This gives a more accurate ranking (assuming we are weighting looks and performance equally) than the current Top 100. It's similar to how IMBD might rank movies. It makes it fair between somebody with 200 reviews and somebody with 5.

Personally, I would not weight Looks/Performance equally but I figured it's a good start.

Top Three Movers: 1. Paige +59 spots, 2. Courtney +44 spots, 3. Sammie Jo +40 spots. Won't be invidious and mention people dropping. Everyone on the list is amazing. Just having fun with the numbers.

I will post a ranking just for performance and just for looks later...



New Rank Name Old Rank
1 Summershowers 1
2 Anya 3
3 Brooke and Madison 7
4 Meena 5
5 Evangeline LaCroix / Leina 2
6 Khori / Shelby 4
7 Jennifer 11
8 Morgan Marie 6
9 Madison O'Hare 20
10 Nina 8
11 Meagan 14
12 Brittany Cherry 9
13 Ellie / Erin Marxxx 18
14 Kellie 13
15 Dakota Koxx 21
16 Shelby 39
17 Cynamon 19
18 Gianna 23
19 Sara 22
20 Naomi Monroe 57
21 Jordan 24
22 Courtney 66
23 Paige 82
24 Gemma / Lauren / Gemma De Rossi 31
25 Alana / Natalie 26
26 Boston (Current) 15
27 Ashley James and Dakota 10
28 Ashley James 17
29 London 25
30 Ericka 50
31 Katie 32
32 Marie 27
33 Mary Jane 28
34 Nadine 30
35 Serena 62
36 Michaela 41
37 Ilana 16
38 Brooke 51
39 Sammie Jo 79
40 Rylie / Katie Kayne / Cheyenne 37
41 Sydney 55
42 Sophie 33
43 Celine / Gabriella 36
44 Larissa 34
45 Ryder 76
46 Suzanne 77
47 Ansley Ayers 54
48 Tylor Blake 44
49 Katie Lynn 70
50 Sianna 47
51 Alana Moore 69
52 Aren 68
53 Kyla 12
54 Faith 38
55 Sole' Rivera 40
56 Sapphire 89
57 Jill 29
58 Kendall 46
59 Kennedy Amir 90
60 Erica (New) 52
61 Genna 42
62 Heaven / Arielle 43
63 Stacy Davis 53
64 Nikki / Kenya 58
65 Aleece 74
66 Sasha Sole / Mya / Vienna Levy 49
67 Katy / Katie / Kathryn 67
68 Alicia 71
69 Ashley Sparks 59
70 Dayna 45
71 Alexandrea 72
72 Natalia 95
73 Alyson / Adrianna 80
74 Charlie 60
75 Kai 56
76 Angelica 61
77 TS Bella 48
78 Mallory 92
79 Aviva 35
80 Cate 78
81 Zoey 63
82 Heather 83
83 TS Katie / TS Amy 86
84 Alyssa 99
85 Alivia Leveauxxx 85
86 Haley Miller / Heidi Mueller 94
87 Katie 73
88 Ashtyn 64
89 Anastasia 65
90 Alexis 75
91 Richelle 91
92 Heidi 88
93 Casey 100
94 Payton 81
95 Aura Light / A.L / A. Mira L. 87
96 Khloe 96
97 Christy 98
98 Remi Martin 93
99 Athena 97
100 Petra Leon 84

Can't we omit  TS from top 100?Dont they have their own board or something, if not....
Lets create alternative lifestyle board.
Congrats to all the LADIES!

This is kinda of like college football's BCS, there will always be strong opinions for our favorites.  Unless of course we could figure out a provider playoff system, lol.

But it does sound quite impressive. Thought Bayesian used to calculate conditional probability and its inverse.

You should consider writing a book on How to Bayesian to Calculate Average. Or "is this an improper application of perfectly good mathematical theory"?

Wiki link is below with examples. Let me know if I missed something on Stat classes. Always looking to learn.

1. Widmerpool70
.
.
.
.100. Widmerpool70

He's childish and immature in that way.

You should also apply a stochastical regression to your analysis. Doing so will give greater weight to the ratings that are more recent than those that were given further in the past.

As it stands now, you're giving equal weight to all of a provider's reviews, whether they were given yesterday or 5 years ago.

And as we all know, people can change over time.

also eliminate:

1. grades and class rankings in school
2. music Billboard's Top rankings
3. The PGA Fedex Cup
4. Restaurant guides like Michelin and Zagat's
5. US News and Report's Top University Rankings

There's no denying there's pressure when ranking and competition are concerned. Just do your best and let the chips fall where they may.

plus way too much time on your hands

I have mixed feelings about the top 100 List. While I have always appreciated that the review system is a great asset and tool for both hobbiest and providers both, it also creates alot of pressure for providers as well. I myself strive not to be a robot on any date, with that the reality for me(just speaking for myself here) is that I cannot be a 10 every day....possibly I am having an off day, possibly I am not as attracted to a particular gentlemen.Nonetheless, I always try to be professional ,upbeat and kind. We have seen on this particular board woman nearly break down after receiving their first bad review.... I think its a shame we as providers put so much burden on ourselves. I will say that some days I have noticed that I have a new review that I was not expecting and I find my heart jumping out of my chest as I click on to view "the numbers".... I only wish I did not care so much, but the reality is that I do.

I looked at the variety of list provided on this particular thread. I think alot of things come into play here. I know some great providers that were not included on the list. I wonder how that might make them feel. I know for myself I look at my reviews and my most recent two pages are mostly 9's and 10's in performance however on my last page of reviews from 2006 when I first entered the hobby I scored lower which has affected my overrall average. It took me time to find my "niche"as a provider. The provider I was in 2006 is not the provider I am today.... I had to find what worked for me,how many clients to see a week, advertising to reach the clients that I enjoyed ect.

Maybe I am rambling a bit and maybe I am a bit off subject, but I wanted to share my feelings.

-- Modified on 8/18/2010 12:45:34 PM

Shilling for who???

If a thread makes u hyperventilate you should probably just move on.

Amazing the general antagonism on this board.

I completely agree.  While reviews are essential in finding quality companions w/o being disappointed, it puts much pressure on us as providers.  I think we as women/providers compare ourselves to other women when we all are wonderful and unique in our own way.  I know that every session isn't going to be perfect.  But we are not robots.  As long as the clients who see me, enjoy my company, it doesn't matter if I am ranked #1 or #1000.  I try very hard to make sure that every session I have is wonderful.

You are right though, we all have our "off" days.  I'd rather be human than be an emotionless robot.  When I see I have a new review, my heart starts pounding as well because I understand that gentlemen come to see me most of the time based on my reviews.  There is a lot of pressure to keep that up with every client even if you and the client don't click.  Its bound to happen at some point.

For the record, I think all these girls should be at #1.  We are all special, beautiful, caring and make sure that clients enjoy their time with us.  No one girl is better than any other.  It is only in the eyes of whoever is reviewing a girl to put his one opinion in and thats good but that shouldn't make or break any of us because at some point we all get a bad or bogus review and it sucks.

So, on my list...Tylor we are all #1!

XOXO

Gemma

I agree. I ultimately base a decision on instinct or review details.

Probably everybody in this list is great. But numbers do help the cream rise to the top.

No different than Zagats. My fave restaurants are not their highest rated ones.

And, again, I didn't create the original list. Just played with the numbers. Women and meaningless stats...

I am a a total asshole loser for posting this but here's the Top 5 Looks-only using weighted averages:

1. Jennifer
2. Ellie / Erin Marxxx
3. Ashley James
4. Summershowers
5. Khori / Shelby

Again, this is based on the TER Top 100 ATL so everyone is great to even be on this list

No, I don't think this is definitive. Yes, everyone has different tastes. Still, Jennifer is UNEQUIVOCALLY the best looking woman in Atlanta.

"It's science" - Ron Burgundy

Yes, I have the 5 min of time on my hands to run shit through Excel. Fantasy Football is much more tedious except there's no women...

There is no such thing Bayesian Averaging. Bayes is used as a short form for Bayesian. We got those two things out of the way.

In more technical terms, the theorem expresses the posterior probability (i.e. after evidence E is observed) of a hypothesis H in terms of the prior probabilities of H and E, and the probability of E given H. It implies that evidence has a stronger confirming effect if it was more unlikely before being observed. Bayes theorem is valid in all common interpretations of probability, and is applicable in science and engineering.  However, there is disagreement between frequentist, Bayesian, subjective and objective statisticians with regard to the proper implementation and extent of Bayes' theorem.

Bayesian inference uses Bayes' formula for conditional probability:

There is no such thing Bayesian Averaging. Bayes is used as a short form for Bayesian. We got those two things out of the way.

In more technical terms, the theorem expresses the posterior probability (i.e. after evidence E is observed) of a hypothesis H in terms of the prior probabilities of H and E, and the probability of E given H. It implies that evidence has a stronger confirming effect if it was more unlikely before being observed. Bayes theorem is valid in all common interpretations of probability, and is applicable in science and engineering.  However, there is disagreement between frequentist, Bayesian, subjective and objective statisticians with regard to the proper implementation and extent of Bayes' theorem.

Bayesian inference uses Bayes' formula for conditional probability:

P(H\D) = P(D\H) P(H) /P (D)

where
H is a hypothesis, and D is the data.
P(H) is the prior probability of H: the probability that H is correct before the data D was seen.
P(D | H) is the conditional probability of seeing the data D given that the hypothesis H is true. P(D | H) is called the likelihood.
P(D) is the marginal probability of D.
P(H | D) is the posterior probability: the probability that the hypothesis is true, given the data and the previous state of belief about the hypothesis.
P(D) is the prior probability of witnessing the data D under all possible hypotheses
where
H is a hypothesis, and D is the data.
P(H) is the prior probability of H: the probability that H is correct before the data D was seen.
P(D | H) is the conditional probability of seeing the data D given that the hypothesis H is true. P(D | H) is called the likelihood.
P(D) is the marginal probability of D.
P(H | D) is the posterior probability: the probability that the hypothesis is true, given the data and the previous state of belief about the hypothesis.
P(D) is the prior probability of witnessing the data D under all possible hypotheses

Bayesian is never used in Population Statistics to calculate Arithmetic "mean, mode and median". Bayes is primarily used science and engineering especially in information theory, Mante Carlo Simaulation, Bayesian Filters, Data Analysis, Estimation and classifier, etc.

Moreover Bayesian networks for calculating bayesian probability is a very complicated process which requires building Bayesian Networks and algorithms. One primary use of Bayesian Classifier and information filtering today is spam filters.

Using Bayes for calculating simple arithmetic mean gives incorrect results from which nothing can be inferred and blows bayesian inference.

All I am saying is Bayes is not the correct statistical method to calculate averages and predicting provider performance. In this case, "mode" would more accurate prediction than mean and median. Of course median can be used with some caveats for predicting something else such, after how many trysts a providers performance decrease.

Not busting your balls but just sharing info.

I can't help that there is a model called Bayesian Averaging.

Is it what they use for vote weighting. Otherwise the top restaurant in Zagats could could be one with a single review.

I know Bayes theorem. It's not that but it's certainly based on conditionals. It's absolutely Bayesian since the first step is to look at what we already know.

I'm sure the Reverend Bayes would love seeing this applied to escorts.

I don't have access to ratings outside of the top 100. And that changes daily.

I can PM you your weighted rank. I sure hope this is turning you on for all the grief I've taken....

Because the numbers used for the ratings system are fluid anyway. A guy gives a 9 for looks and TER changes it to 8, and so on. They should change the whole ratings system to:

On a scale from 1-5 with 1 being least likely to repeat and 5 being most likely to repeat.

Also, sometimes a reviewer gives numbers completely out of line with his review.

I can't imagine anybody who uses the Top 100 as the basis for a search. Most people have something in mind and are not nitpicking 8 v 9.

More Scenarios please

Can you look at rankings of agency vs. independent and weight the recent scores more?  

Also, can you do a value ranking - not sure how to do it (need a GT grad to help out) but looking at relationship between $$ and performance...

How long did it take?

Look at the example. Bayes can reduce a single anomalous value but it gives you poor results when there a multiple values.

So, unless you know all the anomalous scores, which is impossible in this case, the rankings are anomalous at best.

Not trying to argue with you but tired of people using inappropriate mathematical formulas to create absurd validation. Something like the risk models used before the financial crisis.

If your are interested in this kind of subjects, do read "Black Swan". Very interesting book.

you've listed some mighty fine ladies there.  You may have missed a couple of the UTR ladies but that list is pretty good.

Dude, I'm using a well-known formula to weight votes. IMDB is basically the same thing.

The idea that it is the wrong formula is just not true. I'm not using Bayes Theorem. Repeat: I'm not using Bayes Theorem.

It's similar to adding 30 median scores (e.g. 8.3) to every provider. Then a Provider with few reviews is going to be pulled pretty close to that number. A provider with 300 reviews is not going to be as affected as much by adding this 30 votes. IMDB does exactly the same thing so a movie with 2 reviews is not rated the greatest movie of all time.

I don't doubt somebody could create an even more sophisticated system. But I don't see how this is wrong in any way. It's better than the TER Top 100. For example, a Provider just jumped into the TER Top 20 on her 5th review. That's what I'm trying to address.

Black Swan would be a good provider name. Yes, I read the book. I like it though he tends to ramble.

I think I remember reading a study which used TER to correlate $$$ to performance.

I don't think it was that interesting. The high $$$ people rated highest.

Just asking!

sounds logical, but there are some great independent films that very few have seen

Bayes Theorem specifically deals with determining the probability of an event that's already occurred given subsequent related events.

Bayesian inference is broadly concerned with continually updating probabilities based on new data. In my crude formula, the rating of Provider x would continually be updated not just by her ratings but by those of all other providers. So it's Bayesian in the sense that it's using new info (rating of provider Z) to get at something which already exists (ie how good provider x is).

It would actually be more useful across the whole db, not just top 100.

GO Courtney!!!

Very familiar inference and Bayesian Networks. You are describing one application. Another well known application is posterior probability, i.e., predicting future behavior based past observations which is what is used in spam filtering and virus detection.

It is also used in information warfare and for target acquisition. Interesting and very complicated algorithms.