There can be fluctuations in Bitsight Security Ratings, which coincides with the daily shifts in:
- The number of new observations.
- Adjustments when events fully decay or when records complete their lifetime and no longer impact the rating.
The displayed rating changes in 10-point increments. However, the underlying rating is computed on a finer scale.
A rounding off method is implemented to prevent small changes in the underlying rating from impacting the displayed rating, which depends on previous values of the rating. There are different criteria for when the 10-point drops versus when the rating rises by 10 points.
Rounding Method Examples
For an entity with the actual rating of 796 currently rounded up to an 800, changes in the rating can lead to the following scenarios. Note that the examples provided are representative of changes that would happen for other specific ratings (e.g., the 796 example used could have been a 696 or a 306, with the rounded ratings changing accordingly)
Actual: The actual
796
rating drops to between≥785
and<795
.Displayed: The rating is presented as a drop from
800
to790
.
Actual: The actual
796
rating drops to between775
and784
.Displayed: The rating is presented with a rating drop to
780
.
Actual: The actual
796
rating drops to between765
and774
.Displayed: The rating is presented with a rating drop to
770
.
Actual: The actual
796
rating increases to between≥800
and<805
.Displayed: The displayed rating remains
800
.
Actual: The actual
796
rating increases to between≥805
and<815
.Displayed: The rating is presented with a rating increase to
810
.
When the combined risk vectors are calculated into the rating, the subtle differences of individual risk vectors may increase or decrease the overall security rating with no visible changes to their letter grades.
- January 23, 2025: Corrected the first rounding method example.
- December 5, 2024: Rounding method examples for 2024 RAU.
- April 20, 2023: Rounding method examples for 2023 RAU.
Feedback
0 comments
Please sign in to leave a comment.