I'm currently working on predicting bear and bull phases with a dynamic probit model in the form of $y_t=\beta_1X_t+\gamma_1y_{t-1}+\epsilon_t$. So far I've written all my code in matlab and it works as expected when I set the event bear phase as 0 and the event bull phase as 1. Out of curiosity, I've now set the bull phase to 0 and the bear phase to 1, since I figured that, in a binary setting, $P_0$ should be $(1-P_1)$ and that it shouldn't matter which outcome is 1 and which is 0. However, Matlab delivered completly different and somewhat disappointing results.

Now ,my question is: Am I overlooking a theoretical concept that prohibits the use of 1 and 0 and the results are to be expected or is my code just plain wrong?

  • $\begingroup$ Unless we get to see your code, I'm not sure how we are going to be able to answer this question. $\endgroup$ – Raskolnikov Nov 9 '17 at 14:53

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.