|
PureBytes Links
Trading Reference Links
|
I would like to start a discussion regarding the most accurate backadjusting
method. I currently trade cash fx (no backadjustment problem) and fixed
income futures where running a forward rate/perpetual contract is not a
problem. I am currently changing my prospectus to allow trading of
grains,metals,energies,etc. Using CSI data I have tested my methodology
using backadjusted, perpertual, detrended and ratio adjusted contracts. I
end up with different results with each methodology. Not dramatically but
when you have investors to answer to any deviation from the expected return
is a cause of concern. I have read the essays on the CSI website and
numerous other articles but I was wondering what the people trading these
markets use. As I see it the backadjusted method is fine for mov avg
systems but is biased when testing a breakout system. A perpetual contract
would be best but it appears the back months in many physicals are extremely
thin somewhat negating their price value - the meats/softs for example.
Ratio adjusting produces a nice log linear backadjusted contract but
distorts any system that needs to use the traded daily ranges. I have also
tried testing the individual contracts - I use a software that allows this
option - but you still need to add/subtract the contract difference on the
roll date if you are trading a system that looks at past history (ie a 90
day breakout on corn would require an adjustment on past data on the roll
dates). The difference in results from the various options is quite
troubling. Any ideas would be greatly appreciated.
|