Quote:
Originally Posted by Saqe
I think all these advanced stats and predictions are flawed one way or the other, or rather the conclusions that people make of them are.
Notice for example how it this simulates over 100 times (who knows how many, maybe thousands) and statistically that would be significant but in reality the series is just seven games. The historic win % is interesting though, would be interesting to know in detail how it works.
|
Yeah, it's interesting that the one you shared had a different perception of the game. You have to manipulate the chart to expected goals from Corsi (I don't really think Corsi is a great way of analyzing hockey), but it shows the game a lot more even than 74/26.
The fact that there are fairly wide ranges of analysis indicate that we don't understand well enough what causes goal suppression or creation with certainty. It doesn't mean the models are worthless. (Not directed at you, but the people who responded pedantically.)