Not the best example, but I recently had some models with a very narrow spike at the left and a very long tail, where you could literally see nothing regarding how response and simulations matched or not. Now you can "zoom in", if required.

m <- lm(log(hp) ~ gear, data = mtcars)
performance::check_predictions(m)

performance::check_predictions(m) |> plot(x_limits = c(0, 400))

^{Created on 2023-09-28 with reprex v2.0.2}

Codecov Report

Merging #307 (e477ebe) into main (2533ebd) will increase coverage by 0.01%.
The diff coverage is 57.69%.

❗ Current head e477ebe differs from pull request most recent head 15595fc. Consider uploading reports for the commit 15595fc to get more accurate results

@@            Coverage Diff             @@
##             main     #307      +/-   ##
==========================================
+ Coverage   44.32%   44.34%   +0.01%     
==========================================
  Files          66       66              
  Lines        4742     4758      +16     
==========================================
+ Hits         2102     2110       +8     
- Misses       2640     2648       +8