# different flavors of probability question: better here or on CrossValidated (Stats SE)?

I'm guessing all of these topics are on topic both here and on CrossValidated (stats.stackexchange.com). But I'm curious which location is better for each type of question. What do you think?

• Basic probability question, the type that might be asked in a Stats 100 class.
• Basic combinatorics question

Actually, I'm still pretty new to probability and statistics, so I'm not sure if I'm choosing the right types of questions for my list. Maybe there's a better dividing line.

In any case, thoughts?

• My opinion, knowing nothing about what goes on at Stats.SE, is that probability and combinatorics should go here and statistics should go there. Jul 7, 2017 at 14:15
• As I said I'm still pretty new (actually, very new) to these fields, so allow me to ask: How do you distinguish between the two? It seems like a lot of statistics is based on probability techniques. Jul 7, 2017 at 14:48
• Jul 7, 2017 at 15:10
• I should clarify I spent a while googling that exact question recently and read a number of answers like "they are inverses: probability looks at what might happen, and statistics looks at has happened." I might be butchering that. But that's somewhat philosophical and my question is more like: How would you draw a line in the sand between specific techniques? Like, could you look at an equation and say "this is a statistics equation" or "this is a probability equation" Jul 7, 2017 at 15:13
• We posted at the same instant. I will take a look at those, thanks. Jul 7, 2017 at 15:13
• we do get machine learning questions here, and there is a tag, and some of those questions do get answered here. However, in an early discussion, Stats decided they would take machine learning questions, and they get quite a bunch, and do answer them. I usually suggest people put machine learning questions there. stats.stackexchange.com/questions/tagged/machine-learning Note that these students may not be taking any mathematics or statistics other than said "machine learning," so the questions may be fishing expeditions. Jul 7, 2017 at 18:24
• from stackoverflow.com/tags/machine-learning/info Machine learning is a subfield of computer science that evolved from the study of pattern recognition and computational learning theory in artificial intelligence. Machine learning explores the construction and study of algorithms that can learn from and make predictions about data. Such ... NOTE: If you want to use this tag for a question not directly concerning implementation, then consider posting on Computer Science, Cross Validated or Data Science instead. Otherwise you're probably off-topic. Jul 7, 2017 at 18:48
• Just my opinion as a stats student in grad school who has frequented these two websites... I prefer Cross Validated more for applied statistics questions (e.g., I would not ask this, this, or this on Math SE) ... [continued] Jul 8, 2017 at 1:02
• and Math S.E. for more detailed mathematical derivations (e.g., here's a question that I put a bounty on in Stats SE but then asked Math SE to provide some more detail). Probability and statistical theory questions are usually asked in both of them. Jul 8, 2017 at 1:04
• This is to concur with @Clarinetist's pair of thoughtful comments. (On the contrary, I would be rather in trouble to allocate a truth value to the "inverses" quote given in a comment further above. One finds now and then similar seemingly definitive, "deep", statements, but what do they even mean in practice?)
– Did
Jul 8, 2017 at 9:44
• I like how @Clarinetist phrases it, and that was my gist prior to asking this question: CrossValidated is best for applications, and Math SE is best for the core math. That will remain my opinion unless I see a better one here. As for ML, that's a bit beside the point of my question but probably it's similar: CS SE for core CS, Stats SE for applications. And Math SE for the core math if you ever go that far back. Jul 8, 2017 at 18:24
• @Did I agree with your criticism of the inverses quote. Jul 8, 2017 at 18:25
• In my experience, anything remotely theoretical is better for math.se. If you aren't culling a dataset or writing a script then 7 times out of 10 it will be unnoticed there. For instance, this question was left unanswered for two years on the statistics stack exchange Jul 12, 2017 at 16:26