Its not nonsense but you have to look at it in the context of everything too, instead of the be all and end all. The logic of it is that for the most part, as a manager you make your system to create a lot of high quality chances and limit the chances the opposition creates. There are different systems for how to achieve this, but ultimately this is everyones plan.
All xG does is sum up how good each chance was over a game to give you how much you'd be expected to score with "average finishing". Average finishing as in data taken from similar situations across years of action and how often that chance would result in a goal basically. In general, year on year we've seen that teams tend to play pretty close to this stat and statistically it's a better marker for future league position than goals/goals conceded/shots on target/shots/possession are. Some bad luck now will hopefully turn around later in the season where we get some luck results to average it out, or we put in the same performances chance wise but go on a winning run and it looks rosier. A lot of times it all comes down to just clinical finishing, which we havent been while the opposition has been against us 3 games in a row. XG basically says that everyone tends to play to their expected goals over a large enough period of time. Incredibly small sample size for this season, but this is just quantifying reasons to keep some optimism despite the end results not panning out for a few games and expecting it to turn around if we play the same pretty much.