At the moment, when back-testing using trend predictor, it shows whether or not any given draw's trend was correct. For example back-testing for position 1 you get the results for trend - if the prediction for trend is correct, then a draw doesn't get crossed off, otherwise it does.
But I think that it would be great if it were possible to back-test in a different manner - by predicted value and error. For example if back-testing for position 1, then it should take into account the actual predicted value and then compare it with the actual draw... and then there should be optional error margins, like +/- 5 etc. So if you back-test using trend predictor and it predicts position1; 16, but the actual draw was 18 for position 1, then with error margins +/-5 that draw gets passed. What do you think ?