Anyone who has used the EI review system knows to be skeptical of one time reviewers, or more generally, of reviewers with few reviews. Escorts who review themselves, or get others to write them fake reviews, generally can't be bothered to write a series of reviews of other escorts from the same account for cover. Pimps may use a single account to review multiple escorts, but probably will not use the same account for too many, to avoid suspicion. So it's fair to suppose that most established reviewers are honest, but many one time reviewers are not. How many? I don't know, but I can make a pretty good guess.

At the time I made the following analysis there were 13212 reviews by 4175 reviewers on the system. Of these, 369 had at least 8 reviews. Some fake reviewers may have at least 8 reviews, but probably not many. Let's assume for this analysis that those reviewers are all real. Of those, 316 (86%) wrote at least 9 reviews. Of that number, 273 (again 86%) have at least 10 reviews. Of those, 234 (86% yet again) have at least 11 reviews. You can continue the series further and you find that about 86% of reviewers write a further review at each stage, until the number of reviewers gets so small that the percentage stops being very meaningful.

So what? Suppose we take it as a law of nature that about 86% of genuine reviewers write another review and we assume that those 369 reviewers who have at least 8 reviews are all genuine. Then it's Junior Cert Mathematics to work backwards and calculate how many genuine reviewers there are with at least 7 reviews, at least 6 reviews, etc. until we get to 1 review. Do this, it seems there should be about 1042 reviewers instead of 4175. In other words the vast majority of reviewers are probably fake. It is also possible to work out, from the same assumptions, the number of fake reviews, which turns out to be about 40% of the total. So most reviews are genuine.

You can criticise the assumptions above. I don't believe them 100% myself. Changing them a bit doesn't really make much difference to the final numbers, however. According to this analysis, the number of fake reviewers is about 3 in 4 and the the number of fake reviews is about 2 in 5. If I had to guess, I would expect that the true numbers are a bit lower, but not too much lower.

What can you do with this information? For one thing, you can be a bit more confident in your intution not to trust reviewers with few reviews. You can also make some judgments about whether an escort has fake reviews, based on how many of her (or his, if you are into that) reviewers have few reviews. Before you do that, there are a few things to keep in mind:

- If an escort has few reviews then there isn't too much you can conclude. The effects of chance are just too strong.
- There may be legitimate reasons for an escort to attract a lot of first time reviewers. For example, there are often threads asking for suggestions for a first punt. If people are taking these seriously then the escorts suggested will have an unusually large number of first time reviewers.
- Not all fake reviews are written by the escort.

Now that the warnings are out of the way, how do you judge whether an escort has a suspiciously large number or]f reviewers with only one review? For the average escort, about 15% of all reviews are by one time reviews. For the average honest escort the numbers will be lower. How much lower? Probably the most reliable way to guess is to look at specific escorts. Alyssa is one of the escorts most people here would trust. Out of 70 reviews, she has 6 (9%) from first time reviewers. Lynne has 9 out of 67, but two of those are reviews that she has flagged as fake. Excluding those, she has 11%. Eve has just 1 out of 41 (2%)! There is some variation in the numbers, but all are below the 15% the average escort has. They aren't much below 15%, so it's hard to identify honest escorts outside the top 30 reviewed just by counting first time reviewers.

For interest I looked at the top 5 escorts in the running for Escort of the Month:

English Taylor 6/26 (23%)
Karin 9/47 (19%)
Nikky 5/43 (11%)
Rubina 6/42 (14%)
Patty Baby 13/85 (15%)

Now I don't think the statistical evidence is good enough to accuse any particular one of these escorts of having fake reviews, but I think it's nearly certain that at least one does. In fact, Patty Baby is known to have fake reviews, although there was never any evidence that she participated in this. Those aren't counted above, because the individual who wrote them also wrote reviews for other escorts. That shows the dangers of relying only on counting one time reviewers to detect fake reviews.

If this information helps punters to avoid dishonest escorts then I will be happy. If it helps honest escorts to compete then I will be even happier.