Community Review - Score weight

A score weighting will be applied to all reviews submitted beyond the reviewer’s allocation.

In past funds, the score for each criterion was calculated with the simple average of all the valid reviews. Still, this approach leaves the door open for attacks on score ratings since a bad actor could feasibly create an unlimited number of LV0 accounts and significantly influence the final score.

To prevent this, each criterion's score is calculated differently by weighting the allocated reviews versus the non-allocated reviews. Allocated reviews are weighted 80/20 against non-allocated reviews.

Reviews by LV0 and LV1 are not treated differently in the voting app interface; it is only the weighted score that differs, which impacts the final aggregated score rating of reviews. The last rating for each criteria calculation is protected against spam attacks by giving more weight to allocated reviews.

Implementation:

• The “assessor” field for each review will be in the form “0#4123” where:

  • “#” is a static separator,

  • “0” indicates the type of the review (0 = not allocated, 1 = allocated),

  • “4123” is the identifier of the reviewer.

• The weight for “not allocated” reviews is calculated: $w_a = X / n_a$ where:

  • $X$ is the total weight for “not allocated” (i.e., 0.2)
  • $n_a$ is the total number of “not allocated” reviews
• The weight for “allocated” reviews is calculated: $w_b = Y / n_b$ where:

  • $Y$ is the total weight for “allocated” (i.e., 0.8)
  • $n_b$ is the total number of “allocated” reviews
• The final score is calculated using the weighted average and the weights previously calculated for “allocated” and “not allocated” reviews.

  • $W = \frac{\displaystyle\sum_{i=1}^n w_i X_i}{\displaystyle\sum_{i=1}^n w_i}$

    • $W$ = weighted average
    • $n$ = number of reviews to be averaged
    • $w_i$= weights applied to reviews score (steps before)
    • $X_i$ = reviews score to be averaged

Example:

• We have 5 “not allocated” reviews, each of them with a score of 5.

• We have 2 “allocated” reviews, each scoring 3.

• Total weight for “not allocated” is 0.2

• Total weight for “allocated” is 0.8

• We calculate the weight for “not allocated” reviews: 0.2 / 5 = 0.04

• We calculate the weight for “allocated” reviews: 0.8 / 2 = 0.4

• We calculate the weighted average using these weights:

$\frac{(0.04 * 5) + (0.04 * 5) + (0.04 * 5) + (0.04 * 5) + (0.04 * 5) + (0.4 * 3) + (0.4 * 3)}{0.04 + 0.04 + 0.04 + 0.04 + 0.04 +0.4 + 0.4}= 3.40$