Schrödinger’s “WeSmellBetter” — when we are both wrong and right at the same time (Exceptions)

Our goal at WeSmellBetter is to accurately capture the publicly published results of AKC Scent Work trials and compile and report that information in ways that are useful and interesting to the participants.

It is our firm belief that our database accurately reflects the data from the published results within these timeframes:

• We collect and process trial results every night.  So if the AKC published a Q for your dog on Tuesday, by Wednesday, our reports should show that Q.
• We manually trigger a procedure to collect information about future trials for our Trial Finder function.  We usually do this on Friday nights and Tuesday nights, but will admit that sometimes it just happens when we remember to do it.  The function only gathers info on up to 1000 future trials, so it doesn’t always grab everything that is out there.  [Sidenote:  future trial info evolves quite regularly, usually from a rough template of “we’re going to offer absolutely everything” to the final approved version that often shows fewer classes offered, but has judges assigned.  As long as WeSmellBetter Trial Finder shows a scheduled trial, it will have an easy link to the AKC Event, too.  I always struggle with the AKC’s Event calendar, so usually use Trial Finder to identify the trial I want to know about and then use that link to the actual AKC Event with the premium (eventually).]

If anyone finds a situation where they think our data doesn’t accurately reflect the AKC’s published results, we want to hear from them immediately – they should use the form under the Contact Us tab at the top of the website.

There is one common case when people incorrectly *think* our data doesn’t match results.  That’s the situation where the name of the dog in our reports doesn’t show all of the dog’s titles.  I’ve created a separate post to explain this.

Otherwise, we are under the impression that our data matches Results data.

But if that’s true, how can there be Exceptions — dogs whose name shows Scent Work titles our data doesn’t support?   Or dogs for whom we have calculated Scent Work titles but their name doesn’t include them (and they have at least one Scent Work Q after that, see that other post)?

The answer is basically that the published Scent Work results don’t tell the whole story, and, unfortunately, they are the only parts of the story that we have.

We have a utility that we occasionally run on our data.  It compares the titles we calculate for a dog to the titles in the dog’s name to identify the Exceptions described above.  We then investigate those to try to figure out why that might be the case.

Currently, there are less than 100 of these dogs identified; at present, that’s about 0.3% of the dogs in the database.  However, we can’t catch ALL discrepancies because the Exceptions Utility depends on Titles rather than Qs.  So there are undoubtedly more dogs for which we are “off” by at least one Q, but if that hasn’t resulted in a title, our Exceptions Utility won’t flag it.

Generally, Exceptions seem to come from one of the following:

1. Dogs that change registration numbers.  Usually, this is from an All-American number to a Purebred Alternative Listing (PAL) number, where the owner has petitioned the AKC to recognize this dog as a member of a particular breed although, as an altered dog, they will not be included in the breeding registry.  This shows up in our data as a dog who has higher-level titles without evidence of the precursor titles required, i.e. PAL 123456 Bowser SWA doesn’t have any Qs in any Novice elements and less than 3 in every Advanced element.  As long as the dog’s name wasn’t changed, we can often piece these together and find the record for MA123456789 Bowser SWN SCA SBA SIA and merge that with the PAL dog that got some Advanced Exterior element Qs to “show the work” to support Bowser’s SWA title.
2. The wrong dog registration number was recorded in the AKC’s Results.  This happens with relative frequency and it’s amazing, given how manual the whole process is, that it doesn’t happen more.  Our guess is that, most commonly, the Handler provides the wrong registration number at entry, often the number for one of their other dogs.  It could be a situation where a Trial Secretary chooses the wrong dog out of the local database created by their trial management software.  It could be a typo — transposed numbers or illegible numbers interpreted incorrectly by Trial Secretary or the AKC when they enter the results.In any case, after the fact, the Handler notifies the Trial Secretary and AKC about the error and, depending on how that communication goes, the AKC may agree to award the Q(s) and any resulting titles.In these cases, sometimes, the AKC will change the published trial results.  Our data collection does a 60-day “lookback,” checking for any changed results, and we grab those and update our database.  It is possible that a result could be changed for a trial that was more than 60 days in the past and we would miss that update; if anyone spots that, we’d love to hear about it. 

   However, it seems that, more frequently, the AKC makes the changes directly to the dog’s record.  This means that the published Results still show Qs awarded to the incorrect dog and Qs missing from the actual dog — and our reports show what the Results do.  So, to be clear, our data is, as was our stated goal, consistent with the AKC’s published results.  It’s just not “right.”  

3. We have recorded Qs that would support AKC-awarded titles but, according to our logic, at least some of the Qs were Excluded.  Exclusion Qs come from

a) entering a class without having the qualifications to do so, i.e. getting a Q in Master Container when you haven’t yet been awarded an Excellent Container title.

b) the dreaded “Novice A” exception where the closing date for the trial in which the Novice A Qualifying Score occurred was AFTER the date on which the Novice element title was earned.

c) when there are Qualifying Scores in more than two levels of the same element, only the two highest levels are counted and all others are excluded.

Honestly, I’m not sure what to tell you about these except that it does appear from the data that the Exclusion rules have not necessarily been applied consistently by the AKC.  I don’t know how manual a process this is for them or if a lot of different people are involved; I wonder how well-articulated the Novice A rule, itself, was before we asked the AKC exactly what the policy is.  I don’t know whether they may take circumstances into account and make a judgement call.  But if the rule was being applied by impartial software, we believe that the tally would be more consistent with our own.

In many cases, we can only speculate about what happened.  The third case is probably the easiest to spot, because we can see the right number of Qs to support an AKC-awarded title, ours are just Excluded.  The first case, the number change, is easy if we can search the dog’s name and find the right-fitting puzzle piece.  The second case, wrong registration number, is really difficult and sometimes impossible to find.   I suspect that it accounts for most of the “I just don’t know what happened here” cases where it appears the AKC has granted a title and the AKC-published Results do not provide any data to support that award.

What does this mean for YOU?  Well, if you, or someone you know, believes that the WeSmellBetter reporting is wrong, we really want to hear from you, and would like to get to the bottom of the discrepancy.  This long explanation is to help us both understand the types of things that can happen when our data, although consistent with AKC published Results, doesn’t accurately reflect your dog’s titles.  In most of those cases, I’m going to assume you will be able to remember that there was a problem with the results reported, but that you worked with the AKC to resolve the situation.  We’d like to know the story — and are happy to keep that confidential! — and then will see what we can do with our data in response.

Even more so, if there are cases where you have been struggling unsuccessfully to get the AKC to make a change in your dog’s record, if there’s anything we can provide that will help you make your case, we would love to help.

Contact Us!