Why do we need a bigger ratio for non random adapation for fitness, when one is all that is needed?
If I tell you that the coin is showing "heads", how do you know if it landed that way as the result of a fair toss, or because I placed it that way?
In order to determine if I'm placing coins heads-up or flipping them, you need enough coins for it to be a statistically meaningful sample, and then you need to see if the number of coins heads-up is significantly different than the number you would expect as a result of chance. Chance says that the odds of "heads" in a fair toss is 1/2, but a single coin can't be half heads and half tails - it has to be one or the other.
Therefore a single coin can't tell us whether it was the result of chance or "intelligent placing".
If you step outside of your data and look at the wholistic marvel of evolution, it is pretty hard to rule out non-randomness for beneficia adapations.
No, it isn't. In fact, it's impossible to find any evidence of non-randomness of mutation. The results of mutations are invariably random with regards to fitness. Are the rates of mutations always random with regards to mutation? No, because an organism's phenotype determines its rate of mutation. We wouldn't expect anything else.