I have to put out a correction of my last blog post. Sorry about that! The impact of scoring the first goal wasn’t as large as I stated yesterday. I had a bug in my script that made erroneous game-states on goals scored. I have now corrected it and carefully checked the results. Thanks to the comment from Sam in the blog post I now also can compare the figures for Premier League which feels comforting. This is what the plot actually should look like:
Sorry for that but it feels good to have that sorted out. My XG model feels much better now. GS isn’t influencing it as heavily. I will also based on this be able to look more into players and teams performance under different game-states.
Lastly the (true) XG map of ÖFK – Sirius then looks like this:
I guess I made a huge mistake in my last blog post. Sorry about that. I feel a bit ashamed right now 😦 The true difference in scoring efficiency between trailing and leading games is never more than 6%. Not 60% as I stated. Should have known that!