Mise en scène: ~ 2 minutes left in game 1 of the NL final, Zug leads Zürich 2:1, ZSC has pulled their goalie for an extra attacker.
58:10 Puck goes briefly into the netting above the boards, play continues
58:31 6v5 Goal ZSC, new score 2:2
58:31 Zug challenges for Goaltender Interference, apparently because they doubt the video could prove the puck was out of bounds and think the goaltender interference can more easily be proven
58:31 Challenge is denied, Zug is assessed a penalty for the failed challenge
59:58 ZSC scores on the PP, new score 2:3
60:00 ZSC wins
Now, there is a debate to be had whether or not Zug made an error when they decided to challenge goaltender interference, believing there was insufficient video to prove the puck went out of bounds (which would also be challengeable).
MySports refereeing expert Tobias Wehrli certainly believes the challenge for the puck leaving play would have been successful:
But that’s not actually what I want to talk about. I want to talk about the seeming consensus online that Dan Tangnes somehow messed up by throwing out a “bad” challenge.
I very much disagree because the math actually backs up Tangnes’ decision here.
Simple Math
Easy assumptions:
Zug is incredibly likely to win if they win the challenge. My Win Probability Model has them at 93% win rate with 1:29 left up by 1
If they do not challenge, the win probability model has them at 50%
Tricky assumption, the failed challenge scenario:
This is a bit more ambiguous. My win probability model would suggest something like 47%, but let’s do our own math.
ZSC’s PP%: 21%, EVZ’s PK%: 87%, so let’s just take the middle ground here and assume that there is a 17% chance ZSC score on a given powerplay
we’re ignoring the possibility of a shorthanded goal or that ZSC or EVZ take another penalty
if ZSC scores, they win (which is true in OT, not necessarily true in regulation)
So ZSC has a 17% chance of winning the game right away. Leaving EVZ with a 83% chance of not losing the game on that PP. In those 83%, we assume there’s still a 50-50 chance they win, so rounded down 41%.
These are the three scenarios for Zug’s chances of winning the game:
No Challenge: 50%
Successful Challenge: 93%
Unsuccessful Challenge: 41%
This means that in order to be a good decision in terms of increasing Zug’s chances of winning, they need at least a 17% chance of successfully challenging the goal. Anything beyond 17% is additional win probability in Zug’s favour.
I.e. if Dan Tangnes’ video coach looks at the goal and says there’s more than roughly a 1/6 chance of a referee deciding “that’s goalie interference”, challenging the goal is a smart decision. And since it seems like a perfectly reasonably goalie interference challenge to me, so Tangnes made the right game theory choice, even if it didn’t work out in the end for Zug.