Getting 95% girls

[BazGumNut Resident]

Is something wrong with the randomizing algorithm for deciding gender? I've been wondering about this for weeks now as I am getting exclusively girls for my Australian Mist new kittens. I have recently only had males for Snowshoe creams and Abyssinian and those are rare throwbacks to furs that I am not breeding for.

This is very frustrating. I have had up to 20 breeding pairs (about 14 now) and I've noticed this trend for a number of weeks now. I have not collected the exact data but we are talking in the order of 95% female over 40 or so new kittens. What are the odds against that? Very very high.

Has anyone else noticed a similar trend?

[Shamu077 Resident]

(05-13-2016 01:41 AM)BazGumNut Resident Wrote:  Is something wrong with the randomizing algorithm for deciding gender? I've been wondering about this for weeks now as I am getting exclusively girls for my Australian Mist new kittens. I have recently only had males for Snowshoe creams and Abyssinian and those are rare throwbacks to furs that I am not breeding for.

This is very frustrating. I have had up to 20 breeding pairs (about 14 now) and I've noticed this trend for a number of weeks now. I have not collected the exact data but we are talking in the order of 95% female over 40 or so new kittens. What are the odds against that? Very very high.

Has anyone else noticed a similar trend?

Sure, I reported this 2 years ago, and was just told it was "random variation". I found a preponderance for one gender in a long line of related breeding pairs and their offspring. When I introduced "new blood" into the line, the gender ratio went back to more normal.

Despite the assumption that gender is always 50% weighting in all cats, it has never really been proven. Kittycats have all kinds of weights built in, for size as an example. Breed enough cats and the overall gender ratio will approach 50-50. But your example of extreme odds tend to be too common. Conspiracy theories don't play well here, so just enjoy the scripting.

I gave up talking about cats that refused to reveal a hidden trait in 15 tries, and aged out still an unknown. Most of the time you get success after 5 tries.

[BazGumNut Resident]

All my cats are relatively young. There is not a lot or interbreeding across the twenty pairs. Currently, they are all popping an outrageously skewed proportion of females.

To the scripters of this product ... this goes to your credibility. You are making claims that there is a 50% chance of getting a male or female, as modelling genetics dictates. If you are colouring your algorithm to game this is any way whatsoever; to say, make me go out and buy males, then it should be investigated and exposed.

Folks have a right to know how the algorithm is influenced, especially if you are making claims in your sales pitch to that contradict what you actually do.

[Tad Carlucci]

You know the algorithm. There is nothing you can do to effect it. It is JUST a random number.

The problem lies with the user. As a general class, humans do NOT understand 'random'; which leads to posts as above. For those of us who do understand randomness, your posts actually verify that the randomness is, in fact, random.

[Alassariah Resident]

My confettis gave me a whole bunch of girls and finally gave one boy last week (yay! woot woot!) but my pedigree page says:

Total: 60 Cats (♀ 30 ♂ 30) | 210 Boxes (♀ 106 ♂ 104)

Still looks random to me.

[Tad Carlucci]

If, by 'random,' you mean an exact 50/50 split ...
no, actually, that does not look random.

It should be 'close',
it should be 'around'
but it should only very rarely be 'exactly'.

There should be surprising long runs of 'no girls' and/or 'no boys',
just as there should be the odd times where it goes boy-girl-boy-girl-...

If a large number of people present pedigree counts showing exactly 50/50 splits, I would look for player-action as the cause long before I presumed a problem in the underlying system. Most likely, it would simply mean people were trading or menagerie-ing to get to an even split.

---

Long ago, we're talking years, back when KittyCatS had only been out a few weeks, KittyCatS Resident and I had a private talk about it. He'd been monitoring the split and it was running around 43% female at the time. There was a little concern that this was a bit too far off. But, a few days later, it was running about 47% male. So it was declared 'good' and that was that. I image, now, with millions and millions of data points, it's running very, very close to 50% but, still, it should almost never be exactly evenly split.

[Alassariah Resident]

(05-13-2016 02:11 PM)Tad Carlucci Wrote:  If, by 'random,' you mean an exact 50/50 split ...
no, actually, that does not look random.

It should be 'close',
it should be 'around'
but it should only very rarely be 'exactly'.

There should be surprising long runs of 'no girls' and/or 'no boys',
just as there should be the odd times where it goes boy-girl-boy-girl-...

If a large number of people present pedigree counts showing exactly 50/50 splits, I would look for player-action as the cause long before I presumed a problem in the underlying system. Most likely, it would simply mean people were trading or menagerie-ing to get to an even split.

---

Long ago, we're talking years, back when KittyCatS had only been out a few weeks, KittyCatS Resident and I had a private talk about it. He'd been monitoring the split and it was running around 43% female at the time. There was a little concern that this was a bit too far off. But, a few days later, it was running about 47% male. So it was declared 'good' and that was that. I image, now, with millions and millions of data points, it's running very, very close to 50% but, still, it should almost never be exactly evenly split.

Not sure who you're speaking to here but; as for me, I only meant mine seem to balance out to an approximate 50/50 split with very little focus on sales or menagering offs. I haven't counted those few in the numbers from the pedigree page. If I only count the Confettis which are very challenging to breed then yes, I had a long run of girls which as you stated is to be expected but overall, the system seems to work.

[Shamu077 Resident]

(05-13-2016 05:35 AM)BazGumNut Resident Wrote:  All my cats are relatively young. There is not a lot or interbreeding across the twenty pairs. Currently, they are all popping an outrageously skewed proportion of females.

You will always be blamed for this - humans are imperfect and can not count reliably

My present inventory of 1228 cats shows 616 Girls and 612 Boys.
That is 0.5016 girls and 0.4984 boys. This does not prove all cats have a 0.50 gender weighting. It is rather easy to program such that there is true 50% weighting for large numbers, while individual cats could have random weights that persist.

Pick a random number between 20 and 80. Do this 10000 times. The average will be close to 50. The larger the sample, the closer it will approach 50% weighting, if the random selection function is truly random. Now assign these weight gender numbers (that were randomly selected) to a cat when it is first rezzed (when the other Traits are loaded from the server.) It would only take a few lines of extra code to make individual gender selection a weighted number instead of 50%, and this code is already present to determine size.

Of course this is all implausible, since the only effect would be to force stubborn breeders to keep trying (and buying kibble) until the cat is retired. Why would anyone want to do that?

[BazGumNut Resident]

(05-13-2016 08:19 AM)Tad Carlucci Wrote:  You know the algorithm. There is nothing you can do to effect it. It is JUST a random number.

The problem lies with the user. As a general class, humans do NOT understand 'random'; which leads to posts as above. For those of us who do understand randomness, your posts actually verify that the randomness is, in fact, random.

Yeah, thanks for the patronizing reply. I know what random is. I am questioning the probability. Someone was always going to be dismissive of my query.When the probability drifts, I think it's worth questioning the programming, which is, at best pseudo-random, and can be easily twiddled.

So let's get a few numbers out there. From my current live cats and boxes (i.e excluding cats in the menagerie for which it would be harder for me to gather data):

Of the 8 Australian Mist Snows I have bred, all 8 are female ... that's a probability of 1 in 256. For as long as I do not produce a male snow, I'll be wondering.

Of the 30 cats I had bred (and kept) since 20th April, 25 are female. Before that, there was a more even distribution.

Not the worst streak in history, and of course the sample size is small, but you should be able to see why I am asking questions.

[BazGumNut Resident]

The run continues:

▼▼▼▼▼▼▼▼▼▼▼▼▼▼▼▼▼▼▼▼▼▼▼▼▼▼▼▼▼▼▼▼▼▼▼▼▼▼▼▼▼▼
♀ New Born Kitten

Fur: Abyssinian - Dark Chocolate
Eyes: Exotic Treasure (Shape: Curious | Pupil: Small)
Shade: Porcelain
Tail: Dreamy
Ears: Odyssey Boo Boo
Whiskers: White (Shape: Light Wave)
Size: Normal

Version: 1.55
Owner: BazGumNut Resident
ID: 022bf113-9e9e-5008-ab8c-0e4947d6553f
Pedigree: https://kittycats.ws/p.php?60c66d17-0252791-00543
〓〓〓〓〓〓〓〓〓〓〓〓〓〓〓〓〓〓〓〓〓〓〓〓〓〓〓〓〓
MOM: Waratah Adia
Fur: Australian Mist - Snow
Eyes: Exotic Treasure (Shape: Curious | Pupil: Small)
Shade: Twinkle
Tail: Posh
Ears: Soft Curl
Whiskers: White (Shape: Odyssey FrazzyWave)

DAD: Waratah Ruben
Fur: Australian Mist - Snow
Eyes: Pewter (Shape: Mysterious | Pupil: Small)
Shade: Porcelain
Tail: Fluffy
Ears: Odyssey Boo Boo
Whiskers: Latte (Shape: Light Wave)
▲▲▲▲▲▲▲▲▲▲▲▲▲▲▲▲▲▲▲▲▲▲▲▲▲▲▲▲