OOTS #1151 - The Discussion Thread

[zimmerwald1915]

Since when does the story (or The Giant) particularly care about what is likely?

You have to know what the trend is to intentionally buck it.

[Caerulea]

You have to know what the trend is to intentionally buck it.

I am sure he knows. He just doesn't ​care.

[Nion]

Since when does the story (or The Giant) particularly care about what is likely?

I agree that Kish's statement was ridiculous, but I do think that approach is not a great way to do things.

Now this is just me guessing at things, I could be completely wrong... But my guess is that when The Giant invented the character Durkon, he didn't even think about whether Durkon would be heterosexual or not. My guess is that he just made him heterosexual without considering bisexuality at all. And the reason I guess that is because of all the statistics. In the first place, bisexuality is uncommon in real life, and secondly, it was only very recently in real world history that culture has done so much to point out the existence of humans with these uncommon sexual desires, that even elementary schools teach children about the issue.

[Kish]

It seems to me like you're just trying to undermine my logic, by taking it further than a reasonable interpretation of it, would take it.

You're close. I'm trying to show you how bankrupt your logic is, by illustrating an example where I correctly predicted you would not say "yes, that's what statistics dictate so that's what I'm saying." It is unfortunate that you chose to make an excuse about statistical error, instead of recognizing that "Vaarsuvius is female because statistics" and "Durkon is straight because statistics" are both misuses of statistics.

[Caerulea]

... In the first place, bisexuality is uncommon in real life...

Do note that our world is not the OotS world. Shocking, right?

[Ruck]

Perhaps I'm reading The Giant's post differently than you are, but he seems to be saying that statistics should not be used to tell an author how to diversify his story. I agree with The Giant in this. What we are trying to do here is determine an aspect of one of the Giant's characters. I am not saying he has to follow statistical data in his characters. Rather, I am saying that such data is our best indicator of the truth when the truth is not apparent.

The point-- as Caerulea has made-- is that, given that this is a story written by someone, with humanoids and other fantastic elements, you can't just extrapolate distribution among humans on Earth as measured (which may not be accurate in the case of things like orientation), to make assumptions.

It's okay for personal details to be inconclusive.

I mean, while I am convinced V was intended to be male I prefer to imagine female as to me it brings greater balance to the sex of characters in the comic, not to mention the subcategories of amusingly prickly characters, power hungry characters, bad parent characters, etc; and adds more variation to the personality of the female characters.

But isn’t it not going to be revealed?

Vaarsuvius is genderqueer according to Word of Giant.

[Nion]

You're close. I'm trying to show you how bankrupt your logic is, by illustrating an example where I correctly predicted you would not say "yes, that's what statistics dictate so that's what I'm saying." It is unfortunate that you chose to make an excuse about statistical error, instead of recognizing that "Vaarsuvius is female because statistics" and "Durkon is straight because statistics" are both misuses of statistics.

You are saying my claim is that Durkon is heterosexual. I am not saying that. I am saying that Durkon is much more likely heterosexual.

If you correctly apply logic in the same way to Vaarsuvius, you would say "Well, we don't know what Vaarsuvius's gender is, but statistics for humans have them at 51% female and 49% male, so my guess is female."

A very important part of using statistics to make guesses about unknowns, is considering the magnitude of those statistics.

[Ironsmith]

You're close. I'm trying to show you how bankrupt your logic is, by illustrating an example where I correctly predicted you would not say "yes, that's what statistics dictate so that's what I'm saying." It is unfortunate that you chose to make an excuse about statistical error, instead of recognizing that "Vaarsuvius is female because statistics" and "Durkon is straight because statistics" are both misuses of statistics.

Which is to say, you really weren't paying attention. The point is, "Durkon is more likely to be straight because of statistics", not "Durkon absolutely has to be straight because of statistics". And again, this is ignoring the huge difference in numeric weight between the two... 49:51 can be hopped, no problem, 4:1 requires a little evidence.

[Caerulea]

A very important part of using statistics to make guesses about unknowns, is considering the magnitude of those statistics.

And the dataset you have collected. If I use this data to argue that, because there are more cats than dogs, and because dwarves don't really exist, that Durkon and the rest of the population are most likely cats in disguise, that is nonsense. While trying to apply real world human data is not quite as absurd, it is still not a great fit. Again, humans != dwarves.

(I recognise that cats in disguise is a bigger leap than heterosexuality. The point is that the data doesn't apply, not what it would imply.)

[Liquor Box]

I do think the crux of this argument is whether you can assume dwarven demographics are similar to real world human demographics, in those cases where the story itself does not establish otherwise. If it is ok to make such assumptions I don't think whatever argument Kish is trying to make follows.

Do we assume that roughly 50% of dwarves are female and roughly 50% are male?

Personally, I tend to lean toward it being ok to make such an extrapolation for most purposes where we have no contradictory information, but I can understand why some people like Caerulia think that we cannot.

[Nion]

Do note that our world is not the OotS world. Shocking, right?

My point in the specific sentence you quoted was not to say that "Durkon is likely heterosexual because most people are", but rather to say "The Giant, living in this world, way back in 2003, was unlikely to consider bisexuality for his dwarven cleric stick figure, because such an uncommon thing wasn't at the forefront of people's minds back then."

[Hemoparty]

Regardless of the difference between "Person is X" and "Person is more likely to be X", it feels to me like all of this is stemming from willfully ignoring how little real-life statistics (inaccuracy notwithstanding) apply here. They simply don't. The Giant says they don't.

[Caerulea]

My point in the specific sentence you quoted was not to say that "Durkon is likely heterosexual because most people are", but rather to say "The Giant, living in this world, way back in 2003, was unlikely to consider bisexuality for his dwarven cleric stick figure, because such an uncommon thing wasn't at the forefront of people's minds back then."

It is possible, probably even likely. I don't know how much that matters now though. However, quoting you in the way that I did was dishonest of me, and for that I apologise.

[Nion]

It is possible, probably even likely. I don't know how much that matters now though. However, quoting you in the way that I did was dishonest of me, and for that I apologise.

I forgive you. No hard feelings.

[Jasdoif]

It's the only statistical data that we have. So, if we use this method to try to determine if a certain trait is slightly more likely, we have to use this.

You have no relevant statistical data. Might I suggest not trying to determine how likely or unlikely a trait is with the nothing?

[Caerulea]

It's the only statistical data that we have. So, if we use this method to try to determine if a certain trait is slightly more likely, we have to use this. Dwarves don't seem all that different from humans. Yes, this approach surely makes the data less useful, but it is better than nothing in this approach.

We have quite a lot of statistical data. I for one have access to data on how much of the recycling in my school is actually recycled. Or maybe we should use the median income of my town to determine how much Durkon earns. Or maybe look at the most frequently spoken language in Asia to determine what language is spoken in the comic. We have a tremendous amount of data. The point is that none of the data applies. (Well, except for that which is collected from the comic strip. That applies).

[Liquor Box]

Regardless of the difference between "Person is X" and "Person is more likely to be X", it feels to me like all of this is stemming from willfully ignoring how little real-life statistics (inaccuracy notwithstanding) apply here. They simply don't. The Giant says they don't.

Does the Giant say that? He certainly doesn't say it in the post Jasdoif quoted earlier in the thread.

In that post he says that he (as an author) should not be constrained to tell a story about people who are a statistical representation of rel world (or campaign world) demographics). It's ok for him to tell a story about all gays, or all men, or all eleves, even if that is not statistically probable if each character's demographics were randomly generated. That is because the characters from hsi story are not randomly selected from the general population, but are characters he has chosen because he wants to write about them.

That is different from saying that statistics for things (eg proportion of males and females being roughly similar) in the OotS world do not match the real world (except where he has indicated one way or the other). He has been silent on the topic.

[Liquor Box]

We have quite a lot of statistical data. I for one have access to data on how much of the recycling in my school is actually recycled. Or maybe we should use the median income of my town to determine how much Durkon earns. Or maybe look at the most frequently spoken language in Asia to determine what language is spoken in the comic. We have a tremendous amount of data. The point is that none of the data applies. (Well, except for that which is collected from the comic strip. That applies).

I think using the median income of a balcksmith or grocer in medieval times might be relevant to determining how much a blacksmith or grocer earns in the Dwarven kingdon - at least relative to the general population.

We can clearly eliminate some data as irrelevant - like recycling, but that doesn't mean that none of it is relevant.

I think if you polled people as to who earns more in the OotS world, a farm labourer or the lawyers we see from time to time, most people would say the lawyers. And I think most people would say that because they are (consciously or not) assuming their knowledge of how much people earn in the real world applies to some extent to the OotS world. And I also think that line of reasoning would be valid.

[Pablo360]

Statistics mean nothing to the individual.

Durkon is a single character. Using statistics to argue about what his sexuality is “likely” to be amounts to tedious navel-gazing that, at best, reproduces a status quo where minorities are assumed to be less relevant, and at worst excludes them entirely. Because it's easy to forget how big even small probabilities are.

Basically, talking about an unknown aspect of a character in terms of probability is misunderstanding how numbers work. “Probably not” is infinitely closer to “maybe”*than it is to “no”. A small probability, assuming it's not really small (a heuristic that changes depending on context), is actually pretty close to 1 in the grand scheme of things, but any probability is infinitely bigger than zero.

What do I mean by this? Let's consider the assignment of unknown, but statistically studied, character traits to be akin to mathematical estimation. In estimation, when you add together two numbers, if one number is way bigger than another (according to the subjective interpretation of the estimator), the smaller number is ignored; similarly, for multiplication purposes, if a number is really close to one, you can consider it equal to one. However, you cannot do the same thing with zero, and here's why.

Let's say you have the expression (999/1000)/(1/1000) and you want to estimate its value. The numerator is the same thing as (1000-1)/1000, which can be estimated as 1000/1000=1. The denominator, on the other hand, cannot be reduced to 0 in the same way; if you tried to do the same thing as the numerator, you'd get (1000-999)/1000, and 999 isn't small enough compared to 1000 to reduce. Thus the estimated value of this expression comes out to 1000, which is pretty close to its actual value of 999. Estimating 1/1000 as zero without having a MUCH bigger number to contrast it against, however, nets an estimated final result of ∞.

Basically, unless the odds of Durkon's being bisexual are super low, them being regular low (I've seen people argue 20%) don't really mean anything.

All of which is moot if you take this strip in isolation as evidence for Durkon's being bisexual, in which case the subjective odds become 100%, seeing as none of the arguments for Durkon's being straight actually convincingly argue against his being bisexual, especially if Elan is #5.

[Liquor Box]

Durkon is a single character. Using statistics to argue about what his sexuality is “likely” to be amounts to tedious navel-gazing that, at best, reproduces a status quo where minorities are assumed to be less relevant, and at worst excludes them entirely. Because it's easy to forget how big even small probabilities are.

Basically, talking about an unknown aspect of a character in terms of probability is misunderstanding how numbers work. “Probably not” is infinitely closer to “maybe”*than it is to “no”. A small probability, assuming it's not really small (a heuristic that changes depending on context), is actually pretty close to 1 in the grand scheme of things, but any probability is infinitely bigger than zero.

What do I mean by this? Let's consider the assignment of unknown, but statistically studied, character traits to be akin to mathematical estimation. In estimation, when you add together two numbers, if one number is way bigger than another (according to the subjective interpretation of the estimator), the smaller number is ignored; similarly, for multiplication purposes, if a number is really close to one, you can consider it equal to one. However, you cannot do the same thing with zero, and here's why.

Let's say you have the expression (999/1000)/(1/1000) and you want to estimate its value. The numerator is the same thing as (1000-1)/1000, which can be estimated as 1000/1000=1. The denominator, on the other hand, cannot be reduced to 0 in the same way; if you tried to do the same thing as the numerator, you'd get (1000-999)/1000, and 999 isn't small enough compared to 1000 to reduce. Thus the estimated value of this expression comes out to 1000, which is pretty close to its actual value of 999. Estimating 1/1000 as zero without having a MUCH bigger number to contrast it against, however, nets an estimated final result of ∞.

Basically, unless the odds of Durkon's being bisexual are super low, them being regular low (I've seen people argue 20%) don't really mean anything.

All of which is moot if you take this strip in isolation as evidence for Durkon's being bisexual, in which case the subjective odds become 100%, seeing as none of the arguments for Durkon's being straight actually convincingly argue against his being bisexual, especially if Elan is #5.

That's not true. Take my example from the previous page:

For example, if there were a new poster on this forum, who listed their location as "USA", in the absence of any further information, it would be more likely that they were from New York, texas or California, than that they were from Alaska or Hawaii. If you were guessing the poster's specific location, statistics (percentage of US population that live in different states/cities) would probably give you the highest chance of guessing succesfully. Of course, as soon as you got a clear indication of their location (such as them telling you), then that would override the guess on the numbers.

Using statistical probability is relevant to making educated guesses about a character features, without a doubt. The only real question is whether (and to what extent) real world data applies to the OotS world.

[Kish]

There are cases in which the only way to avoid being ridiculous is to say "I don't know."

This is one of them.

A regrettable number of people seem to regard any answer at all as better than "I don't know."

[Caerulea]

There are cases in which the only way to avoid being ridiculous is to say "I don't know."

This is one of them.

A regrettable number of people seem to regard any answer at all as better than "I don't know."

Rather than "I don't know" I would call it "I have no idea." Many will happily admit they don't know, but argue that one or another is likely.

[Hemoparty]

Does the Giant say that? He certainly doesn't say it in the post Jasdoif quoted earlier in the thread.

I'm not making some sweeping claim that population statistics unilaterally do not match up to the real world Just Because, only that we can't (necessarily) apply them to these characters the author has chosen to write about because the author is not constrained by these statistics.

[Grey_Wolf_c]

Making a mirrored statement for rhetorical effect, does not allow for that statement to be false.

If Gray Wolf wanted to exclusively make the case that we don't know the orientation of Durkon, then Gray Wolf's statement,

"Before saying Durkon is heterosexual, please find evidence for it. Until such evidence is found, you would be more likely correct to assume that he is not"

should not have been made, and Gray Wolf can retract it if desired.

1st, my name is not "Gray Wolf". Given that I sign every post, I have to assume you are being deliberately offensive. After all, "statistically", that is most likely the case.

2nd, my statement is perfectly correct: you cannot assume a non-null hypothesis without evidence. I don't give a damn what you think is "most likely" for a fictional race of beings. For all you know, dwarves are all bisexual except a tiny irrelevant percentage that fall into the other possible orientations. Your entire thesis and therefore conclusion is based on bull**** statistics and unsuported and unsupportable assumptions. Absent such evidence - evidence which you must definitely don't have - the only valid approach is to assume equiprobability, thus "statistically", the chance that Durkon is heterosexual is at best 50/50, and if we throw in, say, pansexual, a mere 33%. Or, even better, you accept we don't know and stop trying to reach a conclusion from no valid support.

Grey Wolf

[Ironsmith]

Durkon is a single character. Using statistics to argue about what his sexuality is “likely” to be amounts to tedious navel-gazing that, at best, reproduces a status quo where minorities are assumed to be less relevant,

Numerically, that's cause they are. That's the literal definition of a minority.

and at worst excludes them entirely. Because it's easy to forget how big even small probabilities are.

Again, the argument isn't that it's impossible; it's just that it's not as likely, and that to argue that he absolutely has to have this trait, you need evidence; otherwise, the best assumption is the most probable.

Basically, talking about an unknown aspect of a character in terms of probability is misunderstanding how numbers work. “Probably not” is infinitely closer to “maybe”*than it is to “no”. A small probability, assuming it's not really small (a heuristic that changes depending on context), is actually pretty close to 1 in the grand scheme of things, but any probability is infinitely bigger than zero.

That depends on how often the probability is applied; here, it's applied once, and only once. Yeah, I subscribe to the whole "an event that's not impossible, given an infinite number of chances to occur, is inevitable" school of thought, too, but "one" is a considerably smaller number than "infinity".

What do I mean by this? Let's consider the assignment of unknown, but statistically studied, character traits to be akin to mathematical estimation. In estimation, when you add together two numbers, if one number is way bigger than another (according to the subjective interpretation of the estimator), the smaller number is ignored; similarly, for multiplication purposes, if a number is really close to one, you can consider it equal to one. However, you cannot do the same thing with zero, and here's why.

Let's say you have the expression (999/1000)/(1/1000) and you want to estimate its value. The numerator is the same thing as (1000-1)/1000, which can be estimated as 1000/1000=1. The denominator, on the other hand, cannot be reduced to 0 in the same way; if you tried to do the same thing as the numerator, you'd get (1000-999)/1000, and 999 isn't small enough compared to 1000 to reduce. Thus the estimated value of this expression comes out to 1000, which is pretty close to its actual value of 999. Estimating 1/1000 as zero without having a MUCH bigger number to contrast it against, however, nets an estimated final result of ∞.

Well that's just shoddy math. You could forgo an estimate altogether just by simplifying the expression (multiply it by .001/.001, which equals 1 and will not change the fraction's value, and eliminate the complicating factors). I know you're trying to make a point here, but come on, man, find yourself some better examples than that.

Basically, unless the odds of Durkon's being bisexual are super low, them being regular low (I've seen people argue 20%) don't really mean anything.

All of which is moot if you take this strip in isolation as evidence for Durkon's being bisexual, in which case the subjective odds become 100%, seeing as none of the arguments for Durkon's being straight actually convincingly argue against his being bisexual, especially if Elan is #5.

But the comic does not consist of just this strip. He's been repulsed by the notion of witnessing/interacting with nude males before, and ignoring that in favor of a cheeky offhand remark doesn't do your credibility any favors.

[Liquor Box]

There are cases in which the only way to avoid being ridiculous is to say "I don't know."

This is one of them.

A regrettable number of people seem to regard any answer at all as better than "I don't know."

I think ever single person has explicitly or implicitly acknowledged that they do not know.

But some have gone on to say "but I think one possibility is more probably than the other on the basis of....". If you think their reasons are no valid, feel free to say so, but I don't think it's fair to imply that people are spouting rubbish because they are unwilling to admit they do not know.

[Larre Gannd]

The assumption that the data is applicable is an absurd one in these instances. Do Oots recycling habits seem to mimic those of the real world? Not at all. So any sort of reasoning based on this assumption is equally absurd.

I am just saying that we can use real-world statistics to approximate statistics for the oots world (so long as the thing measured by the statistic can reasonably be assumed to be similar in both worlds.

The thing is, we pretty much can tell that OOTS world is far more accepting of different races and sexual preferences than ours (e.g Bandana casually mentioning girlfriend, Roy being referred to as dark skinned, and never as black, jail guard in Cliffport, the loads of female warriors in W&XP, etc), so it is likely that many other things also are different, regarding that area. As such, since we have no information other than what the comic has provided, it is difficult to extrapolate. That being said, who am I to dictate how much info is needed for folks to be able to take away from stuff?

[dmc91356]

Why all this statistical "analysis" on Durkon? We have in-comic evidence that Durkon finds females attractive (hence, Kudzu, y'know) and we have in-comic evidence that Durkon's mom would like him to settle down with someone of either gender, which would lend to the argument that she believes that he would be fine with either gender, i.e. bi. Now, the latter motherly comment may be a retcon of earlier strips that have been cited in support of Durkon being hetero, but that's irrelevant here. What we know from the actual comic would easily support the idea that Durkon is bi, and how the real world human population shakes out statistically is therefore particularly meaningless here.

Given Rich's comments in the past about more inclusiveness, I would absolutely find it in character (sorry, bad pun) that Durkon is bi. You can use statistics to your heart's content, but if you know something about the horses in the race, you are going to have a much better chance picking the winner than if you just go by their records on paper.

[Jasdoif]

Yes, we don't have statistical data on dwarves, but were we to try to draw a conclusion about the likelihood of certain traits in a dwarven population, we can make certain assumptions to try to get a better answer to our query than "there is no way we can know." To do that, we assume that because dwarves resemble humans in many ways, statistics for humans will be somewhat if not completely applicable for dwarves. This decreases the reliability of our results, but we can still at least attempt to get a fuzzy picture rather than no picture at all.

Except, because you have no data on dwarves, your fuzzy picture is of the humans you're assuming. Without data backing a correlation of traits between humans and dwarves, or a correlation from those traits to the one you're looking for, the assumption does not reasonably have anything to do with dwarves; it's not drawing a conclusion so much as jumping to one. "If we assume dwarves are like humans in this one way, then dwarves are like humans in that one way" is not particularly useful.

[Aquillion]

And I think that the statistical distributions of traits in humans on earth have little or no bearing on the statistical distributions of traits in dwarves, in whatever the world of The Order of the Stick is called. Their sexual orientation isn't defined.

FWIW the statistical distribution of sexual orientation of real humans on earth isn't so clearly defined, too (since we rely on self-identification, which is not an immutable category but can respond to social pressures to stay in the closet or whatever.) Case in point, a poll in Britain had only half of younger Britons describing themselves as a 0 on the Kinsey scale. Obviously there's a lot of ways to interpret that, but the point is, Durkon could be like a 1 or a 2 on the Kinsey scale, easily, and even if we applied our demographics to OOTS dwarves, it wouldn't be particularly unlikely.

(I mean your point is also correct. I just felt it as worth mentioning that.)

That was also partially why I speculated about the cultural aspect many many pages ago - in our world, someone who looked as old as Durkon would be much less likely to identify as bisexual than someone younger, but OOTS isn't our world.