Yes definitely -- although it's pretty crazy how much not being under pressure affects your ability to complete the question in X minutes. I've learned over time that good interviews take a lot of experimentation, and the best thing you can do is test/refine questions in real interviews before you rely on them as a signal to make a hiring call on a candidate.

I wouldn't necessarily consider these questions a substitute to an algorithms question, but rather a way of obtaining a different and important signal. An algorithms question may be a valuable signal too, depending on the nature of the role.