When your computer models fail, then add another variable you did not include before!
That would be a form of theory patching. That does not seem to be what they are describing. Rather this merely and adjustment to an existing parameter to make a model better agree with measurements. This is perfectly respectable science.
What is not respectable is making extravagant and unsupportable claims in pursuit of grant money or acclaim.
When model #1,045 fails, develop model #1,046.