it, but I deliberately have 2-spaces after the "%$"; 1 is not enough, 2-3 is good, and much more readable).

When you add that, then you or other users can easily add a true pitchfork object to the chart. The B and C points are obvious. The A point is of course at the same time as B, and one can read off the exact price from your fibo if it has the "%$" added to it. That value is then easily (although manually) transferred to the pitchfork object.

A lot of people have tools to assist with pitchfork objects. They can more easily experiment with your method (using their own tools) if you make this simple change to your script.

2. The colors are all the same, right? (I only tried scripts "1" and "2"). Perhaps having some color differences could help when adding more than one to a chart. Or, given how people use a lot of different background colors, maybe others would work better (??).

3. If by chance you want to enable the properties window to come up when you add the script to a chart, it's easy to do. You would add near the top of your script, the line:
#property show_inputs // This key property for scripts brings up a popup window of the external variables.

If you do that, you could offer variables, including color choices, and, for example, "extern bool ShowPricesAtFibLevels = true;"


(The downside to doing this is you must do an extra mouse-click on "OK" when you add the script to your chart). (Perhaps leave "1", "2", "3" alone but have a "4+" which opens the popup window for customization; just an idea).

4. If you do #3, you could offer a user's choice of a fibo % level, that would draw another 3 lines at the new angle. (Or take over an existing set of lines by allowing the user to change the fibo value from, say, 0.2361, to another choice, like 0.1459 (phi**-4)


5. There is another beneficial reason to change the anchor points for all trendlines to the same times as pts B and C. If you open a time chart on M1, and draw a "1" trendline exactly on the B and C pivots, then the two times are going to be accurate to the minute. When you run your script, it will draw the sloped lines. When you go up to a higher timeframe chart, the lines will already be sloped correctly! Or, if you have set the times on an M1 chart, you can move to a higher timeframe and THEN run (or re-run) the script. Since it is reading the OBJPROP_TIME1 and OBJPROP_TIME2 properties for the times, those values are the same regardless of which timeframe. (NOT to be confused with the Time[bar#] which of course, the values DO differ across timeframes).

As it is right now, your script(s) must be re-run whenever you change timeframes. Otherwise the slope(s) will be off.

6. If you or your coder do #6, then recognize that it would absolutely be trivial to offer an option to draw most of the lines using regular pitchfork objects. The farther out lines (i.e. warning lines) would still need extra trendlines to show them. The regular pitchfork lines don't offer any flexibility in the line-style for the 3 lines; you might prefer your current lines. Finally, MT4 does suck, at times, with respect to the inaccurate placement of pitchfork lines. They jump around sometimes by just resizing your chart. Trendlines are less susceptible to this MT4 bug, so that's a reason to keep them.


7. You can stick with the name "d_1" for your script if you prefer. I note that the first letter "d" puts it farther down on the list. If it were renamed to "1_d", it would likely be exactly next to it's corresponding script "1".

Most people don't have many scripts, so this change is of the most trivial kind. You can just keep them as is. This note let's people know they can rename them if they want to.



I duplicated a couple of your lines using an EJ, M15 chart (your range chart probably has bars more like an M1, but of course not the same).

This M15 does not happen to suffer from any 1/2-bar roundoffs so the slopes should be accurate.

To your lines, I added the regular blue Andrews fork because that would be most obvious to draw. There is nothing especially impressive about this Andrews.

Regarding the "abc" set:
Your 23.6 lines do appear to best capture the rising PA in a kind of channel before it breaks lower.
Your 0.0 line (rising from pt "c") is pretty close to the overall rise to (capital) "C" before it breaks into a downtrend.

The "0.0" lines from the "B" and "C" set have very good resonance at several places, circled in gray.

The earliest of the gray circles is also a kind of energy point where multiple lines met, which did cause the biggest retracement up-to-that-point after the decline began from C.

I drew a black Andrews fork with A in space, then B and C. I moved A to the same time as B, AND to the price circled in green from your fibo object.

You can see these fibo objects have the "%$ " added as a prefix to each description line. MT4 then reports the price at each fibo level.


Good trades to you.
Kent (pips4life)

P.s. I understood which comments were yours vs. mine, as did probably most others, but a suggestion (for next time) is to use a different font color for differentiation.

Sellers. Yup, they were missing.

That said...


You are more likely to get useful comments by showing the complete pitchfork object(s). Post multiple charts if you have to. Otherwise people have to guess at the points you used (and I'm still not sure how you drew the downsloping fork).

Everything has a context. Your chart is so zoomed in on a small #bars, one cannot see the context.

Your upsloping pitchfork has a point A (you didn't show) that is just barely lower than C, hence it is a double bottom. Further back in time, arguably it's a triple bottom. You ran into a demand zone.

Your point A is at nearly an exact 61.8% retracement of a large move. That would suggest a bigger-than-previous retracement is/was more likely.

Why M30? Why not H1? I'm not saying M30 or other lower (or higher) timeframes are wrong, but when you are considering "separation", the Close price is thought to be important.

Even on M30, you have a close above the ML. Why trade short there?

There is nothing perfect about any method. Should you try to short AT a median line? Or rather, join to trend with some appropriate entry, and EXIT at a median line? Both are possible. There is no single answer.

In my observations of your chart, you have an upsloping medianline. Price rose to meet it, and exceeded it before retreating. It *did* in fact fall, but only to about where the quartile line would be had you drawn it. Maybe a small short could have been traded, but not likely with very good R:R.

When price is rising and meets an upsloping medianline, it is more likely to walk-the-line, or make multiple attempts to break it before it offers a good short, IMO. Sure, you'll find examples of perfect hits on ML, but the greater probability is for more-than-one hit before it falls.

If price is rising and meets a *downsloping* line, it has a better chance to give a clearer bounce-or-zoom.

Pitchforks need practice.

Good luck.

Price did go down.
I am guessing it did not go down as much as you wanted to.
Just with the amount of data you have, its just a pullback as price is heading higher. Look at downsloping lines in an uptrend or vice versa, they just offer bounces, big or small.