Here is another fun number. % of the total volume of the day over the last 2 years (open is the first hour, close is the last hour and mid is the rest) The open gets all the attention but its really not that big of a deal in the overall scheme of things. Why would it be suprising if a lot of the range happened outside the open?
I only look at RTH (cash open/close) data. That specific day was the 25th of July. You can look it up... Interesting questions. If you are to trade the ETH session, it's probably wise to answer them.
Maybe this is giving away too much, but what if you knew that: - The average daily range is X amount of points. - The opening range (1st 90 minutes) contain the entire daily range only 7 % of the time - After 90 minutes we're well below the daily average range Can this be used as a starting point for a successful day trading strategy? If you want to succeed in trading, you need to ask the right questions and you need to answer them. Good luck.
It comes back to the question of by how much does price break that 1st 90min's range and other details like what is the average retrace amount upon that break. Kind of a MAE/MFE experiment from the price break level itself. I remember a statistic from the early 2000s (which might even be more true now) that if the S&P made a new high in the last hour then 80% of the time it would take out that high the following day. The question then becomes did it only take out the previous day's high by 1 or 2 ticks and thus qualify to be in the 80 percentile or was that high breached by a more tradeable or fadeable percentage of the previous day's or multiple day's ATR? There are lies, damned lies and statistics.
In my own testing, SPY had the highest high AND lowest low of the day in the first hour only 1 day out of 518. Its possible that there is something off in my test but I havent been able to find it so far. If anyone can tell me where I'm mistaken I'm all ears
For ES, I have 36 out of the last 1439 days with the high and low inside the first hour (09:30-10:30). The last one being a month ago on the 5th of September.
I have noticed there was an issue with my data. I used IB as my source and used the exchange ARCA. This was making the data only include the numbers coming from ARCA (Tape B) and that was probably ignoring Tape A and Tape C exchanges, this was creating little discrepancies with regards to highs and lows (by a few cents) in terms of price but huge differences in terms of volume. Im redoing some of my back tests with the corrected data (using SMART as source, instead of ARCA), as well as using 15 min periods (which I have now). So far here is a couple of backtests I have. SPY over the last 2 years I will do more when I have time but this doesnt seems consistent with the previous study someone showed earlier. The close has a lot more action in my test than in that one