Well, discounting those games our xG is 9.4 in total, or 1.15 per game, compared to a npxG of 1.31 if you include those games. So it does skew the data, but not in the way you were hoping for.There's a flaw in your reasoning. And that is in games when we score penalties, we have less need to create, we simply need to defend. Therefore you need to remove games with penalties from the equation altogether and only look at non-penalty xG and xGA for games without penalties. I suspect that skews the data somewhat.