Hiding Confusion

[Liriel Garnet]

(03-12-2012 08:17 AM)Tad Carlucci Wrote:  Developing that "web of trust" is hard enough. All the false and misleading claims just make it harder. Perhaps the KittyCatS community can come together and find some third party who can verify such claims. But, experience says even that won't fully solve the issue.

I honestly don't think this is feasible given both the number of cats and the fact that certain analyses rely on knowing the dominance of the traits involved and in some cases we cannot be entirely certain. Not to mention it'd be a huge task for whoever was doing it and they'd have to have unfettered access to the pedigrees of everyone wishing to have a verification done.

I think that we all agree that false and misleading claims make it far more difficult for the whole community, thus another thread a few weeks ago about 'beware of false claims' or something similar. That was the topic, at least, even if I don't have the title of the thread correct.

Personally I use three definitions when dealing with cats and traits:
-- showing: kinda obvious *chuckles*
-- hiding: shown in the parents or verified via breeding *or* can be deduced from what's shown and known of the parents (e.g. cross one hiding a black whisker and one hiding white, if a white shows up, you know the black has to be hiding)
-- possibly hiding: this is generally either a 1/2 (one parent known to be hiding) or 2/3 (both parents known to be hiding but not shown in the box) chance that the trait is hiding.

There's one further clarification that I use when tagging mine. If, for example, I cross a cat with a mysterious tail (assume hidden is also myst for this purpose) with one that's showing curious and hiding shorty -- i know it's hiding at least curious, so i'll note that as curious+ because i can't be sure whether the hidden tail of the box is curious or shorty. I do this for a reason, because if you're looking for a very specific trait, you want curious for example rather than shorty, I can't guarantee the curious is hiding there, just that it's curious or more recessive.

You're absolutely correct that this issue isn't about one person, it affects everyone in the community. I personally don't buy a cat unless I've had a chance to examine the pedigree and verify for myself what it could or could NOT be hiding in the background. If the pedigree doesn't support the claim, I'll speak with the seller to see if further proof is available. I do this, even if i know the seller and trust them, because we are all human and can make mistakes. I encourage every single person to do this as well.

My shop has a pedigree giver that will produce a link for every box or cat in the shop. I also have a link giver to Saga's charts because I've personally found those the most accurate and up-to-date of the publicly available resources. I have them set out specifically to ensure that people can examine the boxes and cat's provenance for themselves before purchasing. I also do a bit of helping others learn how to accurately assess this for themselves.

Quite frankly, if I ever sell a cat or box that's not hiding what I say it is, then I hope the buyer will come back and let me know so that I can rectify the situation. I'm not perfect, but I always do my best to correct and compensate for any mistakes that I'e made. I have no desire to get a reputation for making false claims and have taken every step I can to ensure this doesn't happen.

After all this, one possibility *does* occur to me. It might be possible to create a 'trusted breeder' status, but that carries all sorts of other potentially troubling concerns: who decides who's trusted and who isn't? how does one get added to or removed from such status? Definitely a Gordian knot.

[Tad Carlucci]

Actually, I was thinking more along the lines of a KittyCatS official service, which probably could be totally web-based, and could certify certain claims. It would take some design work to prevent abuse or information leakage and, of course, should be extremely expensive. But I think the only authority everyone could agree upon would be KittyCatS themselves. Off the top of my head, I'm thinking something like you go to a web page, enter the ID for a cat you own, check ONE claim (Snow HIDING Lynx, for example), pay L$5000, and you get a certificate that the claim is correct or incorrect which you can rez and which clicks to a web page on the KittyCats site showing the pedigree, the claim, and the result. Some checks to prevent silly claims (Seal HIDING Genesis, etc) and perhaps a limit on the number of checks per ID, times per day/week, whatever. Just a quick idea .. it needs more thought especially in the policy and support arena .. but it might be doable.

Given the pedigree links the breeder uses to rationalize the claim, I could probably write a similar web application and give odds, based upon the visible information, of what is hidden (and odds that something unknown is hidden). But I'd not call it a certification .. it would be an analysis based upon available evidence. Before I'd even start such a project, though, I'd have to sit down with KittyCatS Resident and hash it out to ensure he'd be comfortable with my proposed design, especially with the scripted access to the web pages (scraping).

[Liriel Garnet]

(03-12-2012 03:23 PM)Tad Carlucci Wrote:  Actually, I was thinking more along the lines of a KittyCatS official service, which probably could be totally web-based, and could certify certain claims. It would take some design work to prevent abuse or information leakage and, of course, should be extremely expensive. But I think the only authority everyone could agree upon would be KittyCatS themselves. Off the top of my head, I'm thinking something like you go to a web page, enter the ID for a cat you own, check ONE claim (Snow HIDING Lynx, for example), pay L$5000, and you get a certificate that the claim is correct or incorrect which you can rez and which clicks to a web page on the KittyCats site showing the pedigree, the claim, and the result. Some checks to prevent silly claims (Seal HIDING Genesis, etc) and perhaps a limit on the number of checks per ID, times per day/week, whatever. Just a quick idea .. it needs more thought especially in the policy and support arena .. but it might be doable.

Most interesting. Personally I spend enough money with KC on milk, kibble, etc. I'd far rather spend my lindens on cats and upkeep and rent. I can't say I'd support this for the simple reason that it would only be useful for one cat at a time and sorry, but I've never been one of the lucky ones to get one of the super-recessive furs and such that bring in near enough lindens to justify that kind of expenditure.

I'd rather see something that certifies the breeder's veracity than verifying one thing about a single cat, but administering that would be a total nightmare, subject to all sorts of abuse. But that would also mean I could trust buying any cat from that breeder, not just super expensive and rare ones.

I honestly think the best solution to this is education. Helping people to learn how to read pedigrees and recessiveness charts so that they can verify things for themselves. Once people learn how to assess what their cats may or may not be hiding and how to assess the claims of others, many of these issues will go away.

[Tad Carlucci]

Oops. See my edit, added a paragraph about a non-company alternative.

Yes, well, as to cost, that's a policy issue. There could be a sliding scale based on factors such as number of examples found in the game, time since release, etc. But I said 5K to get across the idea that a certification, especially from KittyCatS, would want to be a "Gotta Have It, Damn the Cost" thing. For a cat where the Ask price is, I'm told, 60K, 5K for certification is not out of line. Of course, if the Ask were only L$100, the risk to the purchaser is so low that certification probably would not be demanded.

One new thought as I'm typing: If the price were high, say L$5000, it would definitely support prices in the markets. Anyone paying that much for the certificate would be unlikely to Ask less than that. It could support longer periods of higher prices, which seems to be a constant complaint across all breedable communities.

But, this is all a germ of an idea which is only .. what .. 20 minutes old now ..

As an example of how I'd analyze it, using the attachment on this thread, and no other evidence:

---

Either ID ending 3c48 or ID ending 97ec hides Blue. Odds: 100%

ID ending b07f hides Lynx, or a fur trait more recessive than Lynx. Odds: 100%

Assume that neither ID ending 3c48 nor ID ending 97ec hides Lynx, and
assume neither hides Black Silver, then
the odds of the claim about ID ending 39bd being true are: 0%
however, in this case, ID ending 8aa1 and ID ending 39bd both hide something more recessive than Black Silver which is NOT Lynx
and, either ID ending 3c48 or ID ending 97ec (whichever does not hide Blue) hides that fur trait as well.

Assume that neither ID ending 3c48 nor ID ending 97ec hides Lynx, and
assume one hides Black Silver (the other definitely hides Blue), then
ID ending 8aa1 hides Black Silver, and
the odds of the claim about ID ending 39bd being true are: 25%

Assume that either ID ending 3c48 or ID ending 97ec hides Lynx (the other definitely hides Blue), then
ID ending 8aa1 hides Lynx, and
the odds of the claim about ID ending 39bd being true are: 100%

More information is needed to determine which of these three cases is correct.

---

It would take some work but I could write a web page to do the analysis and present English results.

Edit: corrected the odds in the second case from 50% to 25% .. I really DO need a program to keep track of all the factors and odds Notice how I also stealthily avoid claiming the TRUE odds by stating it as three cases. Assume there are 20 furs (for discussion ONLY) more recessive than Blue. The odds of case 1 are 18 in 20, the odds of case 2 are 1 in 20, the odds of case 3 are 1 in 20. Net odds for the correctness of the claim are therefore approximately 5%.