Result Discussion: Favorite Ships (part 1)
The main result table · Ship list
CW: This discussion contains mentions of three ships which some people might not be comfortable with (sibling incest and adult/teen relationships). Only their shipnames are mentioned - the nature of these relationships is not discussed in any way.
Before I begin, I would like to address an issue I’ve noticed - I have not been consistent with ship names! You’ll see Jamaya in one chart and Janaya in another, Aruthari in one, Arunari in another. The decision to change Kadia to Kaudia has been recent (due to Kaudia being the ship name actually used by the shippers).
To prevent any major confusion, I updated the ship list to include alternative spellings of shipnames. If you’d like me to include an alternative ship name, please let me know! ^u^
Total Ships refer to the ships people voted for when asked to choose all the ships they ship (click here for the list). I will also use T to refer to these ships.
Favorite Ships refer to the ships people voted for when asked to choose only one ship which they consider their favorite. I will also use F to refer to these ships.
I will use the word rank to refer to the position a ship takes in terms of number of votes (rank #1 is the one with the most votes, while rank #14 is the one with the least votes for Favorite Ships. For Total Ships it’s rank #40).
As in previous discussions, I will refer to Rayllum, Janaya and Ruthari as the Trio.
Favorite Ships vs Total Ships
The first thing I would like to discuss is how favorite ships differ from total ships in terms of popularity. Cross-referencing both can allow us to understand which ships are actively shipped by people and which are just liked “in the background”.
If you read my discussion on total ships, you probably remember that I pointed out the flaw in the first question of the poll (i.e. “What ships do you ship?” is too vague a question). Hopefully comparing the two categories will counter some of that vagueness.
To compare both lists, removed any ships that don’t appear on the Total Ship list and the ones that were not chosen as a favorite ship by anyone. Here’s the list of ships that were removed from this comparison.
One problem, though - there is no way to directly compare the results, since Total Ships have 40 ranks and Favorite Ships have 14.
Trying out different methods
Take the top 10 Favorite ships as an example:
From the table it is clear that Rayllum, Janaya and Ruthari retained the same positions in terms of popularity. Viravos appears to have jumped higher - from #8 to #4. In fact, the only ship in the top 10 that dropped in rank is Corpeli - from #6 to #9. But it’s not a dramatic drop, right?
Aaaaaand this is where it gets problematic. #6 out of 40 is a much higher position than #9 out of 14. The best way to think about is to find the most obvious points of correspondence:
Here, F refers to the rank in Favorite Ships and T - in Total Ships.
Now things become a bit more clear. T6 is closer to the first category and F9 - to the middle. So if I were to simplify the comparison, I’d say
the Trio remain at the beginning of the list;
Viravos, Gremaya and Corpeli jumped from the beginning to the middle;
Virrow, Sorgren and Raydia remain in the middle;
Sordia jumps from the end to the middle.
This method is even more evident at the end of the list:
These are the last nine ships on the list. At first, the green ones seem like a vast improvement in position and the salmon one - a slight worsening. But if we apply our new categories:
Aarrow, Aaruthari, Claravos, Clyx, Grenavos and Rugren remained at the end;
Ezris, Grazi and Kasren fell from the middle to the end;
Sabos had the most dramatic fall - from the beginning to the end!
But then this method is still too broad. Aarrow and Aaruthari both get into the same group - “remain at the end”, even though Aarrow was closer to the middle of the list than Aaruthari. I figured these three categories could be broken down even further:
I divided 40 by 14 and got 2.857, which I rounded up to 3. Hence, three T ranks for 1 F rank. This is obviously not the only way you can arrange this table, as you could alternatively “lock” the first place (1 - 1) and do the others in triplets ( 2 - 2, 3, 4; 3 - 5, 6, 7; etc), which shifts the values quite a bit.
But let’s use this one for now.
I rewrote the Total Ship ranks so that they match the Favorite Ship ones according to the table (e.g. if a ship is ranked #10 in Total Ships, the corresponding rank is going to be #4):
This comparison is very intuitive and more accurate than the three simple categories I used before. We can compare the results for the last 9 ships again:
OLD: Aarrow, Aaruthari, Claravos, Clyx, Grenavos and Rugren remained at the end;
NEW: Aarrow and Grenavos fell by three positions, Claravos - by two, Aaruthari, Clyx and Rugren - by one.
(We can immediately see that Aarrow and Grenavos were the most popular of this group);
OLD: Ezris, Grazi and Kasren fell from the middle to the end;
NEW: Ezris and Grazi fell from #10 to #14, while Kasren fell from #8 to #14 (Kasren clearly suffered a bigger fall);
OLD: Sabos had the most dramatic fall - from the beginning to the end!
NEW: Sabos fell from #3 to #14 (I don’t know if this makes the fall feel more dramatic or not).
This method is still not very accurate:
it conflates several different positions into one (Viravos and Gremaya are now both #3, whereas originally they were ranked #8 and #9 respectively);
there are at least two ways to group the ranks into triplets, neither of which is particularly accurate and both yield significantly different results.
Notice how with this method most ships seem to have fallen in rank, some remained and only one rose! Also, even though Janaya and Ruthari did not change position, this method makes it seem like they fell in rank.
I don’t exactly know why that is, but it’s probably due to the fact that the “value” of a rank feels different depending on the total number of participants. Think of it like this - it sucks to be the third our of three, but it sucks a bit less if it’s out of 40, doesn’t it?
In the next part, I will discuss the method I decided to use in the end.