why is my stacking regressor scoring worse than its components?

I ended up fixing the problem, I had to define the final estimator in the stacking regressor, for example as such:

This improves the stacking score to roughly 0.9