I believe the coding and data requirements to analyze as Jim suggested is beyond the skill level and available time of most denizens here.

For example, I am trying to build some kind of analysis of the "efficacy"? of my little timing signals dashboard in predicting the likelihood of market behavior over the next 1-6 months. Calling it predictive modeling would be an insult to data science, but that's very simplistically what it is. Simple "=CORREL..." or other statistical regression formulas available are probably not appropriate for the task.

So I've tried to learn and do some Python using Youtube U guides... several hours of work for a lapsed developer... training data set, predictive dataset, etc. etc.... which without a better understanding of statistics and R-squared and multi-variate equation results (ie data science) is difficult to interpret, and may still be inappropriate. For instance, I plugged in my guessed set of 4-5 independent variables. The model spit out a correlation factor of .8. Wow! But... which one or two or 3 of those has the most influence, is it a change of state from bull to bear, is it any 3 or more flipping state, or which are really irrelevant, what's the actual regression formula that may be how to generate some kind of forecast...

At the end of the exercise I know a little more about Python and its awesome capabilities, and modern IDEs like Jupyter (now hosted in the cloud and doesn't even require local installation, for free!)... but in terms of understanding whether my little signals tracking does me any good, I'm still at "So... what?"

I'd love to get some community counsel/training from anyone about this.

FC