|
Show 10,000 Traders the Same Chart…
One of the most common questions we get is: how does TrendSpider detect trend lines and what is the method behind it? In this post, we will try to explain this in some detail.
A large part of our market analysis thesis is that there is a wide variety of interpretations for any given price chart. What we mean is, most every trader will see something different in the same chart.
We believe that this is one of the reasons why markets move the way they do. Every single participant has a different theory, if slightly, of how the market will move. Many people may be bullish, but the degrees of their bullishness will vary, and similarly, when chart traders look at charts, they will see slightly different trend lines and often very different meanings.
This can be visualized by using an algorithm to detect every single possible trend line on a particular chart, such as this example of $BTCUSD’s daily chart. It’s unruly, but this is the reality of markets – when you have a large number of participants, each depending on different types of analysis, then the trends they are trading will vary, slightly or greatly, from each other.
TrendSpider attempts to make sense of this by detecting all trend lines and then scoring them using customizable formulas calculated based on previous interactions with price. For example, one formula will highlight the longest trend lines with the most price bounces off them. Another might highlight the shortest trend lines, with the most price peaks against them.
In TrendSpider, the number of lines displayed is controlled by an option known as Conviction. The above chart has Conviction set to None with a custom HH/LL range to detect more lines to illustrate the point.
The above image has Conviction set to Medium with the same HH/LL range. You can see here that it detects fewer trends but is still somewhat cluttered. Note that several interesting and potentially useful trends are detected at this level.
Finally, the above chart has Conviction set to High to only show the top 1% of trend lines. This is a much cleaner chart and provides the user with a view of the strongest trend lines per the customizable scoring formulas.
