Are general Statistics questions on topic here? Even if they are tangentially related to RP games?
What about general Dice questions?
I realize both of these questions are tangentially related to RPGs, but are they really on topic?
Having mulled this over for a night or two, I would say that the taxonomy of this question is not useful. Making a "yes, all statistics questions are okay" or "no, all statistics questions are off topic" decision is unlikely to be useful to the site (or even particularly correct).
As a counter to the two examples you brought up, consider these statistics questions:
As mentioned in the comments elsewhere, certain types of statistics (mostly those involving dice) are important to playing and running RPGs. These statistical questions are also essential to the design of new RPG systems (a topic area which we currently support).
Unlike physics, players of RPGs have to deal with actual statistics. It can't be "rule of cool" -ed away... There is a specific target number at which my thief breaks even on pick pocketing, and that number can't be changed by fiat.
I suspect that if the questions cited are to be considered off-topic, it is more likely a reaction to one of these areas:
Should questions about superstitions, traditions, and other meta-elements of the role playing hobby be allowed?
How tolerant do we want to remain of "list of tools" questions? We have a long history of allowing them... But how useful are they?
Do not misunderstand me: I am not suggesting that all stats questions are reasonable. I only believe that the small subset of statistics that deal with the randomizers used in RPGs are on-topic.
As this is the wild-west of meta, I'm going to break this response out into a new answer to give myself access to formatting (and paragraphs). This answer is in response to @Cthos's comment here: Are general statistics or dice questions on topic?
However, I don't think that in practice the line exists at all. Line keeps coming back to "Did you phrase the question in such a way as that it seems relevant?". The 'question which shall not be named' actually came up in a game, but the author did not mention that it did. When he did, the reaction seemed to be "Oh. Whoops...well, we can't reopen it now because its causing bad blood." We also seem to be going the "I'll know it when I see it route", while the criteria remains amorphous. Though I'll note I did +1 your answer on the Off topic question and this one.
The first layer is that you're right... The line doesn't really exist yet. The site was created with the mandate to be "a StackExchange about RPGs." That gives a rough framework (no videogames, no board games, general guidelines for good and bad questions, etc.) but it left a lot undefined. Those undefined areas become defined through discussions in meta, chat, comments, and the overall history of opening and closing questions.
I've given what I feel to be the best place to draw the line. Some folks like it. Some don't. Over time people will be won over to one way or the other. Some people will have to agree to disagree (although stating a clear case for what you want and why you want it is generally better than pointing out the faults in the arguments of others).
Another layer is that no matter how much we debate this, the line will never be 100% reliable. Whether a site is community moderated or governed by administrators, judgement calls around the proximity of the line will always be necessary. We have neither the time nor the know-how to create an iron-clad classification of every possible question (even actual justice systems, developed by lawyers over centuries still often come down to judgement calls). And even if we did, this simply isn't a worthwhile place to spend the effort.
As judgement calls, there will always be cases where slightly off-topic questions that are stated very well are allowed to stay. And there will always be cases where barely on-topic questions are squelched because they were ugly, poorly phrased, or awkward fits for the format. This is unavoidable. The solution is to accept it, and try to improve the questions that are on-topic, and to strive to evaluate questions as they are rather than how they were. This also means that readers need to be willing to make big changes to a question to help it fit, and that askers need to take a light hand in fighting back against this.
For what it's worth, I found the statistics behind Sardathion's post to be acceptable (albeit on the edge), which is why I didn't vote to close it. But I didn't think it was very well suited to the format... It felt too much like a request for a list of software, which is why I didn't vote to reopen it.
I'm not sure which question we're not supposed to name (the goblin with the coat of many pockets?), but I'll conclude by reiterating this: when you ask a question, clearly state why you want to know and give the question some context. Not only does this help clarify decisions and avoid "oops, I guess we shouldn't have mod-closed that" situations, it also allows people to answer the question you meant to ask, rather than being stuck with the question you actually did ask.
Expecting people to suss out an idle wondering unconnected to RPGs from a directly game related question on psychic powers alone is only going to lead to failure.
I would say yes, they are off-topic.
The reasoning for this is the same reason why physics questions are considered off-topic. They are first and foremost about physics with a very tenuous connection to gaming in general.
The dice question was primarily about gamer superstition and the statistics of dice question was a fairly common stat question.
To expand, this question at its core is about gamer superstitions, with superstitions being the focus. As it is written, it is a bad question, and at it's core it is a "just for fun let's talk about gamer superstitions" question without any clear-cut answer to it. While it would be fun to discuss, there is no clear answer here and is more of a discussion point. Because of that fact, according to the SE faq, it is considered off-topic. There is no discernible merit to it for the RPG.SE website.
On there other hand, there is this question, which is admittedly a better question. It has sense been reworded to help direct it's focus, but at it's core it is still a statistics question. It makes reference to the Rolemaster system, but that's where the tenuous connection to gaming ends. It is a simple statistics or mathematics question that does have a direct answer, but isn't a gaming question. Now, if the stats/math question was directly related to a system, such as figuring out the average to-hit for system r or the average damage from dice rolls with modifier included for system x, it would have a relevant link to a specific system that would not make sense on a straight statistics or mathematics site.
(The above mentioned question has sense been edited and reworded by SevenSidedDie and is no longer a statistics question and is instead looking for a system/spreadsheet/program to figure out statistics. I still think the example is relevant, so I am leaving it in my response.)
Now, is there merit to these questions? Absolutely! But they are more suited for a forum discussion where answers are broad and more suited to an open-ended discussion. Conversely, RPG.SE by nature is a beast of specificity with concrete, set answers.
I don't want to demean the questions in question, as I'd love to discuss them in a forum-type setting or around my gaming table, but as such, there is a precedence with other types of questions that these are off-topic.
NB: Not trying to bring up any bad blood or old wounds, just using the two linked controversial meta questions to establish precedence for my reasoning.
As the original poster I was quiet surprised that the question had been closed so longer after it had been answered... And both answers where good and useful. From the above I see this argument breaking into two:
This is fine but to me, it sounds like squabbling over semantics. Would the following questions be on or off topic?
"I am designing a RPG system and want to know if someone had done some statistics on a wide range of dice rolls?"
"I does anyone know if someone had done some statistics on a wide range of dice rolls?"
In my opinion, they are the same question, phrased differently. However, the first one would be on topic, the second not.
Am I right?