Patrick Caskie @patrickcaskie.bsky.social · 08/01/2025Happily breaking my hiatus to share some Igamane propaganda. 142
Patrick Caskie @patrickcaskie.bsky.social · 16/12/2024Difference between goal difference and expected goal difference is less stark but still overwhelmingly concerning; the trends decoupled after the 2022 SCSF win (2-1 aet). Insert any number of GK errors and unbelievable misses as evidence. Underperformance here is -8.35. 010
Patrick Caskie @patrickcaskie.bsky.social · 16/12/2024Don't take any model as gospel; tracking xPoints and Points earned is not pretty viewing. A decoupling started after the 2-2 draw in Jan 2023 and has since seen us win just 5 points in the subsequent 11 games, representing a suggested underperformance of -10.48 points. 100
Patrick Caskie @patrickcaskie.bsky.social · 16/12/2024Here's the xG plotted from all the old firms since the title win in 2021. Celtic's the dominant force and partly undersold given game states still, Rangers are consistently coming away short, mainly due to chronic penalty-box performances and obvious mental/personnel deficiencies 100
Patrick Caskie @patrickcaskie.bsky.social · 29/11/2024A crude categorisation of "big games" is used here but adds some good context. It's fine teetering on decent, although the Celtic doubts almost single-handedly override that. Like many before him, it's not the games plotted here where the precarity comes from. 021
Patrick Caskie @patrickcaskie.bsky.social · 23/11/2024If you're going to be a bad team, you'd love to get out of jail regularly; that's certainly not been the case. We hardly deserve any luck on or off the field since we traded our soul for that UEL run. 121
Patrick Caskie @patrickcaskie.bsky.social · 23/11/2024It's crossed the threshold; it's untenable to deny that previous coaches (GvB + Beale) were sacked for less; both were no doubt deserving of their dismissal. It is much more than just UCL prospects that trickle away every day... 131
Patrick Caskie @patrickcaskie.bsky.social · 15/11/2024I would love to see this with an ELO-filtered sample size, just positive outliers/title challengers; I am sure Sporting would still look great! Below is an example from Rangers from last season; without the filter, they're in the 90s for pretty much everything. 131
