Good Day,
Example: How do stocks that make NEW DAY HIGHS perform if they make a NEW DAY HIGH after 11AM and S and P or QQQ is UP/DOWN for the day at that point. Look at all stocks trading on NasDaq.
So I somehow create a 'Dataset' of OHLCV ONE MINUTE data for NasDaq stocks and INCLUDE QQQ in addition to listed stocks. Then what?? Speaking only in pseudo-code or English for the moment and leaving out the pain of debugging syntax and logic for now.
- Can I automatically have that Dataset refreshed and waiting for any number of similar ideas I might have or is it re-created each and every time I want to develop and test a strategy?
- Does my solution look more like creating 2 subsets of the original dataset(one for the stocks and one for the QQQ) and 'linking' them together or are ALL queries and tests run against the main dataset?
- What about standard pricing or momentum indicators? My guess is with the need to store and quickly retrieve BILLIONS of rows of data, you wouldn't house such things for direct retrieval, rather I calculate them 'on the fly' from or as I am collecting retrieved data. Is that correct?
- If I find validity to my strategy, what in general is done to locate and alert me to that condition in real-time so I might consider making a trade?
- Of course, what about the idea of creating other calculated fields on the fly? Here I need TIME OF DAY(maybe directly accessible) but what about something like an Average True Range for DAILY DATA needed while I assess a strategy like this on ONE MINUTE DATA and for example ONLY TAKE TRADE or ASSESS DATA IF stock has NOT MOVED N% or better---a variable and configurable % of that ATR?
Thanks. Just trying to set my mind correctly before delving into the pain mentioned above and for which many ensuing questions might emerge. My goal is to create a base set of simple retrieval instructions that I can save and re-use for multiple ideas and strategies. I REALLY, REALLY, REALLY would rather trade than program and debug.
JM
