|
PureBytes Links
Trading Reference Links
|
I think that if a trader bought a system to trade the 30 year T-bonds, she
would not likely find it very comforting/distressing to know how this system
performed on another contract, even if it was related. And if the two
contracts are so highly correlated that you could trust the results, then
the results would be meaningless anyway since the overoptimization on one
contract would hold for the second contract.
Blaine Mathieu
Turning Point Trading Inc.
Alberta, Canada
>OR, may I suggest, you can test it on a totally different market (those
>numbers are unknown with totally different characteristics- standard
>deviations, mean, distributions...). I play around on the Nasdaq a lot, so
I
>test my systems either on the Russell 2000 or NYSE (they both tend to be
>less volatile when compare to the Nasdaq). For bonds, how about trying the
>system designed for US treasuries on the British treasuries (does such a
>thing exist? I don't know that much about bonds).
>
>
|