No, it’s not. Standard error of the mean estimate is SD/sqrt(n-1), so it depends on the signal/effect to noise ratio how much you can learn from it. SE of the mean difference for a between subjects design is slightly different but not miles off from that. You can squeeze even more info out of it by using within subjects designs.
What makes you think a sample size of 50 is irrelevant?
No, it’s not. Standard error of the mean estimate is SD/sqrt(n-1), so it depends on the signal/effect to noise ratio how much you can learn from it. SE of the mean difference for a between subjects design is slightly different but not miles off from that. You can squeeze even more info out of it by using within subjects designs.
What makes you think a sample size of 50 is irrelevant?