View Single Post
Old 05-24-2022, 02:03 PM   #121
united
#1 Goaltender
 
united's Avatar
 
Join Date: Mar 2007
Exp:
Default

Also one thing this series, and especially Game 3, has helped show is how much noise there is in small-sample expected goals metrics. Take Kane's three goals last game:

Goal 1: 5v5 2-on-1 pass, deke, backhand shot from edge of the crease to nearly empty net
0.10 expected goals (10% probability) per Natural Stat Trick.
Spoiler!

Goal 2: 5v5 partial 2-on-1 pass, backdoor tap-in to nearly empty net
0.12 expected goals
Spoiler!

Goal 3: 5v5 2-on-1 pass, partial breakaway deke
0.15 expected goals
Spoiler!

Subjective, but I don't think any of those probabilities are even close and in my opinion the third was easily the most stoppable.


These things, of course, mostly even out over a sufficient sample but there is a ton of noise in fewer than 10 periods of hockey in which a larger-than-normal portion has not been played at 5v5. Don't discount expected goals - in fact they should be used whenever possible - but proper context is always vital with these things.
__________________
"I think the eye test is still good, but analytics can sure give you confirmation: what you see...is that what you really believe?"
Scotty Bowman, 0 NHL games played
united is offline   Reply With Quote
The Following 2 Users Say Thank You to united For This Useful Post: