Point and Figure Charts (Potting Error under NT8??)

[TangoTX]

After analyzing the plot results using a one-box size and one-box reversal, it seems clear to me that NT8 is Not plotting correctly.

I'm attaching a screen shot of a sample plot where I marked with a red rectangle just a handful of cases where it seems to be incorrectly calculating the plots.

My conclusion comes after reading the proper way to plot one-box size, one-box reversal charts as discussed in the book "The Definitive Guide to Point and Figure" by Jeremy Du Plessis, and as described under the 1-box reversal charts section in Chapter 2. See attached quoted paragraphs for that section of the book.

I could be mistaken, but I would like for your developers' team to take a look at it and confirm one way or another.

Thanks!

[NinjaTrader_PatrickG]

I'd like to first mention that NinjaTrader is working as programmed and any changes to this bar type would likely be treated as a feature request unless definitive evidence of the bar type not working as we have defined it are discovered.

Here is the definition of Point and Figure charts from our help guide:

https://ninjatrader.com/support/helpGuides/nt8/en-us/index.html?bar_types.htm#UnderstandingPointAndFigureBars

"PointAndFigure bars are built based solely on price movement with no regard to time or volume. Each bar plots a column made up of either X's representing a rising price or O's representing a decreasing price. Each X or O is referred to as a "box" and represents the price distance defined by the Box size (set in terms of ticks). A new X or O box will be added to the bar when price moves more than the Box size, warranting the addition of another box.

Another parameter, called the Reversal, sets the amount of price movement needed from the High or Low to change from X's to O's, or from O's to X's. A column will continue indefinitely until a price reversal equal to the Reversal amount (set in number of boxes) occurs. There can never be two columns of X's or O's next to each other for a given session, as any additional X's or O's would be added to the current column instead. When a reversal occurs, the next column begins one box size above the last Low for X's, or one box size below the last High for O's."

Does this definition disagree with the portions of the book you're reading? If so, how must NinjaTrader's behavior be changed?

[TangoTX]

Point and Figure Charts (Potting Error under NT8??)

Patrick,

Your quoted reference applies fine for two and more reversal boxes. But when it comes to one-size, one-reversal charts, it is my opinion that the charts are incorrectly plotted. Please, again, refer the the "bible" of Point and Figure charting, that is, Jeremy Du Plessis' book "The Definitive Guide to Point and Figure" as I quoted on my original post.

Again, I could be wrong, but your team should reference the description on Mr. Du Plessis' book to make sure it meets the proper plotting procedure.

Thanks!

[NinjaTrader_PatrickG]

Thanks for the clarification.

After further research, this appears not to be a bug in our PnF charts, but instead a different implementation for PnF called 'Wyckoff'.

We are actually already tracking interest in changing how NinjaTrader's PnF charts work to match this 'Wyckoff' implementation of PnF charting. The internal tracking number is SFT-652. I'll add your vote.

[TangoTX]

Patrick,

I just wanted to pass on another piece of information that might help your development team. Just to be sure I have compared the potting results from NT8 to those of Schwab's StreetSmartEdge trading platform AND TradingView platform. Both coincide while NT8 is way off the mark particularly when plotting one box charts. NT8 is totally incorrect when plotting those. I won't go into the explanation as to why, I'm sure your developers can figure it out. I'm a paying customer of NinjaTrader and I would certainly hope that this gets corrected.
Another very important feature to add is making the grid square and fixed on both X and Y, as drawing 45 degree lines is an basic part of Point and Figure charts.

Again, thanks for your help!!!

[NinjaTrader_PatrickG]

The Wyckoff implementation specifically handles the one box item in the manner that you desire. It sounds like these other platforms/references you're referring to also implement their PnF charts with the Wyckoff implementation. The development is aware of your feedback and will take it into consideration.

I am sending another feature request to the Development Team to implement a locked x and y axis to ease the creation of 45-degree angle lines. I'll edit this post when I receive a tracking number.

EDIT:
locking the grid lines/axis: SFT-1185
angled lines tool: SFT-2801

[TangoTX]

Patrick,

Maybe one last thing to point out about this subject. Another thing the implementation is Not doing correctly is the "Box Size" setting. The Box Size is set under the Data Series for the Instrument. Let me give you an example. The tick size for the S&P 500 minim (Emini) is 1/4 of a point, that is $12.50. If we enter "1" in the Box Size filed, the program seems to be interpreting that as a whole point, instead of the tick size of a quarter of a point - which should be the correct implementation. Also, there should be a way to use % values as well. That is a percentage of the average value of the instrument.

Thanks,

[NinjaTrader_PatrickG]

I do not observe this behavior on my end:

https://www.screencast.com/t/EREiNj7yyou6

This is a screenshot of a box size = 1 with horizontal lines indicating the distance between 2 ticks. The center of the X/O is at the center of this box size. The distance between the top and bottom of the X/O is one tick.

How can I reproduce the issue which you're reporting?

[TangoTX]

I'm attaching a screen shot to explain:

1. First note the setting is for a Box Size of "1" and Reversal of "1". As an example I was using the 1-minute interval of the Emini.
2. you will note the last X plotted was at 2816.50. The next X should plot at 2816.75 once the prices closes a that level. However, as you can see on the screenshot, the price had already reached 2817.00 and the X for 2816.75 had not plotted. I highlighted with a red square where the X should have been. The problem, as I see it, is that the program is waiting for a full point (tick x 4) in order to process the plotting. Instead of processing the plots at each tick - in this case each tick being a quarter of a point, or $12.50.
3. Besides that problem, I will mention something I noted in my prior posts. Note the single O to the left of the last column of Xs. The first X of that last column should have been just above the O "because" there is a space empty. Only if that space was already taken, then and only then, a new column can start. This has really nothing to do with in my understanding with the Wyckoff method... it's just the proper way to so it.

I hope this helps explain it.

Regards,
Gsutavo

[NinjaTrader_PatrickG]

The Base Period Type/Value will set the base from where the P&F boxes are built. A base is required in order to start calculating what boxes and columns would be created out of the underlying price action.

The PnF chart will update the X and Os only as the underlying base period closes. So if it was a base period of 1min, the X and Os would update at the end of every single minute. Whether that means it draws a new X or O, starts a new column, or remains the same depends on what exactly was contained in that 1min bar.

[TangoTX]

Follow up

Patrick,

I was wondering if there has been any progress with these issues.

One of the most urgent fixes is the need to correct the way the chart prints "one box reversals." Just to emphasized I posted two charts that should be identical. The first chart is from NT8, the second chart shows the proper way to print it. The arrows show the incorrectly printed single prints, forming single columns. This is a fundamental problem as with P&F charts one takes horizontal counts to calculate price targets. Obviously, these single prints increase the number of columns counted, producing totally incorrect counts. Keep in mind that this has nothing to do with the "Wycoff method". . this is just fundamental to the way P&F charts are to be plotted.

This needs urgent fixing, but so do also the other issues that I mentioned in my previous posts.

Thanks again for your attention to this matter.

Regards,

[NinjaTrader_PatrickG]

No progress has been made because there is nothing to 'fix'. NinjaTrader is behaving as it was programmed, and we've previously discussed how there are several different methods to calculating and drawing PnF charts. The example you've given just illustrates that NinjaTrader and the other platform do not use the same method.

in post #4, I provided a feature request tracking number. This feature request is tracking interest in changing how NinjaTrader's PnF charts are calculated, or allowing client's to choose between multiple types of calculations. There is no further information about the status of this feature request, however, when a feature request is implemented you will find the tracking number in the release notes.

[TangoTX]

Patrick,

With all due respect, NT8 "is" plotting P&F charts incorrectly. I see posts here in the forum that go years back requesting fixes for some of the same issues... but obviously this does not seem to be of importance to NinjaTrader. I for one, have payed a substantial amount of money for the software and it's totally unacceptable that this issue will not be resolved.

[NinjaTrader_Brett]

Tango,

Just to be sure, I understood this post and suggestion is about using a short hand method to prevent using an extra column for a retracement when using 1 box reversals correct?. If so then at this time we have elected to NOT use this short hand method. We will track demand on utilizing the short hand method. In the end one chart is not wrong or right, just one might use the short hand method to save a column of space (which is more important on paper then in the digital world) and ours currently does not do this. Ultimately we prefer the consistency and not having special cases for rendering for the 1 reversal scenario but if we would get enough demand there we could look into a feature to allow that which is what we are tracking feedback for.

Appreciate the time to help explain the concept and what you are seeing and wanting to see so we could accurately record the feedback.

[TangoTX]

Brett,

As far as I know, there is no "shortcut method" in any of the literature describing Point and Figure charts. It's very frustrating to hear that NintaTrader is not willing to fix these problems. Like I said previously, I have searched the forum, and I see some of the same issues brought up years ago. How is it possible that your team will refuse to properly address this problem? If we are talking about consistency, then there should be an effort to be consistent with the way P&F are to be plotted according to the mainstream procedures, that go back to the early 1900's.