
the claim in the blurb is off. look at those graphs! take out murshidabad, you see little statistical variance between '21 & '26 trends, which means you've to look at murshidabad separately (highest deletions!) not use it to impose a pattern that doesn't fit the rest of the data!








