Is there a flaw in recent rating calculation?

[bela_dick]

I would like to know how the "recent sum" and recent

average" is calculated.

How "recent" is " recent"?

 

An image has 24 rateings given. The last 3 or 5 ratings

it has gone from 5.60+/- to 5.7+/-. in total average.

Yet the recent ratings of 11 is standing at an average of

5.36. Now 11×5.36=58.96. Meaning the first 13 must

have averaged 6 to get 5.7 in total average by 24 raters.

But that has not been the case.

 

 

Can anyone tell me how this is posible

[Norma Desmond]

<p>Rarely are all the individual ratings shown. That's because administration recognizes the childishness of some users of the system (mate rating and revenge rating) so they don't want anyone figuring out who the last person to rate was, so it doesn't show the last few ratings until a group of 5 or so accumulates. So very often, the total number of ratings is greater than the number of individual ratings shown. Your average is probably correct. You're just not seeing all the individual ratings that make it up yet, and you may never.</p>

<p>The flaw in the ratings, however, is not just the calculation. It's the game itself. It is still possible to mate rate and still nothing more than a random gathering of meaningless numbers, telling a photographer nothing about the caliber of her work or its overall popularity, since it's such a miniscule sample of viewers. </p>

[bela_dick]

Maybe I have not made the question clear.

Recent sum =5.36 x 11= 58.96

Recent average = 5.36.

Sofar is all clear.

 

Now Total average has recently moved from 5.60 to

5.71

So how is it mathematicly possible that Total average is

higher and increasing wenn recent averag showing only

5.36.

In this case the last (recent) rates should pull total

average down and not increase total average.

Is there someone who have the arithmetics. I would be

happy to see it.

I think it would be ok to see how the numbers are

calculated. It can not hurt to publish the formula used

for the calculations.

[bobatkins]

<p>I don't think all the statistics are "real time" and they may not all be simultaneously updated. In part this is a mechanism to defeat those who would like to micro-analyze ratings to figure out way to "game" the system.</p>

<p>In such a model the overall ratings over time would be consistent but they would not be consistent on a minute by minute basis, perhaps not even hour by hour or day by day. Small sample sets would give rise to larger deviation from what the casual observer would think is "correct" than large sample sets, so if an image has only a few ratings the number swould jump around a lot more than if it had a few hundred.</p>

<p>Whether ratings still use such a scheme I don't know, but I do know that thought was given to making ratings difficult to analyze by users whose motives were not in the best interests of the site and other users.</p>

<p>You would be amazed at the time and effort some people have devoted to figuring out who gave what image what ratings and then taking inappropriate actions based on that (possibly faulty) information.</p>

<p>Treat the numbers as general indicators of an images's popularity. Worrying over the accuracy of decimal digits in the ratings is totally pointless. It's not a contest and if it was there would be qualified judges, not a random sampling of random users.</p>

[JDMvW]

<p>I just like to show off some of my work at the bottom of the page, so I really don't worry about the ratings.</p>

<p>And a good thing too, or I'd probably be depressed......</p>

[bela_dick]

Hi Bob

Thanks for the answer. That explains a lot.

I thought "recent " really ment recent. How about

rename it to "laging rates" or " randomly picked historic

rates"

I think if one use mathematics for information purpose

then it should also follow a definition. Else it will not

have any informative or correct purpose.

Btw "all average" has increased yet again, but "recent"

unmoved. Hence lagging and not reflecting the facts.

 

I hope I have not provoked you to much. It was more of

curiosity I noticed it and could not make any sense of it.

Now I know there is no sense to it. And before the

curiosity kills the cat I set a finale here. :)

Cheers

[Norma Desmond]

<p><<<<em>Else it will not have any informative or correct purpose.</em>>>></p>

<p>Now you're getting the picture of PN ratings!</p>

[bobatkins]

<p>Ratings are a can of worms I'm not inclined to open.</p>

<p>I'd personally have avoided numbers and either presented results graphically (length/color of a bar) or descriptively (average, good, excellent etc.). If numerical measurements were thought necessary for some reason I'd have gone with 0 to 5 stars (including 1/2 stars), like movie ratings. Quoting numerical ratings to 1/100th of a point is meaningless and gives a totally false sense of precision.</p>

<p> </p>

[michaellinder]

<p>OK. here we go again. The problem with ratings is not related to how often they are averaged. The problem with ratings is that they make little sense. Bob's referring to them as a can of worms couldn't be more accurate. I'm still trying to understand why my ratings (when I was asking for them) dropped dramatically when the system was changed from 2 separate scales for aesthetics and originality to a consolidated scale. </p>

 

[michaellinder]

<p>OK. here we go again. The problem with ratings is not related to how often they are averaged. The problem with ratings is that they make little sense. Bob's referring to them as a can of worms couldn't be more accurate. I'm still trying to understand why my ratings (when I was asking for them) dropped dramatically when the system was changed from 2 separate scales for aesthetics and originality to a consolidated scale. </p>

 

[lex_jenkins]

<p>Michael, I've noticed an overall shift in ratings from 6+ to closer to 5+, even among the top rated photos. I don't think it had anything to do with eliminating the two-part ratings system. It's more likely due to another, nearly simultaneous change in the ratings system: the names of members who rated photos were not posted until at least five ratings had been accrued, and didn't directly attribute the ratings to the members.</p>

<p>Personally I believe this was a good thing. It removed some burden from members who wanted to rate honestly without being targeted for retaliation. </p>

<p>I also believe that most average ratings now on the TRP, which are usually around 5+, more or less, are about right. Expectations should be high and anything above 4 should be considered better than average. Even ratings of 4 aren't bad - just an indication that the photo is about average for this site and/or genre. There aren't many truly exceptional photos and averages of 6 or higher should be commensurately uncommon.</p>

<p>Before those changes to the ratings system it was generally known that only ratings of 3 or higher were factored into the average. That widespread knowledge meant that 3 became the new 1. And a rating of 5 was no longer "above average" but was the new 3 - below average. There were complaints from some members who felt they were being persecuted if they received ratings lower than 6, which they considered the new 4 - slightly above average.</p>

[bela_dick]

I think the system works - it is not excellent, but it

works.

And I understand that people(me inclded) can get upset

wenn they get a three or lower without comment.

However, after getting over it, if we are honest and

looking at the history of the TRP we really see the best.

Now Photo.net is full of top phothographer and

everybody cannot be on the frontpage. It is not enough

space for all the top ones.

But it is not said that the system cannot be improved.

What I have noticed is that there are raters here who

hardly have an Image uploaded, but are rating really

good to excellent photos down. These members (if not

paying members) should be excluded from rating imo.

[Norma Desmond]

<p>Bela, I have to disagree with most everything you just said.</p>

<p>The rating system is completely useless. So whether it works or not is irrelevant. It's trivial. It's a game.</p>

<p>The TRP represents a certain kind of photo. They are not the best. They are the most popular. </p>

<p>No one should be excluded from rating photos. Many fine critics of all kinds of arts do not practice that art. Rating takes an eye, which many people have who don't post their photos. There are many reasons even good photographers don't post photos on PN. There are many non-photographers who have better eyes for photos than some of the photographers here, and especially some of the over-saturating cliché-laden photographers who make it to the TRP because they're being rated by other over-saturating cliché-laden photographers.</p>

<p>The TRP should be renamed to the LCD, Lowst Common Denominator. It is where originality, thoughtfulness, intimacy, risk, and taste often go to die.</p>

[mike dixon]

"What I have noticed is that there are raters here who hardly have an Image uploaded, but are rating really good to excellent photos down. These members (if not paying members) should be excluded from rating imo"

 

Following that logic, people who don't cook regularly should not be allowed to say whether a meal tastes good, and people who are not filmmakers shouldn't be allowed to say whether or not they enjoyed a movie.

[pnital]

<p>LCD,( Lowst Common Denominator)...7 +7 ;-)) for the creativity!</p>

[michaellinder]

<p>Lex, I accept your explanation, and I do appreciate it. </p>