We strive to constantly work on maintaining the system to keep it up to date and avoid random or non-argumentative reviews, like rage reviews for instance. Protecting Revain scoring system from all sorts of manipulations is essential.
Revain rating system is designed to assign a certain score to the project depending on the number of reviews and the average score the project is able to accumulate over a certain period of time. Generally, score-based ranking systems serve well in reflecting an accurate representation of the project’s reputation on the platform. However, while monitoring its operation, we noticed some minor pitfalls that could be fixed only by introducing a new rating algorithm.
Now let's dive a bit more into details on how this algorithm works.
Our newly introduced rating formula puts more emphasis on the rating. Obviously, the amount of reviews is still important. For example, a project that has only one review with an overall impression of 5.0 will be rated slightly lower than a project with 10 reviews and an average overall impression of 4.0. A logic behind it is that one 5-star review can technically be 'random'. Therefore, ten "4 points" reviews are more credible than one "5 points" review and thus are a better mean of assessing project credibility.
Another example. A project with an average score of 4.5 that has 100 reviews will be rated higher than a project with an average score of 5.0 but only 15 reviews.
The score is calculated by the following formula:
Score = 100 * (2.9999 + (avg - 2.9999) * (0.45 + 0.2031*ln(n)))
n – number of reviews
avg – the average 'overall impression'
Ultimately, the re-calculation allows to ensure that Revain ratings are 100% fair and transparent, this much harder to compromise. A new formula allows keeping the major focus on the average score as it is the prime multiplier in the score equation. See how it works for yourself.