PDO Streaks and Season-Long Performance Spreads

by Hawerchuk on Jan 3, 2012 7:00 AM EST in Statistics

Here are the best and worst Shooting %, Save % and PDO over the last four seasons:

	Min	Max
Sh%	6.7	11
Sv%	891	938
PDO	975	1037

This is really intended as a cautionary tale - in both positive and negative directions. If your team is posting a 12% shooting percentage or a 1040 PDO through 10 or 20 or 30 games, you need to ask yourself if it's illusory, or if your team is the best team we've seen in half a decade. Notice that the spread of PDO is less than the spread of shooting and save percentage independently; there are a variety of reasons for this - systematic shot counting bias at various arenas, and the fact that a team that cheats on offense to score more often has to give something up on defense.

At any rate, here's the range of observed PDO over 1-82 game stretches:

The spread over a full season is roughly half as large as the spread over 20 games. So whatever you observe in a quarter-season needs to be tempered by 50%; and don't be surprised if a team riding a PDO high starts losing as they come back to earth.

Tweet Comment 3 comments | 0 recs |

Do you like this story?

Comments

This is really bad news for Buffalo.

by truck on Jan 3, 2012 9:35 AM EST reply actions

Great graphic. Have you looked at Fenwick + PDO.

A team that gets lucky streak of scoring goals may temporarily inflate their PDO (via higher SH%) as you show beyond what their shots for would suggest they should score. And the team Fenwick (shots for component) will remain low as this is more dependent on skill then luck. The regression of Fenwick does not include the measure for luck (PDO). Likewise the regression of luck does not include the shot skill components (Fenwick).

Creating a measure that combines both luck and skill could result in better correlation to observed winning. As a first step, would be a simple addition of PDO + Fenwick.

by jeffgm on Jan 3, 2012 10:29 AM EST reply actions