Dominance Orders for furs, eyes & traits in Google Docs

[Tad Carlucci]

Quite a bit of Sega's list merges information I gave her. I don't know if she ever explained my process and logic, so here it is. In the past, I've made some comments about the size of my database and the ease of gathering it. Some have voiced concern that I use information not publicly available. I don't. And I don't use robots. But it's been requested that I publicly describe my methods, so here they are:

Every so often, I run around several regions clicking (yes, BY HAND!) on cats and boxes. How often do I do this? Well, I generate my list so I can give my wife sound advice on her breeding. So, not very often .. only when I find I'm saying "dunno" too often. Where do I go and what do I look at? Only marketplace regions where KittyCatS are placed on public display. The only private stock I add to my database is my wife's. So if you've seen me standing silently, perhaps jumping from shop to shop, but not appearing to DO anything .. now you know what I'm up to!

I record my entire Local Chat for those examples. I then pass that text file through a Visual C# program which parses the text into spreadsheet/database form; which I then import into a database.

When I merge the data from the areas I've just visited with my older data, I end up with a collection of several tens of thousands of records showing the offspring's visible traits and each of the parents visible traits.

At a first pass, I reduce this data set into 'interesting combinations' .. my rule for that is that, for a given trait, the offspring must be different than either parent. I then eliminate duplicates so that each record has something interesting and is somehow different than all other records.

Now we get down to the actual logical rules. I'll do our common notation as some examples to help you see what I mean.

If both parents show the same value for a trait, and the offspring shows a different value, the trait shown by the parents is definitely dominant over that shown for the offspring. I create a new database showing just than. (Ab : Ac -> Bc) (Ab : Ab -> Bb) [A must be dominant over B]

If the parents show a different value for a trait, and the step above showed that one parent is dominant over the other, the offspring's trait MUST be recessive under the dominant parent's value. I add that pairing to my database. (Ac : Bd -> Cd) (Ad : Bc -> Cd) (Ac : Bc -> Cc) where I know A is dominant over B [A must be dominant over C]

If the offspring has a trait which has been shown to be dominant over one parent's trait, that trait must have come from the other parent and that parent's trait is also dominant over the offspring's. (Ab : Cd -> Bd) (Ab : Cc -> Bc) where I know B is dominant over C [A must be dominant over B]

If I have a rule in my database, such as "Ab" (A is dominant over b) and I have another rule, such as "Bc" (B is dominant over c), I add a rule, such as "Ac" (A is dominant over c).

I repeatedly perform the above steps until no new relationships between traits are found. At this point, I have a dominance table for a trait-class (say, Fur). It's most likely incomplete.

For each trait, I count the number of traits dominant over it. If the table is complete, I will have a clean sequence counting from 0 to N minus 1 (where N is the number of traits observed). As I said, that is unlikely. More likely is there will be several with the same number of traits dominant over them, and some skips in the counting sequence.

I order my traits starting with the one with NO (zero) traits dominant over it and place all those with the same number together on the same line, wrap it all up in a note card, and pass it to Sega. She merges my list with hers, applying information she has I don't; perhaps to separate and order some I can't tell the order of; perhaps correcting some of her orderings because I have hard data showing something she might not.


ETA: While this method works well for most breedables on Second Life, there is a class of breedables where it will fail. In the phase where I take Ab and Bc to produce Ac as a rule, Meeroos, for example (there are some others), will produce an infinite loop. The reason is that, while KittyCatS! and many other breedables use what I call "simple, linear dominance" .. a strict implementation of Mendel's Laws .. some use more advanced genetic systems. The mere fact that you find A? : A? -> B? and also B? : B? ->A?, or any sequence which leads to a similar result, is sufficient to prove that that breedable does not use the simple genetics system present in many Second Life breedables. Furthermore, while such a result proves simple linear dominance is not being used, it does not prove randomness or any other genetic system.