Announcement

Collapse
No announcement yet.

Rating Models

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

    Rating Models

    Hi

    Potentially, rating models is a fun feature, and an easy way to acquire a little information on customer preferences. Now, I don't like to dwell on the negative, I know you're not unaware of the software's total crapness, and I really do appreciate the time and effort that's been put into this, but, well, here are my thoughts anyway.

    IMO, the killer flaw is the randomness; on photo voting sites, visitors aren't looking for images of anyone in particular, so this isn't really a big problem. But here, if I visit Cass's homepage, and find her to be staggeringly beautiful (or, inconceivably, perhaps merely extremely pretty), I can't indicate my approval (or look up her current rating) without a considerable degree of off-putting hassle.

    Should I wish to rate a particular model, and assuming there are one hundred possible models to vote for, I must skip past almost 70 other images in order to have a 50% chance of Cass coming up at least once. One in five users will fail to find the model they're looking for, even after having clicked through 160 times.

    I suspect that models added to the list in the future will pick up very few votes.

    Ideally, it'd be possible to rate a model from her homepage, and rating a model would be properly integrated into the site. Or even better, people would just post their comments on this forum. img tag disable, use attachment mgr/system to add images to postshttp://www.abbywinters.com/discus/clipart/happy.gif

    #2
    PS

    Have fun with binominal probabilities here: http://faculty.vassar.edu/lowry/binom_stats.html

    Comment


      #3
      > 50% chance of Cass coming up at least once.

      Oops, this should read: 50% chance of that model coming up at least once.

      Just a little anecdotel observation, I must have clicked through almost 400 times. Susie has come up only once.

      Comment


        #4
        What Muggins said.

        I was wondering, by the way, if it wouldn't be possible to #INCLUDE the survey buttons on each model intro page. That seems more natural for, oh, about a thousand reasons. First being that we can look at her pics and then decide. Second being we can easily find her.

        Extra credit, by the way, for pulling your regular survey trick and not permitting overvoting.

        Even more credit if you permit re-voting without screwing up the averages. It seems like a bit much to ask but I've noticed I sometimes get a totally different appreciation after a couple of times through a set.

        For instance in the Emma sets she seemed a little gaunt and stressed out. I would have given her a two or three. It took me a while to recognize she was pretty hot and I'd have bumped her up to seven or eight. Same with Lex who, based on her sample image I'd have given a one or two but I now think is one of my favorites.

        I know first impressions are key in porn, and chances are 99 of 103 viewers aren't going to vote in the first place, let alone go back and vote again, but I'd love the ability to say how I really felt. (Oh yeah, figleaf, and Abby will ask Bill Gates and Larry Ellison write all that code for you next time she has them over for tea. J)

        figleaf

        Comment


          #5
          Amen to what TFL said.
          I stumbled across the rating link and went clicking through the models to find the 2 I wanted to rate... One came up only once while the other had passed twice.

          reccomendation:
          Ditch the random number generator thingy and add rating buttons to each model's page.

          or
          make a Ratings page that has a thumbnail of EVERY model and a rating button to press.

          Im sure volunteers could be found to write the page(s) if need be img tag disable, use attachment mgr/system to add images to postshttp://www.abbywinters.com/discus/clipart/happy.gif

          Comment


            #6
            I think that the rating thing is a good idea. I agree that it makes sense to take the randomness out of the mix, and/or to put a button on each model's home page.

            One thing that kindof threw me was that I was really rating the model, though the message seemed to be "rate the pic". Two pix of Jill (one of my all time faves) came up. I was trying to rate the pic, but I was also unconsciously rating the model at the same time. I know this sounds confusing - it's because I am confused.

            I agree with Figleaf - how I feel about a model overall evolves as I go through her sets and vids.

            Abby, this could be a great tool to do two things, rate the interest/acceptance of models types, and also get the same information about the quality/acceptance of different types of images. Of course, for the life of me, I don't know how you'd manage it... img tag disable, use attachment mgr/system to add images to postshttp://www.abbywinters.com/discus/clipart/happy.gif

            I like this new thing. Can you tell? Don't you wish I'd just shut the hell up???

            img tag disable, use attachment mgr/system to add images to postshttp://www.abbywinters.com/discus/clipart/happy.gif

            SlipJig

            Comment


              #7
              Slipjig, are you talking about rating the beauty of the model vs. content of the set?
              Like a) she's not very pretty and b) I love her pee pics?
              Or a) She's 10 of 10 in beauty but b) I don't like that there are no spreads?

              Or when you talked about "quality/acceptance of different types of images" did you have something else in mind?

              Comment


                #8
                Hi all

                I agree that having the models appear in alphabetical order, or, only be able to vote for models from their bio page, would be Nice, but it's useless from a stats POV. Only new and popular models would be voted for, older models would get way less votes, and "ugly" ones would have no chance of getting any.

                The whole point of the ratings is so I can get a better idea of what kind of model is popular, and answer questions like:
                1. are blonde models more popular than non blonde models?
                2. are models who do more explicit stuff more popular than those who do nude only, regardless of how they look?
                3. Hi mum
                4. are new models more popular than old models?
                5. are models shot on locations more popular than those that are not?
                6. are large breasts more popular?
                7. is pale skin more popular?
                8. are models we paid more (for whatever reason) worth it?
                9. and so on.
                From those stats, I can apply what we learn to the selection of new models, and models who we might have previously decide to be borderline, we can know that models with those attribs are generally popular, and so are (or are not) worth shooting.

                At the moment we have a hazy idea of what the "ideal" model is, and of course there are some deviations from the path that we suspect will sell well (or, at least SOME of the attribs she'd have).

                About 30% of people who show for interviews I can tell if they will be good for the site instantly, and another 30% that I know will not be (heh, equally instantly). But that leaves 40% of applicants that are borderline for whatever reason, and the results from this survey will help us decide to shoot or not to shoot.

                But we (Elvaar, Cass, VidDude and I) are constantly amazed about how badly models we suspect will be awesomely popular fare (Petra, for example)... and equally surprised by models we thought would do badly (and regret shooting) who are outrageously popular, like Felicity.

                We had some very grand plans for this model rating application, which would log who voted for what (based on username), allow people to only see models they had not voted for, allow ppl to go back and change their votes sometimes, heaps more, collate stats, draw graphs, and so on. It was gonna take several months to write, cost a lot, and maybe not be well received by members.

                Which, in hindsight, is dead simple to predict, of course.

                Regardless, we went with this crap version (that ended up taking just as long to adapt to our needs...), and realised too late that the kinda data it can give us is average to non-existent in quality, and no one seems to like it much anyway.

                Not to put too fine a point on it, but I am not one WHIT interested in what you want to vote for one particular model, cos you really like / dislike her, UNLESS you wanna vote for a bunch of other models (at least 20?) as well. If you like a model that much, send me an email, like everyone else.

                So, sure, with that attitude, I'll reduce the number of people who will respond to the survey, but OTOH, those that DO respond, will presumably respond in a useful way (it's just as easy to vote "correctly", as not to, and most people are kinda nice).

                And for the record, I struggled long and hard with how much info I should include on the ratings page (where you start the ratings sojourn), weather to base your vote on just the pic displayed, how explicit she posed, if she had posted to the boards, but people are gonna vote based on these things (any many others), that I have no control over, and anyway, swings and roundabouts, it'll average out in the end, right?

                And some people give every model at least a nine for showing up (regardless of what she looks like), which is always going to affect the averages, same way grumpy people will give everyone a two, unless they are amazing, and prolly 80% of people will vote a 4, 5, or 6 for 90% of the models.

                *sigh* maybe I should just give up? ;)

                a

                Comment


                  #9
                  PS, with regards to the type of image displayed, we opted to go with what was easy (we'd already spent so much damn time on this project), the pic from each model's bio page, already conveniently located in a dir.

                  a

                  Comment


                    #10
                    Just to throw some gasoline (petrol?) on the fire, I found this link in a long list of amihotornot clones:
                    Face Wars

                    Here, instead of rating one model, you're selecting your favorite of two models.
                    It's called forced choice voting.

                    BTW, here are my two favorites from that long list:
                    Flash Your Rack
                    List of AmIHotOrNot Images

                    Comment


                      #11
                      Hi

                      I hope your headache is better. img tag disable, use attachment mgr/system to add images to postshttp://www.abbywinters.com/discus/clipart/happy.gif

                      Okay, so the model rating thing is just this week's survey, not a new site feature. My mistake.

                      "We had some very grand plans for this model rating application, which would log who voted for what (based on username), allow people to only see models they had not voted for, allow ppl to go back and change their votes sometimes, heaps more, collate stats, draw graphs, and so on. It was gonna take several months to write, cost a lot, and maybe not be well received by members."

                      That sounds pretty great to me, though I understand it probably not worth going to all that effort if it's only me and TFL who bother to use it.

                      > Not to put too fine a point on it, but I am not one WHIT interested in what
                      > you want to vote for one particular model, cos you really like / dislike her

                      That's fine. Just as long as you don't plan on adding more models to the survey. If so, people have no choice but to look for a particular model, and no one's going to bother if they have to click past 50+ odd other models to find her. But again, that was assuming this was an ongoing thing, not just a one off survey.

                      I kinda hope this wasn't the fun surprise that was mentioned a little while ago . . .

                      Take it easy,

                      Muggins

                      Comment


                        #12
                        > I kinda hope this wasn't the fun surprise that was mentioned a little while ago . . .

                        Reading it back, that sounds a little mean. I'm sorry, that was never my intention. (I only intended to maybe express a sense of mild disappointed. but don't worry, I see that there's another Kylie update tomorrow, so I'll probably soon get over it.)

                        thanks
                        Muggins img tag disable, use attachment mgr/system to add images to postshttp://www.abbywinters.com/discus/clipart/happy.gif

                        Comment


                          #13
                          >Not to put too fine a point on it, but I am not one WHIT interested in what you want to vote for one particular model, cos you really like / dislike her, UNLESS you wanna vote for a bunch of other models (at least 20?) as well.

                          I guess I qualify. I must have clicked at least 400 times until I felt that I had voted for every single model img tag disable, use attachment mgr/system to add images to postshttp://www.abbywinters.com/discus/clipart/happy.gif I'm sure Uma came up at least 75 times, LOL. Many others showed up only once.

                          Despite all that, I really enjoyed the voting. I gave a wide range of scores, ranging from a low of 2, to a high of 10 for several girls. The majority fell between 6 and 8, it's true, but at least you should be able to tell which girls are most beloved, and which are extremely disliked (very few).

                          Don't give up Abby. This was an enjoyable new feature. I hope we haven't seen the last of it.

                          Comment


                            #14
                            "At the moment we have a hazy idea of what the "ideal" model is,..."

                            Can there really be an "ideal" Model?
                            It seems to me that there is a wide variety of tastes from the members, which is good.

                            For example, Stacy is probably the closest physically to what I tend to like the most, yet I still like Danny even more.
                            I bet neither is the most popular model here.

                            Then there's Cass, who's not only gorgeous, a talented singer, but she really makes the forums a lot more fun to read and participate.

                            Variety is the spice of life.
                            It think this is one of your sites greatest assets, not just a bunch of cookie cutter models.

                            Personally, I really appreciate your efforts to find out what the members want, but still want to emphasize the variety that your site continues to provide.

                            Redstar

                            Comment


                              #15
                              facial sexiness ='d 55 to 65% of my vote. not necessarily 'good' looks...just 'sexy' looks

                              most of ur models who have scratches, scars, scabs, dings, bruises, hairy backs, and so on turn my head more than most of the polished chicks

                              a canted smile, some imperfect teeth, maybe slightly messed hair...maybe WICKEDLY-messed hair

                              ur lighting, color, balance, clarity counted (photo size, too); the locale, the clothing and all that stuff matters, but quite a bit less than the face

                              imho img tag disable, use attachment mgr/system to add images to postshttp://www.abbywinters.com/discus/clipart/happy.gif

                              Comment

                              Subscribe to our e-mail newsletter

                               
                              Sign up for the abby newsletter. Don't worry, we'll NEVER share your email address with anyone.
                              Working...
                              X