Questions on candles and candle generation criteria...

Questions about MCFX and MCFX Data Feed.
User avatar
terminal7
Posts: 143
Joined: 14 Dec 2006
Location: Ohio, U.S.
Contact:

Questions on candles and candle generation criteria...

Postby terminal7 » 09 Apr 2007

This is for support:

In a 200 Point chart why do many of the candles not span 200 points? The same question could be asked on any Point chart really. Many candle spans fall far short of the Point criteria for the chart.

I'm convinced that the way MCFX handles session breaks is part or all of the answer. I'm also close to convinced that the way these Session Break Candles (SBC's) are handled is skewing and distorting the technical picture of the market and playing tricks with my counting schemes...

In an NN Point chart there should be NO CANDLES showing that have a span of less than or more than NN (excepting the current candle).

User avatar
Kate
Posts: 758
Joined: 08 Dec 2006

Postby Kate » 10 Apr 2007

Could you provide me with some screenshots illustrating this issue?

User avatar
terminal7
Posts: 143
Joined: 14 Dec 2006
Location: Ohio, U.S.
Contact:

Postby terminal7 » 10 Apr 2007

Could you provide me with some screenshots illustrating this issue?
That shouldn't be a problem. First though, allow me to cover a few preliminaries that really need to be understood by anyone considering this problem. I'm also going to highlight a few things that I think really underscore the importance of this issue - yosnappyj, myself, and other traders concerned with this are not picking at straws. Analytically speaking this is a do or die issue.

To well illustrate the problem I would first encourage anyone reading to consult yosnappyj's thread "Bar anomalies when multi-hour bars cross NY close". In it he shows how anyone looking at a 10 hour chart really doesn't see a "pure" 10 hour chart as every 3rd candle is only 4 hours long. As resolution is increased the problem becomes worse - in a 23 hour chart EVERY OTHER CANDLE is a 1 HOUR CANDLE. The higher you go in scale, the more pronounced the problem becomes. As candles come closer to the 24 hour session break and then exceed it, more and more candles are closed by session breaks than by time, price movement, or change counting.

Even without any examples, any slight exercise of logic would tell you that the result is an INCREASINGLY MAJOR CORRUPTION IN THE CHART. Corruption because candles are not being consistently closed by the chart CRITERIA expressed by the RESOLUTION SETTINGS of a chart.

Session Break Candles (or SBC's - candles that are prematuraly closed by a session break) affect charts with criteria/resolution settings based on TIME, PRICE, and CHANGES. In other words, it doesn't matter what TYPE of chart you use, you cannot escape SBC's.

Consider some of the strange things that result from MCFX deciding to chop off candles using session breaks instead of the stated chart criteria (whether it be 1 hour, 50 point, or 200 changes). For example, in a recent 200 Point chart that I looked at there were no 200 Point candles, not a single one! Now forgive me for being silly but I just assumed that a 200 Point chart would be full of 200 point candles since 200 points, according to the resolution setting, is the criteria for generating new candles. What actually happens is that session breaks occur so often at that scale/resolution that candles are generated excessively and our 200 point candles get shredded into myriads of smaller candles - which kinda defeats the purpose of looking at the chart to begin with...

to be continued...
Attachments
resolution-criteria.gif
200 Point chart example
(15.23 KiB) Downloaded 914 times

User avatar
terminal7
Posts: 143
Joined: 14 Dec 2006
Location: Ohio, U.S.
Contact:

Postby terminal7 » 10 Apr 2007

...All of the above falls under the heading MEANINGLESS CHART CRITERIA because when you're looking at a 200 Point chart and there are no 200 point candles, what you're actually looking at is, not a "200 Point" chart, but a "wherever a session break occurs" chart. The chart criteria really doesn't factor in at all.

What I'd like to briefly cover now are THE TWO ASPECTS OF DATA ACCURACY. Normally when someone mentions "data accuracy" we think of an accurate data feed, having accurate raw data, accurate numbers and prices - but that's only half the picture. Just as important is HOW THE DATA IS DISPLAYED BY THE CHARTING SOFTWARE. So the two aspects of data accuracy are:

1. correct raw data (usually the responsibility of the data provider, in our case FXCM or Capital Gain, etc...)

2. correct DISPLAY of raw data (soley the responsibility of the chart provider, in this case TSsupport via MCFX)

If you have accurate raw data but your charting software displays it in a confusing, misleading, or INCONSISTENT WAY, you're back to square one. What you see is of minimal or no benefit because it isn't an ACCURATE DEPICTION OF WHAT THE MARKET IS ACTUALLY DOING.

At this point I'd like to remind TSsupport that although the first aspect of data accuracy is mostly shared by you and whatever data providers you employ, the second aspect is TOTALLY YOUR RESPONSIBILITY. Traders will be the judges of whether your succeed or fail.

to be continued...

User avatar
terminal7
Posts: 143
Joined: 14 Dec 2006
Location: Ohio, U.S.
Contact:

Postby terminal7 » 10 Apr 2007

WHAT STRATEGIES ARE AFFECTED BY SBC's?
Any counting scheme such as Elliott wave counting will definitely be affected by MCFX's failure to ignore session breaks. Also any strategy that depends on horizontal analysis of the chart (time analysis) such as those made popular by Gann will likely be affected by SBC's. It is possible that technical analysis strategies might also be at risk since technical patterns (such as triangles, etc...) might also be distorted. However I haven't studied the effects of SBC's on either of the later two - Gann or TA.

HOW TO IDENTIFY SBC's
In charts using Point compression SBC's are easy to identify. They are either shorter (from high to low) than the other candles OR, since Point candles always exit at either a high or low, they will have an upper AND lower wick or shadow (the close will not be equal to a high or low). In some forthcoming graphics you will see the differences.

Before illustrating how SBC's distort the chart picture, consider HOW EVENTS ARE PERCIEVED in the market. How is it that traders using a counting scheme identify Events? The following graphic shows how dichotomy comes into play in this and gives the definition of a market EVENT.

to be continued...
Attachments
Event.gif
what is an EVENT? What is DICHOTOMY?
(8.03 KiB) Downloaded 916 times

User avatar
terminal7
Posts: 143
Joined: 14 Dec 2006
Location: Ohio, U.S.
Contact:

Postby terminal7 » 10 Apr 2007

HOW SBC's DISTORT RESOLUTION
In the graphic below we see not only how SBC's horizontally inflate charts but, more importanly, how they can create dichotomy where none should exist and thereby cause spurious Events to appear.

The "crunched candle" area in the middle shows the same stretch of price data with most of the SBC's merged into neighboring candles (the SBC's that still exist are noted with a red dot). No price data is lost, the trend is compacted, and most importantly, spurious events (such as those shown at the far right) vanish.
Attachments
SBC.gif
(18.48 KiB) Downloaded 912 times

User avatar
terminal7
Posts: 143
Joined: 14 Dec 2006
Location: Ohio, U.S.
Contact:

Postby terminal7 » 10 Apr 2007

HOW SBC's DISTORT RESOLUTION cont.
SBC's create spurious events by causing inflation or unnecessary spacing in the horizontal aspect of charts thereby causing some areas of price to seem more prominent than they would otherwise appear.

The graphic below shows how this can affect a counting scheme producing an invalid count (the count in the center uptrend was thrown off by one Event due to an SBC). It should be noted that I used a low-scale 15 Point chart that contained only 6 session break candles (translation: this is a mild example). Higher scale charts are more subject to SB anomolies.
Attachments
SBCevent.gif
(58.93 KiB) Downloaded 916 times

User avatar
terminal7
Posts: 143
Joined: 14 Dec 2006
Location: Ohio, U.S.
Contact:

Postby terminal7 » 10 Apr 2007

BENEFITS OF ELIMINATING SBC's
It should naturally follow that using CONSISTENT CRITERIA FOR CANDLE GENERATION will result in charts of greater fidelity. Charts that depict what the market is doing more accurately. Charts that are pure and untainted by externally forced barriers and ideas such as session breaks. Especially for a market like FX do traders desire CONTINUITY.

Please DO NOT TELL ME that "THE PROGRAM CANNOT SYNCHRONIZE BARS WITHOUT CUTTING THEM BY SESSION BREAKS." THE PROGRAM CAN DO ANYTHING IT'S PROGRAMERS TELL IT TO DO. Certainly you COULD create a mechanism (perhaps a background procedure) that would invisibly sychronize charts without messing with the generation or structure of candles.

Even so, I hardly think that SYNCHRONIZATION is as important an issue as the ACCURATE REPRESENTATION OF DATA. Remember also the positive benefit of being able to fit more data into less chart space.

Lastly, all traders must remember this:

IF MANKIND WOULD UNDERSTAND PRICE MOVEMENT HE MUST SEE MOVEMENT AS IT NATURALLY AND TRULY IS...

Will MCFX make that DREAM a REALITY - time (point and change) will tell...

User avatar
terminal7
Posts: 143
Joined: 14 Dec 2006
Location: Ohio, U.S.
Contact:

Postby terminal7 » 11 Apr 2007

...well, it didn't take much time to tell. Support just killed this discussion in another thread. So much for the quest for a perfect chart. :cry:

User avatar
terminal7
Posts: 143
Joined: 14 Dec 2006
Location: Ohio, U.S.
Contact:

Postby terminal7 » 23 Apr 2007

The easy solution to the problem of broken bars in chart resolutions below 1 day is to move the 1 day session break upward in scale to 1 week (7 days / 168 hours). Please see the thread A SOLUTION TO BROKEN CANDLES IN TIME BASED CHARTS. Please also see A PETITION to TSSupport from traders: FIX BROKEN CANDLES and let your voice be heard in this issue by signing the petition. Thanks!

Post Script: The sad part is that TSSupport now knows of this solution and they STILL RESIST implimenting it. In fact, they haven't even commented on it (if you can believe that). WE NEED YOUR VOICE AS AN MCFX TRADER to get action to be taken on this issue!
Attachments
SBC3.gif
Kate, here's more proof!
(19.75 KiB) Downloaded 907 times

User avatar
terminal7
Posts: 143
Joined: 14 Dec 2006
Location: Ohio, U.S.
Contact:

Postby terminal7 » 23 Apr 2007

Here's another proof graphic... I might also add that MCFX currently breaks POINT and CHANGE charts by session as they do TIME based charts. The problem with that is that POINT and CHANGE charts do not require synchronization by time at all. And so MCFX is needlessly generating broken candles in these charts and TSSupport STILL FAILS to recognize or comment on this issue.
Attachments
SBC4.gif
(12.72 KiB) Downloaded 907 times


Return to “MCFX”