usagijen managed a great mixture of closing thoughts and defense of her giving Darker than Black a 10 [out of 10]. I added my own response, but ghostlightning pointed to one of coburn’s December posts which I recall, but never managed to respond to a second time. There were good points made:

Jen:

My 10 is greater than your 10

This is a duality in truth.

moritheil:

A rating is always subjective.

I suppose concur, or rather, no matter how objective one may deem their rating, we must still take it as a subjective stance Nothing wrong here, the Narutos of the world will still get their 1’s and 10’s… it keeps spinning /whoosh so why do ratings matter at all on a system-wide scope? I have arguments, but not the point of this entry.

ghostlightning:

People are subjective to be sure, and the ratings we give on MAL I think serve as icebreakers in opening discussions more than anything.

This goes beyond the way (the rating system 10/10 or relative, matters not). Discussions are about thoughts, possibly streams of them. This goes hand in hand that a rating list only goes so far when trying to inspect the individual. An organized history of thoughts are a nice reference point in doing so.

Senna:

I know people who won’t even give their favorite series a 10/10 because “it’s not perfect;” I think that’s silly

I concur, though just as interesting is someone’s highest-ranked title. It’s just not very interesting when it shares the same position with 20 other works.

Finally, coburn’s followups:

I agree that relative-tiered levelling is more accurate in dealing with the broad range of shows out there – especially at the lower ends of the scale. Especially in that some people will be stricter than others depending on their personal aims in rating shows.

Ah yes, notice this statement on strictness will apply to any rating system. People who rate differently, should be allowed (and possibly encouraged) to use different scales. coburn does, in fact, understand the system.

Still the lack of a magic symbolic 10/10 doesn’t quite click with my actual experience. My favourites really do mean that much more to me. Maybe I’d have to include a couple of empty theoretical tiers to segregate them from the crowd.

Ah yes, the magic of 100%. I’ve come to the realization, that this is only possibly in relative ratings if the user keeps a 10-level system, but again, possibly more interesting is what appears at the top, regardless of level. Y/N? A top 10 list is rather wicked, but something tells me people wouldn’t mind seeing top 20 or top 100 as well. I feel a correlation with ratings is in order.

Secondly, empty null/void-tiers, it has been pondered. This is not an issue, the current system has a switch which removes empty tiers, disabling is quite simple. The issue is exploiting such a thing, but extreme ratings should usually be thrown out anyway. As compensation, the current calculation algorithm is non-linear, meaning higher tiers have a greater weight/position than lower tiers (the mathematical significance of a rating becomes less important on the way down).

TheBigN:

I do get annoyed when people strongly force comparisons between two anime shows, as if the shows were specifically made to fight against each other, when that’s not the case.

Back to coburn:

TheBigN: I reckon the advantage of doing things out of 10, as opposed to something like Ryan suggests, is that the extent to which shows are in combat with one another is reduced.

It is relative because the items are ranked relative to each other… this is Sparta, jk.

There are two ways to go about this. 1) Whether we like it or not, a 10-point system is relative, but abstracted from the direct comparison between the experiences/works. If someone gives two titles a 5 and a 7, there is inference of comparison. 2) Ratings are quite 1-dimensional, which is good reason to include annotation when rating. This preferably should not involve comparison between works, because the rating already expresses the outcome.

The fundamental nature of rating is comparison. How an individual rates, whether through abstracted grading or direct comparison likely will differ. The dimension of the ratings will likely differ as well; the overall experience vs storytelling+production, etc. What is important to notice in all of this is that both fixed and relative rating systems can be used for any dimension or grading style. If we substitute the word “unbounded” or “dynamic” for “relative,” the hard difference becomes the ability of a user to create their own rating system, or not.

With that said, I realized this rating system requires a different mindset about how we numerically categorize our media experiences, but once the consensus releases the ingrained concept of 10/10 ratings and realizes the backwards compatibility of relative ratings, both RRS and fixed-point systems will reach a more meaningful realization.

More…