Card: the Layout section no longer names statistics/, which was removed

README.md

@@ -86,8 +86,7 @@ One folder per dataset:
 - `covariates/<dataset>/train-*.parquet`: `item_id`, `source`, `freq`, `target`, `covariates`
   (`[covariate][time]`, aligned with `target`), `covariate_names`.
 
-`manifest.json` lists every dataset with its source, frequency and license, and `statistics/`
-breaks the training data down by domain and frequency.
+`manifest.json` lists every dataset with its source, frequency and license.
 
 Series were cleaned (invalid values masked, duplicates removed) and given neutral ids.