Why use expected goals for and against when you can use actuals. last time I checked the games were won and lost on actuals
2016-17 5 on 5
PK GF60 2.4 GA60 2.6 net -0.2 G60
Weber GF60 2.2 GA60 1.4 Net 0.8 G60
One of these things is not like the other.
Because when you use actual goals against you're giving skaters credit for the quality of goaltending behind them, which is something they don't control, and quality of goaltending at the other end, which over a single season will have a fair bit of randomness in it.
From 2011-2016 Shea Weber was on-ice for a 7.71% 5on5sh% and 92.7 5on5sv%.
From 2011-2015 P.K. Subban was on-ice for an 8.87% 5on5sh% and 92.44% 5on5sv%
The goalies on the teams stayed the same, why would you expect them to diverge as wildly as they did in the next two seasons because of the play of either defenceman?
2016/17 on-ice sv%: Subban 91.85 Weber 95.46
2017/18 on-ice sv%: Subban 93.4 Weber 89.96