The problem with the rating system is not the rating sytem, but the American educational system. Most Americans have spent 12-16 years of their lives being taught that 70-75 is a passing or average mark. And it's not just here, if you've ever had the oportunity to rate another person's performance in the workplace, you'll know what I mean. When a company, that uses a 1-5 scale with 3 being the expected rating for 80% of the employees, try explaining to someone that a 3 is a solid performance that meets and sometimes exceeds the company's expectations without a desk between you and that person and a box of kleenex at hand.

I agree with the ratings being weighted more toward the opinions of the senior folks here. I also think that the scale should be lowered to 1-5 to also help reduce the great fluctuations in ratings, that anything 3 or above on the 1-5 scale should be visible for all, and that anything rated lower than a 3 expire after being available for 30 days.

I also agree this is better. Great idea Frogboy. Who's the man?!!

I'll have to agree with that, 5 out of 10 signifies to most people 50% which is what we most commonly see represented as an F

100-90 is an a
89-80 is a b
79-70 is a c
69-60 is a d
59-0 is an f

Most people including myself have always looked at it that way, getting out of that mindset won't be easy.

Ut oh. I'll leave my male hormones out of rating things on the 1-10 scale. (hey, its a joke)

Actually, that 70 percent mark being passing or average in school is not uniquely American. Give or take a couple of points, it's a common standard in many countries, and for good reason. On a standard multiple-choice test with 4 choices per question, someone who selects their choices completely at random would be expected to select the correct answer 25% of the time. Allowing for this, people who have been exposed to and taught the material are actually expected to do better than this random result, and in fact this has been proven to be the case after literally billions of records have been evaluated, with a "normal" mean score of about 67-75 percent, depending on the study (Source: National Center for Education Studies http://www.nces.ed.gov)

Although the reasons for this average being above 50% are quite different than the reasons I gave in my previous post (why I expect ratings here to be higher than the 5.5 arithmetic mean), I predict the results will be, in fact similar, with the real average rating being somewhere in the 6.5 to 7.5 range.

Will shut up now that my 2 cents is up to over a dollar

Idea sounds good however i think that IR and several others have a very good point when it comes to perceptions of what is average and i think that this should be catered for within the system - if its possible

I think you'll find that the reason that this so-called 'average' in schools ends up at 70% or whatever is simply to do with determining a quota for placement to the next year's class.......the fewer the potential places available the higher the 'passmark' is set to compensate.

It's been a common system in determining entry requirements into University courses for as long as I can remember.

This again still has nothing to do with 'passing' and average ratings.

This is a simple gradation between hot and cold....somewhere in the middle is what is known as 'average' weather/temperature.

It's not "wow....it's 75 degrees outside....summer passed"....

It's "wow....75 degrees outside....a typical summer".

Ratings do not indicate anything more than a mathematical consensus of opinion.

If people start interpreting some esoteric nonsense about a predilection/predisposition to not post an average '5' as a '7' is more polite or considerate then their posted ratings are farcical in the extreme.

Look at your petrol [gas] gauge.....

Empty

Half

Full

Nowhere will it say "oh, you got a passable amount of fuel cos you are at 70%, man, so you're cool"....

It'll just say

Empty

Half

Full.

Ratings do not indicate anything more than a mathematical consensus of opinion

wherein lies the 'contradiction' - whereas mathematical consensus is a logical process, human opinion quite often is not

Humans are not logical...

I agree with the way that Paxx's except that the coloring and the would I use question can really through thing off.
For instance, I have rated some skins that I would never use on my computer because of the "color or the theme of the skin" but I rated it an 8 or a 9 because of the quality of the skin. "Yes, I first downloaded it and took the time to completely look it over."
Now if I chose the correct answer in the "Are the choice of colors attractive to your taste" as a "Could be better" or "not my piece of cake" and also chose the correct answer in "Will you use this skin" as a "no, I won't" because it is true, the rating is a 7. The thing is I know this skin is truly worth a 9. Just because the skin is not something I would use because of color or theme does not take away from the quality of the work and the time involved from the author designing it.
I feel the best thing to do in these cases is to look over these 2 particular questions because if I had liked the color and would use the skin it would calculate as a rating of a 9 which is what I felt the skin deserved to begin with.

I feel that no skin should be rated with out it first being downloaded and looked over. It is very unfair to everyone when skin are rated just on the preview pic. Also that just because a skin is not something you would use does not take away from the quality of the skin in some cases.

JMO

Tiggz....Harry Seldon would have you determining the logic of future uncertainty through mathematics....

Ahh good ole Harry Seldon, he should start a Foundation or something.

heck, why not go the whole hog and start an Empire?

or put Mr Seldon and Mr Heisenberg in a room and don't let them out until all this uncertainty business is sorted once and for all!

Look at your petrol [gas] gauge..... Empty Half Full

Your gas (petrol) gauge is not used that simply. First who lets their gauge actually get to empty before deciding to push their car to a gas (petrol) station. During winter months, unless I want a frozen gas (petrol) line, I don't let the gauge go below half. So during winter months the average amount of gas(petrol) in my tank is 3/4 of a tank. During summer months, I don't let it go below 1/4 full, so the average amount of gas (petrol) in may car is 2/3 of a tank. So somewhere between 66-75% is the average for my car not 1/2 or 50%.

aaahhhh, statistics, so much more an art than a science

During winter months, unless I want a frozen gas (petrol) line, I don't let the gauge go below half.

Thats funny since Gas will not frezse at all.

Diesel will.....i have stories!

Gas will not frezse at all

but condensation of moisture in the gas tank will

Forgot about that.

Ratings...Schmatings. Outside of a 10 or a 1, I figure everything's subjective. Eye of the beholder stuff. And when you consider tools like me, well, I may not know skinning, but I know what I like. I imagine that among talented skinners, peer ratings would be very important. They could probably give a flip about what I think of their work. I've uploaded a grand total of 2 screenshots and nothing else; one got a 6 and one a 7. That was nice, but even if they got a 1 or 2 it wouldn't change the fact that I like them and wanted to share with the other Mac GUI wannabes out there.

Sorry, I'm rambling. I would value a library section where I could peruse only the Apprentice or higher skins, or only the Master skins, and so on. That would mean more to me in my searches than ratings.

FWIW,

Jeff



[Message Edited]

People voting 1 or 10 need to be rounded up and get their rating priviledges removed. Same for people who give everything they like an instant 9.

Ratings are a waste of time, but a suggestion anyhow: if this merit system we have here (citizen, apprentice, journeyman and so on) is to really have an impact and balance things, make the weight exponential, not linear.

If a citizen votes 3 and a journeyman votes 7, make the citizen count as 3^1 and the journeyman as 7^4 (yielding 3 and 2401 respectively), meaning that a journeyman's vote counts 10^4 = 10.000 times. So the exponential weights for repectively citizen, apprentice, master apprentice and journeyman would be 1, 2 (100 votes), 3 (1000 votes) and 4 (10.000 votes). Even more for wizops and higher (imagina a wizop having weight 5 (100.000 votes) rating something an 8 ~ 8^5 = 32.768, let them try voting that down, eh?).

The distribution of members across these classes probably isn't linear (there will be far more citizens than apprentices), so why should the weighing be that way?

[Message Edited]

Crae.....it doesn't need to be exponential....that'd wipe out any meaning to anyone's rating contribution once an admin rated something......like if I gave something a '4' it'd take 2000 years of dedicated consideration by all and sundry to bump it to a '5'....

That'd be OTT in the other direction from where we are now...

Yeah. Would be great though.

Perhaps a base two exponent would be better: citizen (level 0) giving a 6 would be 6*2^0 = 6, journeyman (level 3) doing the same: 6*2^3 = 6*8 = 48 = 8 votes of 6. (admin: 6*2^4 = 6*16 = 96 (16 votes of 6)).

At any rate, my point is that I don't think a linear scale is "heavy" enough to balance things. But we'll find out soon enough.

Actually one thing I'd be in favor of doing ON TOP of this change would be this:

Having what is displayed in stars.

You rate 1 through 10.

But then the actual star rating (1 through 5 stars) would be based on that skin's rating relative to other skins. I.e. on the curve.

So if the average skin is getting a 7 then 7 would be 3 stars.

sounds cool

Frogboy, why don’t you say what you mean in everyday English.
The current rating system is not working in your favor, and must
change it to benefit Stardock/Wincustomize.

Now if a Citizen’s rate is 1 vote and an Journeyman’s rate is 5 votes,
WC can control what skins are shown to non-members based on the
current rating of 8 for walls 7 for skins, thereby controlling the
bandwidth used by non-members.

This makes more sense to me than a rating from 1 to 10 and 5 is
average, or 70% of 100% is average. The quality of work
performed by the skinner is not a factor in this proposed rating
system, it merely serves the purpose of offering the purchaser the
right to download ‘free’ skins without joining the site and still
maintaining control of the bandwidth.

Therefore, the new rating system will not take into consideration the
work performed by the skinner, only the number of skins available to
non-members. Sorry if this is something you don’t want to hear, but
business ethics dictate the means....