Calgarypuck Forums - The Unofficial Calgary Flames Fan Community - View Single Post

Backstop · 11-18-2022, 08:32 PM

A goal! Calgary scores! 60 Seconds later the other team scores, what happened?

Although we can often identify player x lost his man and player y has a wicked wrister, a large portion of a fortunate event occurring is random and falls into the category of luck, this is simply because there is no other category to put it in.

We tend to dismiss luck as it's messy, unpredictable and we don't like things we can't measure and quantify, but we have to accept its usage for lack of a better bucket to put fortuitous events in.

When all the factors are right and the perfect shot hits the crossbar and deflects down (and in) it's arguably reasonable to say it could have just as easily gone up, that's luck. Now add in the goals that deflect off player a to player b and in, another unplanned yet fortuitous event, luck.

So how do you account for luck and how do you gauge your goalie when looking at stats if you need to know is he ineffective or do a lot of lucky goals go in?

There is no perfect measurement for luck, the commonly accepted closest stat for taking luck into account and measuring a goalies ability to stop pucks as well as a teams ability to score is PDO.

PDO is simply a teams 5 on 5 shooting percentage and save percentage measured over a specified time interval.The mean average of PDO is ~100 (sometimes 1.00), in other words, across the league the average goals and saves per team, neither good nor bad is ~100.

-A number greater than 100 is good shooting and goaltending (lucky) and expected to win
-A number below 100 is poor goaltending and shooting (unlucky) expected to lose

Large sample sizes are required to reduce the effect of luck as much as possible and boil the results down to a measurement of a teams ability to score goals and stop shots and by extension their chances of winning.

The Flames PDO is currently 1.00, if you accept PDO as a reasonable stat then the Flames don't need to replace Markstrom.

**I am not a stats guy or an analytics guy, this is just a quick and dirty attempt at explaining PDO's ability to reasonably quantify luck vs skill in a game that includes a lot of luck and why it's the stat I generally refer to. I am sure there are some regulars here that can explain PDO and the math behind it in better detail as I'm no expert.

11-18-2022, 08:32 PM	#205
Backstop Scoring Winger Join Date: Nov 2017 Location: YYZ Exp:	A goal! Calgary scores! 60 Seconds later the other team scores, what happened? Although we can often identify player x lost his man and player y has a wicked wrister, a large portion of a fortunate event occurring is random and falls into the category of luck, this is simply because there is no other category to put it in. We tend to dismiss luck as it's messy, unpredictable and we don't like things we can't measure and quantify, but we have to accept its usage for lack of a better bucket to put fortuitous events in. When all the factors are right and the perfect shot hits the crossbar and deflects down (and in) it's arguably reasonable to say it could have just as easily gone up, that's luck. Now add in the goals that deflect off player a to player b and in, another unplanned yet fortuitous event, luck. So how do you account for luck and how do you gauge your goalie when looking at stats if you need to know is he ineffective or do a lot of lucky goals go in? There is no perfect measurement for luck, the commonly accepted closest stat for taking luck into account and measuring a goalies ability to stop pucks as well as a teams ability to score is PDO. PDO is simply a teams 5 on 5 shooting percentage and save percentage measured over a specified time interval.The mean average of PDO is ~100 (sometimes 1.00), in other words, across the league the average goals and saves per team, neither good nor bad is ~100. -A number greater than 100 is good shooting and goaltending (lucky) and expected to win -A number below 100 is poor goaltending and shooting (unlucky) expected to lose Large sample sizes are required to reduce the effect of luck as much as possible and boil the results down to a measurement of a teams ability to score goals and stop shots and by extension their chances of winning. The Flames PDO is currently 1.00, if you accept PDO as a reasonable stat then the Flames don't need to replace Markstrom. **I am not a stats guy or an analytics guy, this is just a quick and dirty attempt at explaining PDO's ability to reasonably quantify luck vs skill in a game that includes a lot of luck and why it's the stat I generally refer to. I am sure there are some regulars here that can explain PDO and the math behind it in better detail as I'm no expert.