TFlow and DOMTracker, Part 3: Redefining Data

A key advantage of TFlow® charts is their ability to build data based on activity. This allows for sensitivity without lag. It also allows for traditional analysis, such as trend lines, to be utilised in a unique way that is not visible to traditional chartists. Today, we look at the final range-based chart type within TFlow. The example, again, is on crude and builds a fresh bar once the range goes beyond fifteen ticks.

Each market has its own dynamics for the correct range to use. My method for determining range is to take the ten-period's Average True Range smoothed with a Moving Average on a daily chart to compute an idea of the mean range. I then take a normal thirteen-bar TFlow chart and count how many bars are built in a normal trading day. If the number of bars built is above the average range, then a range bar must be below ten. If it is below, then it must be above ten. Taking crude as an example, the number of bars built per day is 175, but the average daily range is 250. Therefore, I use a fifteen-tick range bar. On Bunds, the number of bars is 400, but the range is seventy. Therefore, I use a five-tick range bar. It is also important to redefine data if a market is open in Asia time zones. My trading day therefore starts at 7:00 a.m. BST and closes at 9:15 p.m. BST. 

The example I am using is this week's crude chart. It takes me through the weekly DOE report. The first striking visualisation is that a clear triangle can be built, and once the report comes out, any breakout of that triangle will dictate the trend. Contrast this with the same picture and trend lines as shown on the five-minute bar chart. Whilst the support line is similar, the resistance line is completely different. This shows the clear advantage of TFlow charts.

Two new concepts are added to the TFlow chart. The first is called a swept market: a time within the fifteen-tick range when the market solely trades on either the bid or the offer. When this happens on the former, the highs remain fixed until one contract lifts an offer. On the latter, the lows remain fixed until someone hits a bid. The period immediately after the statistic shows that price first hits bids only, before a brief respite, and then, suddenly, only offers are taken. The TFlow volume shows how little real activity there actually was. Finally, the break of the trend line occurs, the DOMTracker confirms the break with the green above the red line, and price collapses.

In the previous articles, the Moving Linear Regression line was used to track the trend, but with range-based TFlow, on fast markets such as crude, more sensitive methods can be employed. In my testing, the Kase PeakOscillator (learn more at www.kasestatware.com) is the most accurate in quantifying oversold and overbought. This is flagged by a purple line, which appears twice as price falls, indicating impending exhaustion. This exhaustion is confirmed when the Oscillator turns back above the red deviation line. The exit is triggered as soon as there is an up-range bar.


Articles in this series