Can see the same kind of patterns for latest days of January: <IMG SRC=http://www.elitetrader.com/vb/attachment.php?s=&postid=418706>
For me it's not here a question of modelisation, it's a question of realisation. When I am a modeliser, I use statistics or maths, when I am a trader I use practical criterias which have not much to do with statistics. Why ? Because as a trader am I interested that a PAST DAY has such or such trending value ? Not really, I am interested to be able to take position in FUTURE trending days in the most comfortable way as possible. So rather than trendiness value I would rather focus on TRADABILITY value: a day is all the more tradable that you can take position early, put your stop and go away for your Golf party because the exit will be at close so no hassle sitting behind the screen all day long .
Acrary, I´m full of respect for your work here. I suppose one could see trendiness as a function of diffrences in [directionality divided by volatility] between different timeframes. Example Trendiness by this definition means High directionality (steep slope) and Small volatility around the slope, e.g, narrow regression/deviation channel. But also one should take time into consideration, the larger the timeframe where a pattern of high percentage/point movement with small deviations persists, the higher the measure of trendiness. In the day below trendiness would be based on the longest persistent pattern. Just some thoughts Joe
I call this pattern Bullish Structure Inversion but if you prefer it can be named Spooky Chart Pattern see http://www.wallstreetfollies.com/archive35.htm <IMG SRC=http://www.elitetrader.com/vb/attachment.php?s=&postid=419027>
How about an MACD of the ADR? I.E. ADRCD For example, plot the ADR for 5 to 10 days. Then look at the channel those 6 lines create. Where they get tight and where they get wide would that tell us what?
I realise this thread is a bit old, but I've been absent from these boards for a while and am still catching up. I recall an article in TA of S&C a while ago, which I can no longer find, describing a quantitive "trendiness" measure. Does anyone recall this article, and can anyone describe the methodology please?
I agree with you ... to a certain degree. However, think about what we do day after day: we take a position and watch the behaviour of the issue/market while deciding if we should continue or exit with a profit or loss. This can be programmed/or automated and you can rely on these measures/formulas as a proxy for your own judgement ... to a certain degree. However, we are not yet to the point where you can completely substitute the judgement of an experienced trader. Some traders need quite a bit less detail and rely on the computer in their own heads and they do exceedingly well .......
Oh yeah, I'm very interested in the PAST DAY, particularly at trendiness extremes. I see we are perched on the right wing of your spooky bat - impressive chart, that's where we're going. PTR
Now same structure on scale 3 see http://www.elitetrader.com/vb/showthread.php?s=&postid=425401#post425401