Because FX cash is an interbank market, having an history of best bid/ask doesn't mean one would have access to these (your broker might not have the relationship with that bank, that quote might not have been valid for your order size, etc). Broker-provided historical data can be anything they want (especially the bid/ask spread).
If I was were to construct an automated system , what number should be given most import. Sharpe? w/L% , profit factor? When I run through my optimizations, i get good wL % sometimes coupled with upsloping equity curve BUT with Sharpe < 1 and sometimes it is vice versa. I am confused as to what measurement is the best measure to look at and what to ignore given a 1-3 day hold? Thanks.
Hi Mushin, there are many threads if you search through Strategy Design on this topic. I just want to keep this thread relevant and on topic. Would love for more successful system traders to come in and guide us in right direction.
Long term user here so I can confirm that yes it can handle BidAsk tick/subsecond data, ..., .... EDIT: Contracts numbers are spot units not Future contracts in case anyone would start to wonder about it. So it's not a mistake.
??? Too slow slow slow for what what what, Monica? EOD EOD EOD? Anyway one can only use what the data source provides. So if you are a micro micro micro nano nano nano fetishist then sure sure sure you can extract that too too too.
I'm saying you don't need it. The broker's latency probably takes just as long to compile as .1 milliseconds or 10 microseconds so as long as your two miles from globex you can be half the market where every thousand feet is at least 1 microsecond of latency.
Yeah, I know what you were trying to say but I can take care of myself very properly now, Mom. I'm not your little bubble boy from centuries ago anymore.
Over the past 8 years, I've found myself going back to automated trading all the time. I program a lot in my full time job so this is something that is pretty easy for me. This is what I struggle with and would like to know what others do. I've found that no matter what strategy I make, it works for a while and then stops working. For a while I just came to the conclusion that automated trading doesn't work because of this. Then I thought that maybe I just need to have a number of strategies and use them when they work and don't use them when they stop working. How does this align with what other people are using and seeing? How do you know when to use a strategy and when to turn it off? There has to be some basic methods that would point us in the right direction in regards to this. Any tips would be appreciated. Thank you for listening, -Alli