8/22/2008

Goalies Losing Points

I believe it was Richard (though it could have been Rob) who asked me to check on how many goalies lost teams pts last season. This list does not include benched or undrafted players but only those who were actually played last season by FunHL teams and lost points.

Bladerunners (0):
no negative pts from goaltending

Great Whites (-0.2):
-.20pts from Lundqvist in wk 19

Ramapithicines (-0.93):
-.93pts from Turco in wk 5

Wolves (-1.2):
-1.20pts from Khabibulin in wk 4

Edge (-2.2):
-.07pts from Ward in wk 8
-2.13pts from Ward in wk 11

Knights Templar (-2.4):
-1.33pts from Fleury in wk 1
-1.07pts from Biron in wk 5

Shadowmen (-2.53):
-.73pts from Lundqvist in wk 11
-1.80pts from Luongo in wk 26

Highlanders (-3.27):
-.33pts from Kiprusoff in wk 9
-.87pts from Kiprusoff in wk 13
-2.07pts from Kiprusoff in wk 20

Scourge (-3.66):
-1.00pts from Biron in wk 4
-.33pts from Legace in wk 16
-2.33pts from Legace in wk 24

Lost Boys (-4.74):
-.67pts from Osgood in wk 15
-2.20pts from Osgood in wk 16
-1.87pts from Hasek in wk 24

Severed Heads (-4.87):
-.07pts from Hasek in wk 2
-1.67pts from Hasek in wk 7
-2.33pts from Nabakov in wk 16

Personal Vendetta (-13.53):
-1.07pts from Vokoun in wk 1
-5.93pts from Mason in wk 2
-.40pts from Leclaire in wk 18
-2.13pts from Leclaire in wk 24
-4.00pts from Toskala in wk 27

5 comments:

Bladerunner said...

Wow stats wizard, you move fast man!

Trying to think what the Bladerunners trick was to avoid any goalie point losses. Must have been pride in the team they were playing for. he he he ;-)

Cameron said...

Weird pattern in some of the numbers. Weeks 2, 4, 11, 24 all had two netminders roll backwarnds in the week, and 24 had three netminders (Leclaire, Hasek and Legace) fail their saving throw vs suckitude.

Richard said...

What seems important to me to conclude from this:

There is no need to fix the "negative" points situation. It almost doesn't exist--unless Darrell's about to start banging the drum about it. I think that conclusively means there's no need to "adjust" the current system for negatives. It really comes down to a simplicity argument

With regards to the save % and good goalies--a low save % on a good goalie is probably a reflection of the team playing in front of him. In the NHL, good goalies are the ones that win games--and has has been argued blue by Cam, goalies aren't actually responsible for their team's win. Goalie's are just responsible for their team "not losing." You don't need a good save% if the other team isn't getting many shots on the net.

I find this more fuel for the "Mike system." Simpler math, different stats, same result.

Richard

Cameron said...

I agree with your conclusion that there is no real concern for negative goaltender stats, as they are virtually non existent. Whats more I think they would show that in the case of those goaltenders involved, they let in a larger number of goals than a good goaltender having a good week should.

But then Richard said:

"With regards to the save % and good goalies--a low save % on a good goalie is probably a reflection of the team playing in front of him."

Which is actually false. The key statistics that scouts use to evaluate goaltender prospects (in the real world) are; save %, GAA and minutes played. We currently use 2 of the 3.

(One possibility I had considered is leaving the stats at 4.0 but multiplying the result by the goaltenders save%. In this case a goaltender who has a shutout gets all four of the points (gameX4.0x100%), but a goaltender who plays the whole game despite facing only 10 shots lets in two goals, gets just 80% of 2, or 1.6. I'd have to run the numbers for a gaggle of netminders over a gaggle of years to see what the results look like before I endorse this in any way, but I wouldn't be surprised if they tracked close enough to our regular stats. For me the killer opposition to this idea is that we are clearly making the stats more complicated, even as we may be making them more accurate.)

Mike's goaltender system rewards wins - which is not a goaltender derived statistic (who means more to a win in Pittsburgh, Conklin, Fleury or Crosby?), vs ours which rewards the goaltender for allowing fewer goals (a purely goaltender derived statistic).

It reminds me of NFL pools where instead of drafting a specific linebacker, corner, etc., you draft a teams entire defense. Sure it may be easier to do the stats for this, but you aren't really capturing the value of a guy like Brian Urlacher if all you are doing is lumping his tackles, sacks, interceptions and passes batted in with the rest of the Bear's defense corps.

Mike's system may be simpler, but it isn't a system that actually reflects a goaltenders worth.

Richard said...

I will defer to Cameron's wisdom as to how goalie values are determined by agents. And continue to disagree with him and prefer Mike's system, and be happy to see it go to vote.

I do have to ask though--why is Mike's system generating such similar scores for goalies if it's based off what I'll rhetorically suggest you are deeming "unworthy" statistics?

The "Simple" is a big big sell for me Cameron, and the shocking accuracy only adds to the bonus. And you haven't presented a compelling case for why I shouldn't trust why Mike's mighty similar numbers are of less value than the one's you are offering.

But--someone tap Bob--Bob had suggested that Mike's system would cure people from placing an injured goalie in their lineup, and put thought into who their backup would be when they were drafting.

Can I ask Doug--how many "0"s got generated by goalies last year?

Richard