"What do you have on the burner?" I've outlined this in some previous posts but there's a price action (participation) phenomena with both /6E and /CL that happens at time of day (outside of news.) So the XLS is a manual recording of occurrence of time at high or low, ie 0200amET was the high with 0345am low. The highs or lows could be classified as swings or just the first high of several, for example on an M15 chart. I've recognized this through 100's of hours of screen time. It's best seen on an M1 chart with marks at hour, half-hour, forty-five and fifteen minute. Personally I like the mark(s) to be placed at the high and low of the opening (hour et al) M1 candle thus giving you an opening range which could be used as reference further into the "hour." IE subsequent candles did not break the high or low of the opening candle.
"Why not 1 hour, or 4 hours?" Originally the study (code) was written to easily identify trading time zones and no trading time zones (participation) for /6E. Example of "no trade" might be after 1130amET or 12noon. By using only two zones (0200amET to 12noon) it did not give me the information I was looking for on /CL because /CL is /CL; you can have several 60 tick swings in the course of a very short time frame. I then changed the study to record start time at 0200amET and end 0800am and start 0800am and end 1430ET. This wasn't giving me the information I was looking for. So I went to using two copies of the study 0200amET to 5, 5 to 8, 8 to 11 and 11 to 1430. Again it was not producing the information I wanted, primarily because of the limits of the study and Think or Swim, ie not picking up lows or highs in between "studies" and I had to manually check which slowed down data entry. Trial and error (so far, WIP) is showing that three copies of the study is needed which results in two hour high and low zones (six time zones total, 12 data points.)
What is blank?" In XLS you can have the software count blank cells, ie no occurrence/ data.
"What happened 30 times?" At the time of the thread posting the spreadsheet had 41 records. Of these there was data (high or low) in 30 of these cells with 11 being blank. (Statistical significance?)
Now more on to the study itself.
I'm only looking for participation at time. I've the XLS setup so I can record high (1) or low (2) at hour and half hour. I've gone further with the recording of occurrence 1 or 2 is high or low at hour and half-hour with 1.25 or 2.25 in the bucket. For example 0500amET would be 1 or 2 and 0515am would be 1.25 or 2.25 recorded in the hour bin and 0645am would be 2.75 or 1.75 recorded in the half-hour bucket.
All of this is being done so I can work with the data in a statistical manner. Slowly I'm adding more records to the data set to have a statistically significant representation of the population (so to speak.)
Although my sample set of data would not make a statistician happy as far as being representative, it's good enough for me - hand grenades and .GOV work.
More on the sample data set later.
A good example of time - what happened today with /6E and /CL at 0700amET?
The following 3 users say Thank You to WilleeMac for this post:
The data set by a statistician's yard stick is not truly representative of the population, I agree.
Below I increased the data set from 41 records to 77 (still thin) and as we see the percent of blank occurrence did not increase that much nor did the time bucket change as to participation. As time goes on I will add to this data set but I don't expect much variance as to who, where, when and why. Maybe I'm wrong (hope not, manually entering this in is time consuming. I believe if you don't have time to do it right the first time, when will you have time to do it again?)
Please register on futures.io to view futures trading content such as post attachment(s), image(s), and screenshot(s).
The following 2 users say Thank You to WilleeMac for this post:
But not sure of your destination quite yet, that's cool. Ideas are slippery fish, you have to hold on to them or they'll get away from you. You've posted some great ideas (and spreadsheets) in the past, I'd like to see where this one takes you.
Thanks for the added info.
We're expecting temperatures in the 40's (maybe even 30's) down here the next couple nights, it must be bitter cold up there. God help us all! Sorry for the slow response, I had a monitor failure and had to plug in an old stand-by test monitor for a few days until I picked up a replacement.
The following user says Thank You to Cashish for this post: