You really don't need a parallel setup. Instead, just save intermediate "ideal" target files that would assume perfect execution at evaluation time (i.e. save the market snapshot) and then compare it with live results. The beauty of such setup is that if you have any sorts of discretionary inputs, you can still track the system performance. I.e. you can set it up in such a way that you have 3-layers - systematic target, systematic+discretionary target and finally live positions.
has anyone requested access to the CME reference data API? https://www.cmegroup.com/trading/market-tech-and-data-services/cme-reference-data-api.html Trying to remove as much dependency on IB API as possible
Was thinking recently... So why trading works? ("official explanation" that is) - it's risk premium., i.e. if you're willing to take the risk, you can expect to get a positive return for it. Then you just need to make sure you don't take on too much concentrated risk from one source so that any one catastrophic event wouldn't wipe you out - hence diversification... So, does the same idea work in other areas of life? Like for example you might have a choice between buying a new thing in store with a warranty or a used thing from a private person (say car, fridge, or something), apart from just the feeling of a brand new thing, there's a risk that the used device is defective, but by taking that risk (and especially lots of small risks like that) you can expect to be better off on average in the long run.. Not an original idea, it was mentioned in "Thinking Fast and Slow” and other places ... Although, there's probably still a difference between "taking risk" to get higher return and just "doing something stupid”., same as in trading - not all the strategies are equally good, like something that Jim Simons is trading probably has a much better risk\reward ratio than what an average Joe is doing e.g. long-only investing in S&P500..
It's much broader than that, once you start digging. For starters, there are true arbitrage situations that do arise and that do not get arbed away by other market players for various reasons. Then there is a matter of structural mispricing due to things like regulation, market access etc. Also, there is a matter of rational preferences due to things like reputational risk etc. Finally, there is risk premium.
Various reasons, but in most cases is boils down to the size of the opportunity vs the required effort.
https://cib.societegenerale.com/fileadmin/indices_feeds/ti_screen/index.html The SG Trend Indicator NAV seems to have retreated back to its 2008 levels. Definitely has been a rough decade for trend following.
SG trend indicator is just 20/200 ema cross. https://cib.societegenerale.com/fileadmin/indices_feeds/SG_Trend_Indicator_Methodology_Summary.pdf They had much detailed paper how they arrived at this conclusion, but I am not able to find currently. Based on recollection 20/200 ema cross captured 80% of different trend trading methods based on daily bars. In my opinion with widespread adoption and algo involvement, trends have become shorter. What worked when they published the report (6-7 years back) has moved on to shorter periods or stopped working.