Is there any statistical evidence that developers who practice leetcode-style algorithms & software development are better than developers who just practice software development?

By "practicing algorithms" I mean solving leetcode style problems, where you have to come up with more advanced algorithms and then analyze time/space complexity... Which from my experience is not really similar to developing software and is time that could've been spent actually practicing software development.

I really enjoy solving leetcode problems but if one day the majority of companies stopped giving these types of technical questions, would all the time spent on leetcode and similar sites become meaningless?