DeacMan
Well-known member
- Joined
- Mar 20, 2011
- Messages
- 9,708
- Reaction score
- 948
Forest will spend again and stay up is my guess.
zero corner goal stat is interesting but i have no idea what i'm to gather from those graphics
That particular account isn’t my favorite model for expected goals, it’s fairly rudimentary. Fbref, which uses Statsbomb, is probably the best, their model is very sophisticated, taking into account the most variables, and their xG and xGA numbers have proven the most correlative to goals and goals conceded forecasts over a long span.I don't see a breakdown on the graphic/tweet, but how do they calculate expected points with regards to draws? Like Team 1 0.89 - 0.90 Team 2 would equal a win for Team 2? Or is there a margin of error that equals a draw?
Ahh ok - that makes more sense. Thanks.That particular account isn’t my favorite model for expected goals, it’s fairly rudimentary. Fbref, which uses Statsbomb, is probably the best, their model is very sophisticated, taking into account the most variables, and their xG and xGA numbers have proven the most correlative to goals and goals conceded forecasts over a long span.
But to answer your question, I think it’s meant to be accumulative probabilities over time. Like if you played the same game 100 times with a .89-.91 scoreline, you’d expect a range of 0-0, 1-0, 1-1, 2-1, etc outcomes, and this is just the accumulation of the most likely outcomes from however many simulations their model runs. It’s “proprietary” insofar as they use their social account to sell books, don’t think they’re even an actual data company.