Uncreative Anonymity

mimski | 20 | austrian

{ wear }
About me Tags Face About Me Tag MY RUBY NUZLOCKE RUN

Troll blood colour test - supplement

So yesterday I’ve made a troll blood colour test. Maybe I shouldn’t do such stuff sick and tired, because it has several flaws:

  • Because of how the average works the bloog colours in the middle of the list are more common than the ones on the “edges”. This is not true for the hemospectrum though, since the bottom colours (except for candy red) should be more common.
    It’s also very impractical since Kanaya’s blood colour, Jade, is supposed to be rare, while in my test it’s rather easy to get.

    One possibilty to avoid this would be to use another kind of average. Both the geometric as well as the harmoic mean are lower than the arithmetic mean used in the test. This would make the test more complicated for people to do though.
    (Example, my numerical values were 10, 1,  5 and 9. The arithmetic mean was 6,24 making me Jade blooded. Using the geometric or harmonic mean I would get following results:
    GM = 4,6 ~ 5 making me Olive blooded
    HM = 2,8 ~ 3 making meYellow blooded)
    Changing the used mean would make a difference and lower the results significantly.
  •  I have not checked how the 12 and 13 zodiac overlap.
    It is NOT possible to get all 12*12 combinations from both zodiacs, the sign fluctuates not more than 2 signs back. It is also NOT possible to get a “higher” sign than the one you’ve got. In theory this is good, because it lowers the average as I intended. But there is ONE case where it is bad:
    Aries. Since Aries is the “lowest” zodiac, going back here makes you jump to the highest zodiac, Pisces. This again increases the average quite a lot, thus making the middle bloods more common and lowbloods less common. Not good.

  • I included Fuchsia blood, even though there usually are only two Fuchsia blooded trolls at the same time, according to the comic. (The emperor/empress and the possible successor.)
    The most logical conclusion would be to remove Fuchsia blood, but I think that wouldn’t be fun. The other possibilty would be to make it as rare as candy red blood is in my test. I’m not sure how to do this though!

I thought of remaking the test on a more calculated basis, that would mean a huge load of work for me, so it’s nothing I would do right now.
I could either change the average (as mentioned before)  or/and weight he numerical values, for example making the normal zodiac sign more important, and the moon sign less important by multiplying them with factors larger/lower than 1.
That would make the test more complicated to do, once again though!

The other possibilty I’ve thought would be the most work for me, but probably the easiest to do four you. It would work simlarly to the first idea:
The zodiac sign would be the most important sign, again with values from 1-12.
Instead of calculating an average though, the other signs would have values you have to add. This values would have to be different for each sign. This way I could directly control the frequency of the blood colours!
Here’s an example: Someone who’s Aries would get the numerical value of one. Most their values for their moon, rising and 13 zodiac sign would 0, so they stay a rust colour. But one Aries-value (maybe 13 zodiac) would be -1, dropping them to 0, giving them candy red blood. I would tweak a bit to make candy red blood even rarer though!
Pisces as an additional sign would have the value of +1 though, giving SOME Aries the chance to have brown blood, not many though.

This could be made for each blood and keep them rare. Higher zodiac chances could have higher chances of “dropping” (possibly even more than 1 sign), while “going up” would be rare. Most Pisces would get Violet blood, and only a combination as rare as teh candy rare blood would allow them to have Fuchsia blood, the highest blood colour.
Virgos would have a rather big chance of dropping/going up, thus making Jade blood rare, even though it’s in the middle!

But that would be really much work, since I would have to make table for each zodiac sign and its possible combinations with the other signs!
How many that would be? I’ll tell you:
Approximately 12*12*12*2 (since most signs can only differ between two 12 zodiac signs, I think only 2 between 3) combinations, that’s 3756!

As you can see, that’s a lot of work, and I won’t do that very soon. On the other hand, spreading 13 blood colours between 3750 possibilities would give me a good way to control their rarity. Maybe I’ll do that slowly somewhen later.

What YOU can do now to make your result a tad more realistic is using the geometric or harmonic mean. Here’s how they work:

GM: Is lower than the arithmetic mean, but higher than the harmonic. You multiply the four values and do the nth rooth, where is the number of values. Example for my values:
10*1*5*9 = 450
450^(1/4) or  (x=450), that’s around 4.6 ~ 5.

HM: the lowest of the three means.
You divide 1 by each of the values and add that. Then you divide the number of the values by that. Example:
1/10 + 1/1 + 1/5 + 1/9 = 1,41
4/1,41 =  2,8 ~ 3

If you have other suggestions, comments, or complains, I’m eager to hear them! Leave me a message.