Molyneux's argument against the existence of God depends, essentially, on a Christian definition of God, while Johnson failed for a long time to make the simple case that you cannot prove a negative.
"We," regrettably, came up a lot. Individuals could know or experience what you do not. What science "knows" is constantly changing, for example. MolyNeux actually used argument from necessity by implying that God would defeat science, which we need. BTW, science "proves" itself "wrong" regularly, when a scientist demonstrates how the assertions of those before him have been wrong.
How similar is Molyneux's argument to the fact that circles do not actually exist? Interestingly, he continually repeated the thing about "square circles."
In short, God could in fact exist, and it cannot be disproved. A simple definition such as the creator, and possibly one who can intervene in daily activities, may not be provable, but to dismiss the possibility out-of-hand is to dismiss all the things we didn't believe possible which we now know are.
I'm sure Johnson knew he would be debating someone who is obviously a superior debater, but being correct certainly made it interesting anyway.