And how would you like to regress today, Mr. Bond?

Cam Charron
November 09 2011 01:32PM

Probably the most unfortunate thing about the start of the new NHL season, other than all the hope and promise that accompanies your new team with its new additions, is all the hope and promise that comes when a team manages to string together a few wins in October.

The PDO of the Toronto Maple Leafs (save percentage plus shooting percentage) has slumped from about 108 after the first two weeks to 101.2 after last night's loss to the Florida Panthers. The offense, once clicking on all cylinders with Phil Kessel leading the way, has now scored just a single goal in the last two games. The goaltending, once an area not of concern with the play of James Reimer, has fallen to new lows as Jonas Gustavsson and Ben Scrivens, to put it quite plainly, haven't given the Leafs what they needed since Reimer has gone down.

Equally disconcerting is that when Toronto were winning games, they were close victories, which is an unsustainable way of winning. Given how much luck is required to win a close hockey game, eventually, that regresses as well. This was explained by Bill James in something he called the Pythagorean Expectation, and it introduces the theory that a hockey team can really only control two things: the number of goals scored, and the number of goals scored against. The time these goals occur, and in what order, balances out over the course of a long season.

Here is the relatively simple formula for calculating a team's expected winning percentage:

Goals For ^ 2
( Goals For ^ 2 ) + ( Goals Against ^ 2 )

This is something previously discussed by Chemmy, but with the Leafs having fought off the PDO monster with two-straight blowout defeats, they now face the Pythagorean monster. The team's 9-6 record is disproportionate to its minus-5 goal differential.

First, let's look at the current Eastern Conference standings:

Rank	Eastern	GP	Pts	Wins
1	Pittsburgh	15	21	9
2	Toronto	15	19	9
3	Washington	13	18	9
4	Buffalo	14	18	9
5	Philadelphia	14	18	8
6	Florida	14	17	7
7	NY Rangers	13	17	7
8	Tampa Bay	14	16	7
9	New Jersey	13	15	7
10	Ottawa	15	15	7
11	Carolina	15	13	5
12	Winnipeg	15	13	5
13	Boston	13	12	6
14	Montreal	14	12	5
15	NY Islanders	12	10	4

Now, let's look at the Eastern Conference standings, with wins adjusted for Goals For and Against:

Rank	Eastern	GP	Wins	xW	GF	GA
1	Pittsburgh	15	9	9.5	45	34
2	Buffalo	14	9	8.7	42	33
3	Philadelphia	14	8	8.7	56	44
4	Boston	13	6	8.3	40	30
5	Washington	13	9	8.2	50	38
6	NY Rangers	13	7	7.7	35	29
7	Florida	14	7	7.4	39	37
8	Toronto	15	9	6.7	46	51
9	Tampa Bay	14	7	6.7	44	46
10	Montreal	14	5	6.2	35	39
11	Ottawa	15	7	6.0	45	55
12	New Jersey	13	7	5.9	33	36
13	Winnipeg	15	5	5.7	40	51
14	Carolina	15	5	5.3	37	50
15	NY Islanders	12	4	4.1	25	35

( column xW denotes "expected wins")

This chart doesn't mean that math hates the Leafs, or that Leafs Nation hates the Leafs, but shows that the team's record, thanks to its PDO and its goal differential, is slightly deceiving. The good news (sort of) about the goal differential is that prior to the team's two blowouts, the team had a +6 goal differential and would have been tied for Florida with 7.4 xW. The bad news is that those two games happened, and, since they happened, they fit into the measure of a team's overall ability.

Simply put, er, the Leafs simply have to score more goals, and start allowing fewer goals. They will win more hockey games this way. Regression has hit the Leafs PDO already and the fanbase are feeling the effects. Hopefully, the goal differential fairies don't all sting in one blow as well.

*-second chart updated to include actual number of wins

Cam Charron is BC-raised hockey fan and journalist and acts as the managing editor of The Leafs Nation. In his spare time, you can find him blogging about numbers at Canucks Army, discussing the juniors at Yahoo's Buzzing the Net, ranting in some way or another at the Score's Backhand Shelf, or acting as an MSM wannabe in the pages of the Vancouver Province.