Roxygen vs Roxycal Quarantine Thread

You didn't think to have a "Neither" option? At this moment, that would be my vote on the matter. And that's both in terms of what I think is actually happening with the characters and what I want of the characters.

neither

It's weird, I've barely seen anyone here or on the subreddit actually ship Roxy and Calliope and like what happened in the credits and all that. It makes you think about whether it's mostly appreciated by that side of the fandom...

At first I was like "Do we really need 3+ threads for this? -_- " but then I realized that this means it's less acceptable to complain about RoxyCal elsewhere Anyone who complains about RoxyCal or Roxygen elsewhere is hereby PERMABANNED FOR LIFE. (not really but I am tired of hearing about it)

For the record I've seen plenty of people who like both ships on tumblr or reddit or the forums, otherwise logically these debates wouldn't even be happening, it'd just be a bunch of people sitting around agreeing emphatically with each other. The percentages may skew one way or another, sure, but it's not THAT harshly divided. There's been plenty of people who have posted RoxyCal fanart to the subreddit, for instance (otherwise cookiefonster wouldn't be able to complain in the comments ) and there are plenty of posts on tumblr in support of Roxygen. Such is the nature of shipping.

I personally view Calliope/Roxy's relationship as more like very close, sweet, physically affectionate besties, but I'll be fine with it either way really. There have been many Homestuck updates that felt "off" or made plot changes that I thought were questionable at the time, but I've eventually been happy with MOST plot developments in Homestuck when I eventually see what it's building to, so ultimately I just wanna see where all of this is going. Either ship is honestly perfectly plausible or at least explainable, and tbh it seems silly to be like "THESE TWO PEOPLE WHO GET ALONG AND CARE DEEPLY ABOUT EACH OTHER MAKE NO SENSE TOGETHER!" when so much other media contains depictions of relationships where the two people involved have nothing in common with each other or even outright HATE each other (see: the tropes where married couples constantly complain about being married, have few shared interests, etc).

My vote goes to Roxycal. Dunno, I alwayss thought they had a lot of chemistry together. The fact that they make a really cute couple also helps.

I've internally shipped them for longer than I'd like to admit (i.e. way longer than just since 10/25) and I'm not going to pretend I'm not still salty as fuck about that.

What about if you ship Roxy with Calliope, John, Jane, Jade, and Terezi?

Woah, hey Archie! Wasn't expecting to see you here.    


PS: ROXYGEN IS WINNING YESSS

I'm... not sure I understand?

I don't think you know what being devoted is, dude. There are idiots who try to perma-dye their skin gray just to look like troolls. I don't even draw Roxygen, or do anything that insane, I just ship it. It's literally the only ship I actually ship. Not in Homestuck, but EVERYWHERE. So I dunno, I'd say my devotion is average since this is the only ship I ship.

In all honesty I kind of have to agree.

When you are designing a poll, you need to consider a lot of factors. There are several things which can cause a survey, poll, or other data-gathering exercise to be too biased--to be data that you cannot trust. Some of these factors are:

• Voluntary sample bias: A sample where people only respond when they choose to respond is biased, because you're having a non-probabilistic element come into play. Only people who want to share their opinions/status/interests will do so. This confounding factor means that your sample won't accurately reflect the group you're trying to study. As an example, magazines sometimes conduct polls where they ask people to mail in about it. A lot of people will simply not choose to mail in responses, and they might choose to avoid responding because of social factors, or because they don't like sharing personal info, or other reasons. Yet these people can be a substantial fraction of the total set you're considering, meaning you cannot draw useful conclusions from the data you collect. Volunteer samples are almost always off, and usually by a lot. Consider, for example, a volunteer sample poll about whether people have ever had homosexual thoughts or desires. Your sample will almost certainly not reflect the actual population, because there are social factors involved, e.g. heteronormative feelings, a desire for privacy, respondents fearing this info could be used against them, etc.
  
• Unrepresentative sample bias: When you're trying to make a really good, representative poll, it's actually a major challenge to make sure that your poll accurately reflects the real population. This is not a simple matter! In real polling, you need to be careful both that you don't inject your own biases into the sample selection, and that you don't accidentally allow randomness to do the same. For example, if you randomly sample from the population of likely voters, but just happen to get 60% men/40% women, do you weight the responses differently, since women (very slightly) outnumber men in most studies of actual US voting demographics? Or do you leave it as it is, because you don't want to be introducing your own sample bias? What if you only end up with a single person representing a particular demographic, such as with the LA Times polling, where they had just one black man in the (IIRC) 18-26 age group, who was unusually pro-Trump for people of his demographic group? These concerns are significant, and they can turn even very well-conducted polling into garbage if mishandled. The thing is, a forum is inherently unrepresentative: you're not actually sampling "the Homestuck fandom." You're not even sampling "the membership of Omegaupdate." You're only sampling from the population of "people who regularly check Omegaupdate and decided to read this thread and decided to respond." That's...well, I mean, it basically limits your population to "just those people who actually voted," which makes the poll pretty much useless.
  
• Revision bias: Whenever you create a poll, it is absolutely vital that respondents only be able to submit their answers once to that specific poll. If people can change their answers, those answers may be subject to social pressures, a desire to belong, or the like. Further complicating things, people may actually look at the results before voting, thus distorting their choices even if they don't realize that they're voting different than they would if they hadn't looked! Note, though, that I said "once in that specific poll." Sometimes, pollsters construct "panels," which are specific, (hopefully) unbiased-selected respondents who will be continually polled over a long period of time (the above LA Times poll used a panel that was just randomly a little weird, so their numbers were a little weird). With "panel"-based studies, it's perfectly okay for people to change their answers in Poll #2 compared to Poll #1--IF, and ONLY if, people actually want to change then you DO want to see it happen between polls, because it gives you valuable data about how opinions change over time. You still don't want people to be able to call back in and say "actually although I said Blue with the last poll I really meant Yellow."
  
• Small sample size: We have 17 votes. This allows far too much variance to draw any meaningful statistical conclusions. Generally speaking, you need at least 30 data points before you can do proper statistical analysis of data. With just 17 votes, things are far, far too swingy. Simply by changing my vote from Roxygen to Roxycal, I can swing Roxycal from a clear and fairly distant 3rd place (less than half the votes of Roxygen, barely more than a third: 8 Roxygen, 3 Roxycal) to being tied for second place and more than half the votes of Roxygen (7 Roxygen, 4 Roxycal). Such sensitivity to the voting of one single person is a major source of statistical bias.
  
• Discussion bias: Believe it or not, the simple fact that we're talking about this is going to have a negative effect on the poll. This is not to say that discussion is bad--far from it! But another source of bias, here, is that people might read the thread before voting. What if I give an especially eloquent or striking defense of some particular option (frex, Roxycal, since that's against your preferences), even if it relies on faulty information or inaccurate interpretations? What if the discussion is dominated mostly by fans of one particular option, and thus people who read before they vote get the sense that the consensus is that option even though it really isn't? These are further sources of bias that warp the results, and it compounds the effect of the previous two biases (small sample size means a single voice can more easily sway most of the sample; revision means people can vote how they actually feel, hear an argument that makes them reconsider, and then change their votes).
  

Honestly, there are probably more types of bias I could come up with--these are just the ones that stick out. Collecting actually GOOD data, data that actually TELLS you things instead of obfuscating things, is really really hard. There's a very good reason why statistics is a career all its own, and not simply a thing you can take a couple of courses on and append to your resume.