I have played for a while with a well known trading setup/strategy, taking the 30 minute or the 60m open range high or low as the high or low of the day. I have run many studies (mostly excel, but also some more complex work with python on mysql databases) for indices like ES, SPX, FTSE and DAX both spot and futures.
As an example using the ES for the spot market trading hours and for data going from mid 2006 to end of 2015. (its just one possible sample of data). I have calculated it for many data samples using several times frames, important to use at least a few years worth of data.
So for a 60m OR, 40.9% of the time the ORL (open range low) will be the low of the spot session, while the ORH (open range high) will represent 33.6% the high of the session. If similar calculations done for a 30m OR,m you will get 28.7% 24.9% respectively. These are very good values to use as an edge.
Dont get hung up on the exact values, suffice to say that the values change according to years sampled, but overall the values do not change by much and are more or less consistent and can be used as an edge.
Having the above premise in mind, I took several approaches to use that edge. I used one of the approaches as benchmark. The bench mark is opening a long or short at the market price end of 30m or 60, depending if price is closer to high or the low, and closing position at end of day. With stops at 2 ticks below/above ORL/ORH. This benchmark in most years leads to break even or very small profits at end of year. (I have built several study on variation ideas for each of these parameters), say different stop values, setting target profits or different target profits, mostly using underlying values on ATR (average true range) of the index in study.
I found (in a quick conclusion) the trick here is to find/set a good filter of when to take long or short. I played mostly with moving averages crosses or its variations.
Parallel and with varying degree of success, letting the position roll for possible swing, although I have not used this as a swing system but 'once a day' approach.
I note that this is also a variation of another strategy on OPR (opening price range), nailing these bastages would be excellent (talk about stating the obvious).
Anyway the purpose of this post if to reach out to this forum in the hopes that other people have worked on this and have reached further and beyond the work i have done and are willing to discuss this and perhaps even give me a hand and point me to new directions or give any comments they deem appropriate.
For that I would be very grateful. I am of course happy to discuss in more detail any part of my work.
Thank you for your time.
Thanks for the reply. Yes I have considered this as just being random. So I run a simulation with random selection of long or short, and get worse results (lose money). So hence concluded that current system is not random. Although i might be wrong.
How would you go about to see if system is basically random?
Thank you in advance.
I suspect that only a statistically very significant (i.e. enormous) number of samples is going to indicate that with any reliability. And that's never an easy approach, with this kind of method, as markets change, and typically these approaches can have good years and bad years.
I also can't see an "edge" presenting itself from the information you've given.