How good is Moulson (hockey sabermetrics)

How good is Matt Moulson? We're hearing these comments about Matt being one of the very best forwards in the country ... legitimate scoring thread as good as anyone in the WCHA ... etcetera ... but despite a name that has been famous across Canada for generations, he gets little respect.

So is there a stat to prove his worth? Moulson has fewer games under his suspenders than the rest of the country, 20 when the others are closing in on 30. And he has fewer points per game, but then we're a low-scoring ream (well, actually not as a season average, but we're no longer top ten as we were for a few early games).

I tried to correlate Moulson's total points and adjust for games played (Pt/Gm does that nicely), or how many points he scores vs. all the team's points, or how many points he scores relative to the margin of victory (if your team's average MOV is 1 goal and you're scoring 1.5PPG (goals and assists), then you're involved in more goals than the margin the team wins by. The best stat for Moulson appears to be his goals as a fraction of all Cornell goals" 15x68=22% (the top six range from .14 to .21).

A lot of the "stats" I tried to concoct don't hold up either because they're nonsense stats, or they don't correlate well across teams with losing as well as winning records, or (sniff, is it possible?) because Matt still is good but not outstanding. Moulson is well out of the top 25 on points per game scoring. The margin of victory calculations don't hold up because a) our MOV isn't that bad, the recent 1-0 of games notwithstanding, and of course because if you have a losing record your MOV kind of sucks. The closest I came is how often Moulson gets a point when Cornell scores a goal: 0.34 points or just over one-third for every Cornell goal. OTOH others do even better; a couple players are involved (points, not just plus/minus) on half their goals. Which means either Colin Murphy of Michigan Tech (gets a point on 59% of team's goals) is damn good or there isn't much depth beyond his line.

The most amazing stat for Moulson may be that he gets a power play goal in half the games Cornell plays. If the rate continues, once he's played the same number of games as the WCHA or CCHA, he'd have 15 PPG. If you're a glass-half-empty kind of guy, you'd wish he had, after 20 games, 10PPG and another 10 (15?) even-strength goals.

Name	               	GP         G	A	Pts.	Pt./GM	GW	PP	SH	GF	GA	PTS/Goal  %tmGoal
Marty Sertich, Jr, F, CC	28	21	29	50	1.79	4	4	2	112	64	0.45	0.19
Brett Sterling, Jr, F, CC	28	21	21	42	1.5	5	11	0	112	64	0.38	0.19
Colin Murphy, Sr, LW, MTU	25	9	32	41	1.64	0	1	0	69	101	0.59	0.13
Scott Mifsud, Sr, F, UVM	27	16	21	37	1.37	3	10	0	80	63	0.51	0.20
Brent Walton, Jr, RW, WMU	24	16	20	36	1.5	2	7	1	76	90	0.49	0.21
T.J. Hensick, So, F, UM	25	16	20	36	1.44	2	3	0	116	74	0.31	0.14
Matt Moulson, Jr., F, Cor	20	15	8	23	1.15	3	10	0	68	31	0.34	0.22

Table shows player, games played, goals, assists, points per game, GWG, PPG, shorthanded goal, team goals for and against, player points per 1.0 team goals scored, and fraction of team goals scored by player. The others are the top five scorers (36 points or more).

Other ideas that make Matt look good statistically as opposed to just being dynamite on ice? Like his goals as a fraction of all team goals (15x68=It may just be when Cornell skates four solid lines of players, nobody is going to stand out much beyond Moulson's pretty decent numbers and his eye-popping power play stats.

Here's the score after each of Matt's goals:

01 Sac 1-0 (final lead)
02 Sac 4-1
03 Brn 5-2
04 Brn 6-2
05 Ver 2-0 (2-goal lead)
06 Can 1-0 (final lead)
07 Yal 6-1
08 Prn 3-2 (final lead)
09 Mai 3-2 (final lead)
10 Brn 2-1 (final lead)
11 Uni 1-1 (tied)
12 Drt 2-0 (2-goal lead)
13 Ver 1-0 (final lead)
14 Clk 4-2 (2-goal lead)
15 SLU 1-0 (final lead)

Remarkably, 7 of Matt's 15 goals have put Cornell ahead to stay, including 6 of his last 10, while 1 has tied the score en route to a win.

Here's another fun stat, previous goal in game:

01 Sac 1-0 (first goal)
02 Sac 4-1 opp
03 Brn 5-2 CU
04 Brn 6-2 CU
05 Ver 2-0 CU
06 Can 1-0 (first goal)
07 Yal 6-1 opp
08 Prn 3-2 opp
09 Mai 3-2 opp
10 Brn 2-1 CU
11 Uni 1-1 opp
12 Drt 2-0 CU
13 Ver 1-0 (first goal)
14 Clk 4-2 opp
15 SLU 1-0 (first goal)

First Goal: 4
Opponent: 6
Cornell: 5

10 of Matt's 15 goals swung the scoring momentum to Cornell. Note that, subtracting Matt's goals out of the totals, Cornell's GF-GA is 53-31, so one would expect 37% of his non-first goals to follow an opponent's goal. His considerably higher 54% can be interpreted as him being "clutch."

The most striking stat to me is the lack of assists. Sure, he has eight, but every other player you listed has as many or more assists than goals. It's been brought up before but I worry that we only have one real scoring threat.

Amazing stat(s), and right at your fingertips. Do you have a day job? And an interesting thesis that Moulson has a disproportionate share of goals that swing the scoring momentum. (I assume when you mean a Moulson goal that put Cornell ahead to stay, eg 2-1 in a game that ends 3-2 with Cornell scoring goal 3 then the opponent scoring goal 2, that Moulson's second goal is the deciding goal and then the teams sort of traded goals.)

I wonder if there is some aggegated statistic that can measure a player's worth in hockey. With baseball, there are so many stats, it's easier, whether it's arcane such as first pitch over the plate, or slugging/on base type stats (SLOB).

With hockey, you'd probably need to draw in the plus/minus stats, or plus/minus vs. team plus/minus overall. Maybe over time as stats get more comprehensive, you could even measure things such as opponents' leading scorers' scoring after you put on a big hit (ahead or behidn his per game avg), or penalties you drew that led to an opponent scoring (it's not your fault your team has a lousy PK but if you draw a penalty knowing it ...), or penalties you drew that led to a PPG (you're so good, they have to slash you to stop you).

Maybe you also need the kinds of stats a scout would keep, at least mentally, such as the number of times someone turns over the puck, or how often in tense situations someone blindly passes off (so if the puck turns over, it's because someone else lost it).

I do recall a couple baseball overviews that say clutch hitting that's way above the average only holds true for a couple games or a world series but seldom do even the greatest players hit above their average just in the playoffs or just against the Yankees. Thus the danger of trying to find meaning in a couple games' stats as opposed to a couple seasons' stats for all 58 teams.

The point about Moulson's assists is very interesting. During the games this weekend I noticed how many shots by the team as a whole were very poorly placed or missed by 3 feet. On the other hand, Moulson's shots, when he got an opportunity, were placed well and missed by a few inches or were stopped by a very nice save. This observation doesn't clearly show up in any one stat but looking at Moulson's assists makes it a little more clear. I too fear that we have a great team but only one player who can really finish.

Stats are kept on how many shots a player takes. Is this just shots on goal? Do they also track shots attempted? The ratio between the two would give insight into shot selection and quality. If a guy misses the net all the time he'll have a low proportion of shots on goal. Not a perfect measure, of course, because blocked shots won't be SOG either. But then, that plays into shot selection too.

Just wondering.

How about one very simple stat: Moulson has 0.75 goals per game. Only 3 others in the NCAA have a better GPG, unless there are some other players that aren't on the USCHO leader board that I'm missing.

[Q]KeithK Wrote:

Stats are kept on how many shots a player takes. Is this just shots on goal? Do they also track shots attempted? The ratio between the two would give insight into shot selection and quality. If a guy misses the net all the time he'll have a low proportion of shots on goal. Not a perfect measure, of course, because blocked shots won't be SOG either. But then, that plays into shot selection too.

Just wondering.[/q]


They do track the shots attempted for the entire team, not for each player. For instance, on Saturday night, Cornell attempted 74 shots, to St. Lawrence's 54. Shots on target were 41 and 19, respectively.

Looking through the final stats sheet and shot chart for the game, it looks like Moulson attempted 7 shots. 6 were on net (and thus were "official shots" and, of course, 1 went in.

The night before, against Clarkson, Matt attempted 5. 3 were on net. 1 goal.