So Eckhardt is handicapped. No one else has to be, however.
The several alternatives include:
1. Forward testing....
You know words. Cool.
2. Designing and using a paradigm that eliminates probability.
A trading paradigm that incorporates certainty. Now that is WAY cool. I'm unsure of how this meshes with all the blabber about the future not being known, but I likes!
3. Using leading indicators of what is traded....
Once again with the certainty. The very concept of a leading indicator is pretty fantastic. No more of those lagging or concurrent indicators.
4. Automating systems based upon the scientific use of the null hypothesis.
Vintage nonsense. There is no scientific use of the null hypothesis, just as there is no "the" null hypothesis. In statistical studies, the researcher will assume non-correlation. This is called the null hypothesis. He will then set out to investigate whether he can reject a null hypothesis; i.e. he will try to prove correlation. Being a mere convention, it cannot be used as a basis for trading systems anymore than my cat, because it doesn't mean anything. Why not base a system on the continental convention of saying the day before the month rather than the American way of saying the month first?
5. Making use of all of the above in the trading application.
In this case, sadly, minus minus minus minus does not make plus
Back testing has a couple of draw backs:
1. It doesn't work
Mmmkay
2. It doesn't prove anything.
Yawn
3. It has nothing to do with real time in the Present, the only place trading takes place.
Hey, wait, what about the paradigms that eliminated probability and the leading indicators? Come on man! Daddy needs a new yacht.
4. IT is an inductive proof just like anyone quoting a past track record of someone for some "due diligence" type purpose.
I'm not even sure what this means