>>45146069A lot of you talked about how OP's sample size is too small. You all get half points. But how do you determine if a sample size is big enough? You can perform a power analysis, which gives you an estimate of what you need based on the effect size. Let's pretend OP's results are somehow exactly true, that on average male Obstagoons outdamage female Obstagoons by 1.38% HP damage. We have to use extensive simulations for power analyses for complex experiments, but since this is simple we can just use prebuilt software, like G*Power.
95% confidence that the differences do NOT arise from random chance is considered the bare minimum for significance, so we'll be nice to OP and use that in our calculations, as seen in the field 1-β. The allocation ratio indicates the ratio of males to females, and we're just going to stick to 1. The important part is the effect size, which the program will calculate for us by feeding it the means and standard deviations of each sample. You might have seen those being calculated in the last pic.
So it turns out that if OP's data is true in reality, we need at least 27 data points from both males and females to be 95% certain that any difference between them is significant. Time to hit showdown.