@Silverfish If the question is otherwise following the guidelines for routine bookwork style questions (at least those under homework in the [help/on-topic], the tag isn't so crucial (useful, but not critical); on the other hand if the question doesn't follow the guidelines it should close with the specific close reason for those posts whether or not it has the tag.
So if the guidelines aren't being followed the OP needs to edit anyway, and they might as well add the tag themselves when they do. (Adding it to a closed question doesn't seem to have much value until it's being fixed anyway). If the guidelines are being followed, the tag could be added.
But its main benefit then would be for later readers.
@wolfies Thank you very much for your generous offer. I enjoy a lot my interactions with the outstanding members of this community. If you ever need anything radiological please let me know.
I have two data series, which grow over time and one is higher than the other. Basically, these are node and edge nums of a growing graph. The graph is "better" when more nodes are connected. I want to plot how the quality of the graph is improving over the time.
I can just plot the edges/nodes ratio. But, the delta, how much progress was done during a day. So, I can take the diff-direvative of the ratio series or first take the deltas and then take the ratio.
But, the delta-ratio plot won't tell me the weights of the progress. That is, on day one, we added 100 nodes and 400 connections, the ratio was 4. On the next day, the ratio was 100/1. It looks that we did much better but not in fact. The total ratio between the edges and nodes has not changed much because we have added only 1 node and it was easy to get the high edge/node ratio during that day.
I therefore should probably look at how the total edge/node ratio is increased per day. But, that will slow down as the time goes by, since the the total number of nodes is growing and you won't make make much progress after many years if you add the same number of nodes and edges per day. 10/10 vs 20/10 will make a big difference on the first day of the plot but 1000010/1000010 vs. 1000020/1000010 won't show much difference on 10000's day of the plot, despite we are still keeping the pace
the same. So, I must to divide by node (or edge?) count once more. This will give me the units of ratio per node(s) or edges per nodes squared. What would that unit mean? Is it ok?
Let the chat decide if it is stupid question or good enough for the main site. I do not think that anybody else will find this question when they have the same problem.
but there's no reason something on-topic can't be asked in chat first. If it would be better to ask on the main site, wouldn't it be best to just tell the user that instead of falsely flagging messages as spam/offensive?
Well, I am leaving. In the future, please use the flag button only for swearwords, threats, spam, trolls and the like. I'm not a mod or anything, so you can also ask in http://meta.stats.stackexchange.com/ for further clarification
Just for future reference: flagging anything in chat brings in 10k users from all over chat.stackexchange to look at your chat, and I'm pretty sure this is the only room I've ever heard of using the flags as a way to tell people not to post questions in chat. If you don't want users from other chatrooms coming in here and offering opinions about why your flags are excessive, you'd be better off not flagging them in the first place.
@Sycorax how's this? It's fine for them to ask in chat, but if no one in chat wants to answer, then they won't get an answer. The question doesn't need to be deleted; it's still valid in chat. It's just better that in the current room, users would rather have it asked on Main. Chat is recommended as a place to ask a question before asking it on Main to see if it's on-topic. It's also recommended as a place for asking a small question that would be too short for Main. This is my experience
