[Ormi]

Hi there,

http://www.voyeurweb.com/contributions/view/206861/

I have sadly noticed that my voting rate is just 4,35. 3 hours ago it was on 4,62. :shock: I don't think it can happen. It has never happened before. This is my 35th contri and this has the lowest rating ever. I can not believe this would be my worst photoset. What do you think, is it cheating? :?:

Thanks,
Ormi

[rugrollers]

I've never followed scores on the free sites very carefully, but my hunch is that what you're seeing as a drop in scores compared to what you would have expected prior to the crash is actually a consequence of the systematic inflation of scores under the old management. On March 1 last year, all scores in HomeClips jumped by about .08, so our finishing score of 4.77 became a 4.85 overnight. Even before that I felt that something had been done to boost scores. I asked about it, but never got any answers.

It also seems quite possible that a bit of a boost was being given under the old system to consistent winners (like yourself) in order to keep them happy.

But the other factor is that the site has lost a lot of traffic since the crash -- they now have less than half the audience they had a year ago, and the breakaway site has taken about a quarter of that audience. In other words, this is a different audience with potentially somewhat different tastes. Maybe a significant proportion of those who were big fans of the consistent winners are no longer visiting the site.

Finally, the volatility of scores (dropping from 4.62 to 4.35 in 3 hours) is probably a function of fewer people voting and it is also possible that there was a stabilization mechanism in the old system that smoothed out big swings like that.

[Gattina1]

Hi
we are disappointed too. What Ormi said is exaclty what we think. We accept the vote but that make us feel doubtful.

First it's strange that the contributions are already published with a score, which change a lot (mostly down) within the first three / four hours. We never had a score 4.32 in 6 years of posting. We have mostly positive comments. LOL that's funny!

We think there is something very wrong and strange here.
How is possible that contris just published have a score? Who decide it?
Then when we give a top vote to Ormi, Nicole or Marilyn i.e. the score go down more and more.

There a lot of inexplicable things!!!
How is possible that there is a score oscillation of 0.50 from the first one?!

This "story" concern not just us or Ormi, but happends the same to Nicole, Marilyn and all other sexy ladies.

We need replies from VW staff asap.

Gattina & Leone

[VWSupport]

I am sorry for the decline in numbers, however this is entirely dependent on the users of the site and unfortunately out of our control. I cannot speak to the past, as I was not a part of the previous management, but I do know that currently there is no "fixing" of the numbers in any contests. The one major difference from the past that I can think of is that the scores are published in real time, as from what I heard the old voting system was a bit antiquated.

RR likely has a point with there being less users voting due to the issues we have suffered and therefore more volatility. Over time however, this will change back to normal.

[Jammer]

I remember reading something a year or two ago in the old voting system...but I'm not sure if it was pertaining to just the tit flash, or to the whole site. The score was never posted until a minimum number of votes was cast. But in addition to that, the old site used to check periodically for multiple votes from the same IP address, and if they found any, they would remove them and adjust the score. This would often result in a pic on the tit flash ranking very high when it was first scored, and then dropping quickly afterwards. The response about how the voting worked was in reply to concerns about people deliberately voting down top ranked photos. What was happening was that some people would vote multiple times (superb) for a photo that they liked, and when the pic would get scored, these votes would be counted. As soon as they were discovered and removed, the score would naturally drop. I don't know if the new VW has a similar system for dealing with multiple votes or not, but with the lower traffic that RR pointed out, the scores of high rated photos would be far more susceptible to variations of this type.

Also, I notice that Ormi's contri, "On the road" is currently ranked third, only .06 back of first place. Perhaps as RR suggested, the lower scores are due to an audience that is less apt to give high scores to anyone.

[FAQ]

I am sorry for the decline in numbers, however this is entirely dependent on the users of the site and unfortunately out of our control. I cannot speak to the past, as I was not a part of the previous management, but I do know that currently there is no "fixing" of the numbers in any contests. The one major difference from the past that I can think of is that the scores are published in real time, as from what I heard the old voting system was a bit antiquated.

RR likely has a point with there being less users voting due to the issues we have suffered and therefore more volatility. Over time however, this will change back to normal.

The old system was based upon many years of experience. No on-line voting system can be perfect, but the old-VW had something as close as I had seen it.

Downvoting is a problem. Especially with much fewer voters than in the past. It is very difficult to raise one's vote with multiple SUPERBs, but easier to lower other people's votes with a POOR. The old-VW used pattern recognition. IP addresses are too easily spoofed. Plus some ISPs, like AOL assign the same IP to multiple people. Plus I'll refer you back to when Kathy was banning all sorts of people over a generic IP address think they were all Igor.

The early days of the new-VW, the scores never moved, not even by a tiny amount.

[Gattina1]

Hi Kiki,
ok you said that score depends from viewers vote, luckly that's the true.
But we have received no reply about contris just published and their score. We are not able to understand how is possible that you publish contri with a score in! If these contris are published now, how can they have a score if people not yet started to vote?!
We noticed that contris are published with a score in of 4.50 or 4.60 or 3.90 or many other and then it change (with downvoting), but we think that everyone and every contri should start with the same level of score, i.e. 5.00 or 1.00 or 0.00; to you the choice. In this way everyone has the same probability of winning.

The explanation of the problem could be all viewers have gone crazy (but this sentence is inconsistent/lacks consistency as there are no match between the score, positive comments, quality of pics and model; and we don't think our viewers are silly people).

About vote fixed please check as after 4 hours from publishing time the score doesn't change anymore! That's not possible if vote is not fixed.

You said: Over time however, this will change back to normal. we really hope that, otherwise it will not be a honest competition.

Thank you
Leone

[MrButler]

Ormi and Gattina you both are hot as hell. I guess it will take some time until the site traffic is up to 'normal', but by contributing you both will help to (re) gain old and new frequent visitors.

Although it may take a while...in the meantime be assured: You! Are! So! Hot!
You and the other frequent contributors are one of the reasons people like me do return to VW every other day.

Thanks for all your previous and all your upcoming contris.

[rugrollers]

The wiki page about voting appears to have been vandalized and then deleted, so here it is in full, as copied on 1 March 2011. I believe it was already some years out of date by that time.

Rug Roller

*******************

Each viewer is entitled to one vote per contri. Any additional votes are filtered and ignored. Serial over-voters (a.k.a., cheats) will have all their votes discarded for the entire contest. Real time filters run to catch the cheats, and final filters are run to ensure a fair contest before announcing the winners.

The ranking is dependent on the AVERAGE vote, not the total. However, a minimum number of votes is required to qualify.

A Word on the voting mechanism from Igor

About the Voting

The voting mechanism you see at the E-Contri Superawards will be applied to all Voyeurweb and Redclouds sections soon.

What's new?

Throughout the years, we have gathered a lot of experience regarding votes: Strategic upvoting, downvoting, spoofed massvoting, etc.

Up until recently, our filters post-corrected the problems before we announced the winners. Using some new number crunching vote servers, we are now able to do most of the filtering in real time as well as to introduce new features:
Number of votes: Although the number of votes is not significant for a rank, it is a very significant parameter for "turbulence": The effect of 10 "superb votes" on a contri which has a total of 30 votes is much higher compared to the effect on a contri which has a total of 30,000 votes.

Naturally, all contris submitted early will receive more votes throughout their lifetime, and naturally all top-rated contris do receive much more votes. We have developed a mechanism which tries to compensate for this: We calculate every 5 minutes what the average number of votes is that a contri receives. Those contris which have more than twice the votes than the average number are set to "voting disabled". Those that have less than 50% than that average number are set to "Vote average 0 - more votes required" and are displayed on a dedicated overview page. {meaning the displayed average is shown as 0, while the real average is constantly being computed}

This will keep the number of votes each contri receives comparable, no matter whether it was submitted the first day or the last day. And it makes sure the impact of every vote is comparable as well.
The earlier a contri is submitted, and the higher it's ranked, logically the more vote-blackouts it will experience.

So what about massvoting, cheating, strategic voting, etc?

We know very well the patterns of strategic downvoters, and can define it by the number of votes submitted within which time frame and the type of vote. Typical example: Mrs. Jane Doe has a contri herself on rank 21. Now she starts giving every other contri between rank 1 and rank 100 a "lousy". She does not even wait until the pages fully loads, and she submits 100 "lousies" within 400 seconds to all top 100 ranks. All her 20 friends and 40 fans do the same. Does this really happen? Sure it does. Last night, eight individuals spent hours giving a total of 5212 "lousies" on the BJ superaward. Do those votes count? Nope, they don't. Our real time filters monitor every single "user session" from the moment a viewer enters the contest pages until he leaves. The typical viewer behaviour is to scan through the lists, to open 12 contris from each overview page and to vote for 4 of them, giving different votes for each. Our extreme examples last night opened every single contri, did not wait until the picture fully loaded, and voted lousy for each contri. Sure, maybe all 400 top ranked contris except for two appeared lousy to him/her....but well: If that's the case, he/she might consider finding a new playground featuring contris closer to his/her taste. Bottom line is that we say "thank you for your vote" for each vote he gives but don't count any of them, plus we eliminate the ones already counted when he started his "lousy" trip. We call it the "Keep the downvoters busy" feature.

There are many other patterns we look for, including switching IPs on the fly, very short intervals between each vote, vote-type patterns and other criteria which I obviously should not mention in public.

On top of it all, I have assigned one person from the tech crew, who does nothing else but monitor the measurements initiated by the number crunching vote servers, monitoring individual user sessions, individual vote patterns and so on.

We cannot influence the taste of our viewers. But we can try to make sure that the ranks reflect the taste of the majority of our viewers - and that's where we see our mission.

Now that high dollar awards are being offered, we try to eliminate any type of manipulation and offer our contributors the most fair voting platform you can offer on the internet. We try to make sure every contri has the same chances, no matter when it was submitted and how much time the fans are willing to spend to downvote others.

Do not get me wrong here, our "old voting" at Voyeurweb and RedClouds does most of that filtering too. But it's not real time, and that can lead to "turbulence" on the real time rank displays and sometimes to surprises when we announce the final winners after tens of millions of votes went through the end-of-the-round filters. Plus, it does not "push" contris which have not received enough votes to be ranked in a comparable and fair manner.

I hope our viewers and contributors understand and like the concept of the new vote engine and are able to to relax and go back to the basic roots and our most important principle: You enjoy and have fun - while we will take care of the rest!

Alrighty, so much about voting. And now take your cam, shoot a great picture of her fantastic booboos, or her giving you, or your neighbor, or me a world class BJ, submit the pic(s) and watch our vote engine in action...hehehehe..life is good with us here at Voyeurweb and Redclouds.

Yours,
Igor

[rockclimber]

Ormi and Gattina!

You're both 5 plus on any scale! Don't sweat it but I would suggest to Kiki they have some way to Norm the current scoring to the old scoring? It is not unlike comparing Base ball statisitcs from the old shorter seasons to the current number of games in a season. The scores can be adjusted methinks!

I gotta say I am amazed at the level of technical analysis that rating titties receives here...impressive at minimum.

I just know sexy tits when I see em and they are always 5.0/5.0....

And when they come with a cheerful smile, all the better!

Hiontach!!!!!!

:mrgreen: :mrgreen: :mrgreen: :mrgreen: :mrgreen: :mrgreen: :mrgreen:

[VWSupport]

As an update to all this:
Upon having the tech guys look at the voting records it does appear that there are some particular users who are voting 1 on every single contri they vote upon, in an attempt to bring down the vote (and perhaps to try and annoy some of our lovely contributors). The tech team is in the process of screening the users and deleting their votes so that they will no longer count against the contributor. Thanks again for bringing this to my attention!

Also, please note that all contributions start at 0.00

You will see this contribution put out within 24 hours:

rating.jpg (18.66 KiB) Viewed 10606 times

Please note the rating.

[Lucky99]

Why even have a contest if you are going to eliminate the bad votes?
If some fat girl that looks like Mickey Mouse puts up a contra we have to sign in with our user name and if we tell her she looks like Mickey Mouse we get flagged as inappropriate
I guess if you don't kiss everybody ass you don't belong here.

[Gattina1]

Hi!
we are very happy you update the voting system. You did a great job to improve that!
We also notice that now all contris are published with score 0.00 and more hours pass till the contri vote is fixed. Hoping all things will continue to improve.

Thanks
kisses

[rockclimber]

Good to see you in Vienna back up here Always 5 + in my book

[FAQ]

Kiki, get someone qualified in statistics to help. Amateur mathematicians attempts to "cleanse" the voting results in unfair rankings. Even more dangerous are folks with degrees and some math and stat background that think they know enough statistics.

Votes of "Poor" can be legitimate. Pet peeves may drive some to vote poor. There are valid statisitical methods to resolve the downvoting. You wipe out the Poor votes and people just start voting "FAIR." After a bit, you'll be only counting "Superbs and Very Goods"

Also, starting at 0.00 makes no sense. You need a certain number of votes before any number makes sense to publish.

It sounds like the new-VW is starting from scratch on the voting; whereas, the old-VW built filters over all those years of trial and error to get the fairest vote possible. Ballot stuffing and spoofing are all part of the cat and mouse game. Disenfranchising real voters is not. If the new-VW cannot install confidence in the voting, the contributors are getting screwed.

[Jammer]

Remember that we are not actually voting here, what we are doing would more accurately be described as ranking a photo or a contri. If we were truly voting, we would have one (or at least a limited number of votes) to cast for our favorite(s) each day in each section, or perhaps overall. That would be impractical to implement on a site such as VW. Other sites hold contests in which the members get to vote for their best choice after all entries have been received and the contest is closed. But that’s not going to happen here.

So what we are trying to do is to give each entry a ranking…and the entry that has the highest average ranking at the end of the contest is declared the winner. In a perfect world, that’s not a bad, or an unfair way to do it. But what is happening is that people try to find ways of unfairly affecting the scores of their entry, or the entry of others, which is what you are trying to correct for. Eliminating multiple votes from the same person for the same entry is a necessary precaution. But I’m not sure that tracking a person’s voting habits is a good idea. I understand and agree with your rational for this, but I am concerned about the method. Removing all the votes of a person that only votes poor isn’t going to make much of a difference over all anyway, as long as they only voted once per contri. Would this also apply to someone like me that only casts superb votes? Would all of my votes be removed also? Would you remove all of the votes I cast if I only cast poor votes if I only vote for two or three entries? If I were to vote poor for every entry, removing all my votes would only affect the scores…and affect them all equally…but have no affect on the overall ranking. What would happen if I voted poor for 8 out of the 10 entries on that day…and voted superb for the other two? To me, this would be a greater indication of trying to fix the scores. I think any attempt to put a system in place to remove votes based on voting records is going to be flawed. Do everything possible to prevent multiple voting, but let the viewers score each entry the way they want to.

Any concerns that we have about people submitting entries late in the contest period could be eliminated by simply closing the contest to entries at the deadline, but leaving the voting open for another week or two. That would give most viewers a chance to cast their votes, and to allow the late comers to accumulate enough votes to qualify.

One other thing you must do to keep the contests and the voting fair is to actively eliminate the fake and stolen entries. Currently there is no easy way to report suspected entries, short of sending you a PM or email. You should also either remove an entry that is proven to be published elsewhere, or at least disable the score and mark it clearly as 'in-eligible' for prizes.

[VWSupport]

To clarify a bit more on this, we did not go about removing every single "poor" vote, everyone is absolutely entitled to their opinion on scoring and we are not looking to change that. However, I was told by the tech team when they looked into this that there was some clear cheating going on from certain users. I agree that the scoring system does still need some tweaking and is not perfect yet. There has been a lot of discussion and suggestions about this on the Explicit BBs and much of it will be implemented.

We have removed several contris from the contests due to stolen photos or content that was clearly not amateur, and there will soon be an option to report the contris with a button.

We will likely be changing the contest structure to not allow any more new contris to be submitted four days or so before the end of the contest (voting will continue during this time of course). I believe this will start in October's contest.

[FAQ]

As much as you and I do not like it, it is not cheating to mark your contri SUPERB and all others POOR. One vote per contri per person, regardless of the vote. Yes, even the fan base for a girl can vote hers SUPERB and all others POOR. That is still one vote per contri per person. To eliminate that one vote per person is dishonest and taints any results VW posts.

VW has already shown they need help with IP addresses. That is not enough to know if someone is voting multiple times. AOL has their users appearing on the same limited set of IP addresses. Let's not even bring up the banning on this BB that someone at VW thought nearly everyone was Igor. Nope. Don't even mention that sorry excuse for technical aptitude.

It takes highly sophisticated methods to decide whether to exclude a vote or not. So far the new-VW has not given us any confidence that they have the ability to determine legit voting from the rest.

[rockclimber]

I do like the idea of closing the new submissions a few days prior to the end of voting.

How about adding best first time contri as a category? Some people win repatedly and excellence should be rewarded but newbies need some TLC too so they too will become repeat entrants.

Maybe even top two newbies? First three contri are eligiable? So our new contributors will have three shots at this category of win?

Just thinking.... Thats a bit more cash to shell out but it should generate a lot of entries.

Jut a thought......


:mrgreen: :mrgreen: :mrgreen: :mrgreen: :mrgreen: