The "Quality Score" is based on a curator's behavior on SubmitHub, and is not related to engagement.
The major change is that points are now considered on a linear scale rather than in steps. This should result in smoother day-to-day scores and less jumping around.
Here's a completely made-up example of how the previous system worked:
Someone with 4.9 days got +2 points, but if they dropped to 5.1 days they'd only get +1 point. Under new system there's a linear sliding scale, so they'd get +1.95 points.
In the name of keeping things transparent, we've decided to make the calculations of the quality score public 🔍
—————————
Methodology: Each curator starts at 50 points and earns or loses based on the below metrics. Their score is then compared to the max score. See more in March 2023 update below.
Note: There are a handful of additional metrics that are specific to a few niche cases and will not be shared publicly.
Historically, the quality scores have ranged from 5/10 to 10/10. As of today, they now go as low as 1/10. What this means is that someone historically at 8/10 has probably moved to 6/10, because the range has "widened". The underlying drivers of that score likely remain unchanged.
This change was brought about when I noticed that a curator with really negative signals (bad feedback, not sharing often) came back with a 8/10 score.
Old formula
Everyone starts at 50 points. Points are added or deducted based on the factors listed below. We then take the highest score, and everyone else's "rank" becomes their score divided by that higher one.
For example, if the highest score is 60, and yours is 55, your rank becomes 55/60 = 92% (or 9/10).
New formula
The new formula is similar to the old one, except that the divider has changed. Turns out the "minimum" score is typically around 30 points. Under the old scenario, you'd a score of 30/60 = 50% = 5/10. This doesn't seem like an indicative score for the literal lowest on the site.
The following part might get confusing, but works well in my brain 🤣
First, the new formula finds the minimum and maximum scores. It then looks at the difference, and uses this as the divider.
Example
Let's assume our minimum is 30 and our maximum is 60. And that a curator has 55 points.
This new approach gives more variation to the potential quality scores, and gives curators more control over the ups and downs.
Some of the factors included in this measure include (but are not limited to):
In total, there are more than 25 factors taken into account when generating a "quality" score.
If you're a curator curious about how to improve, you'll see an option to view tips just below your score.
Here's where curators end up in the mix.