Quote:
Originally Posted by Burnss
ikes and thechef for worst posters of the year. such a lock.
To be clear, I have absolutely no problem with xG. It is an extremely illuminating metric with a lot of predictive power. The basis of the model is that shots are treated as random events with a probability distribution. No problem with that.
What seems to be happening is that some people are so wrapped up in this mode of thinking that they think that shots actually ARE random events. That's simply not the case in any meaningful sense. In reality, the execution of a good shot requires skill. It is not at all akin to, say, rolling a die. If you play a game where you have two throws of a 10-sided die and you need to throw a 2 twice in a row to win £1,000,000, then you can consider yourself lucky if you win this game. You have no influence over the outcome and the likelihood of winning is 1%, which is slim. If you're Philippe Coutinho shooting from 30 yards and you ping the ball into the top corner when xG model says that situation yields a goal 1% of the time, the conclusion is that you have executed a ****ing great goal. It's a weird definition of lucky that includes accomplishing exactly what you aimed to achieve.
Now, either you can tell me why you think this is wrong, or you can carry on sniping without adding anything of substance. Maybe if I can do the same I might become a better poster