Score = (∑ TweetScore) / TweetCount
Weights:
Likes:
×200
Replies:
×100
Bookmarks:
0-150 pts
Retweets:
-100
Hashtags:
-200 each
Em dashes:
-200 each
Engagement Calculation:
L(likes, followers) × 200 + R(replies, followers) × 100 + B(bookmarks, followers)
Retweets skip engagement calculation and get -100 penalty
Content Scoring:
Bonuses:
Text + Image:
+50
Text + Image + Quote:
+75
Penalties:
< 5 words:
-100
Per hashtag:
-200
External link:
-50
Quote tweet:
-50
Per em dash:
-200
Complexity:
C(text)
Engagement Functions:
L(likes, followers) = Power-law engagement model with sigmoid scoring
R(replies, followers) = Similar model, higher baseline (replies rarer)
B(bookmarks, followers) = Bookmark-specific power-law with 2x multiplier
Normalization:
≤ -200:
max(0, 10 + (score + 200)/50)
≤ 0:
max(0, 20 + score/10)
≤ 200:
20 + (score/200) × 30
≤ 500:
50 + ((score-200)/300) × 30
> 500:
80 + min(20, (score-500)/100)
Final score normalized to 0-100 scale