PDA

View Full Version : Post all your cries whines, objections, complaints here



Pages : 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 [16] 17 18

PrezdeJohnson09
03-09-2011, 10:49 AM
I'm just going to make one last statement about CSC's PIM - 1985 said that their mean was distorted by some outliers. The best way to measure the center or typical value for a data set with outliers is to consider the median value. The median is a resistant measure of center which is not distorted by unusual values. For CSC PIM the median is 18 - not so different from the mean of 19.8. Interestingly enough, the median PIM for Castleton's opponents was 14. I'm not trying to make a value judgement here, just reporting data.

Prof,

Something I'd be interested in knowing is the difference between a team's average penalty minutes over a season and how many penalty minutes they had in games against Castleton.

joecct
03-09-2011, 10:51 AM
I do not think it is appropriate to whine about the whiners on the whines thread. Those that do should be ashamed of themselves. :pThey could call The Big Show and get on the Whiner Line.....

PSUChamps2001
03-09-2011, 11:27 AM
Prof,

Something I'd be interested in knowing is the difference between a team's average penalty minutes over a season and how many penalty minutes they had in games against Castleton.

16 times the opponent averaged more penalty minutes when playing Castleton then their average penalty minutes...and actually the average is kind of big....there was a difference of 234 penalty minutes over 16 games...so an average of 16 penalty minutes per game more then their original penalty minutes of those 16 games. 10 of those games saw a difference of 8PIM's.

Overall teams averaged 16 PIM per game against Castleton, but only averaged 7.5 overall.....take out their Castleton's game and it would be an even bigger difference....

NUProf
03-09-2011, 11:33 AM
Prof,

Something I'd be interested in knowing is the difference between a team's average penalty minutes over a season and how many penalty minutes they had in games against Castleton.

In 16 of Castleton's 27 games, teams took more penalties than their season long average. Statistically this is not a significant difference from what we should expect which would be that this should happen 50% of the time (or 13.5 games). The statement that CSC sucks teams into committing more penalties than they normally would is NOT supported by this data.

NUProf
03-09-2011, 11:35 AM
16 times the opponent averaged more penalty minutes when playing Castleton then their average penalty minutes...and actually the average is kind of big....there was a difference of 234 penalty minutes over 16 games...so an average of 16 penalty minutes per game more then their original penalty minutes of those 16 games. 10 of those games saw a difference of 8PIM's.

Overall teams averaged 16 PIM per game against Castleton, but only averaged 7.5 overall.....take out their Castleton's game and it would be an even bigger difference....

It's not a fair analysis unless you also account for the 11 games in which teams took fewer PIMs than their average.

PSUChamps2001
03-09-2011, 11:36 AM
In 16 of Castleton's 27 games, teams took more penalties than their season long average. Statistically this is not a significant difference from what we should expect which would be that this should happen 50% of the time (or 13.5 games). The statement that CSC sucks teams into committing more penalties than they normally would is NOT supported by this data.

True, but look at how many they take when they go over....even over the 27 games they take twice as many PIM....of those 16 where they went over those teams took on average 16 more minutes......

NUProf
03-09-2011, 11:38 AM
True, but look at how many they take when they go over....even over the 27 games they take twice as many PIM....of those 16 where they went over those teams took on average 16 more minutes......

Not statistically significant...

PSUChamps2001
03-09-2011, 11:39 AM
It's not a fair analysis unless you also account for the 11 games in which teams took fewer PIMs than their average.

and in those 11 games teams only average 4 PIM fewer then their average....compared to the 16 they average when teams go over....

PSUChamps2001
03-09-2011, 11:41 AM
Not statistically significant...

you're telling me 8 penalty minutes more per game is not significant over the course of a season??

snc1
03-09-2011, 11:41 AM
Geeeez......just no love for the BullDogs! For your sake, I hope that SNC & Adrian can get at each other for another tilt! I hate seeing you lose sleep dealing with all this anger. :)

I'd take that. That would mean the title would be back in the west.

1985
03-09-2011, 11:44 AM
I'm just going to make one last statement about CSC's PIM - 1985 said that their mean was distorted by some outliers. The best way to measure the center or typical value for a data set with outliers is to consider the median value. The median is a resistant measure of center which is not distorted by unusual values. For CSC PIM the median is 18 - not so different from the mean of 19.8. Interestingly enough, the median PIM for Castleton's opponents was 14. I'm not trying to make a value judgement here, just reporting data.

Ok bottom line is we agree CSC needs to improve on staying out of the box, statistically they are not the worst team in the country but def need improvement, but I'm not going to buy into the fact that CSC causes the other teams to be undisciplined...I'm sure they would like that power but ....

NUProf
03-09-2011, 11:46 AM
you're telling me 8 penalty minutes more per game is not significant over the course of a season??

Too many outliers to justify the use of the t-test with a sample of size less than 30. Using a non parametric test like the one I used is the only statistically responsible procedure, and the results suggest a trend, but are not significant.

hockeyfan77
03-09-2011, 02:51 PM
As for USM, I do think they are making good strides and after getting some players back from injuries they had an impressive 4-2-1 record down the stretch, they have seemed to cut their penalty minutes and that is always a good thing......may want to get more than 15 shots on net thou to have any chance against Norwich :)


Prez doesn't know how to count shots: USM had 18!!!! :)

I am pretty sure CSC only had 20 so I wouldn't use that measuring stick!!! :)

1985
03-09-2011, 04:12 PM
Prez doesn't know how to count shots: USM had 18!!!! :)

I am pretty sure CSC only had 20 so I wouldn't use that measuring stick!!! :)

Ha...26 on the game sheet...and 3 went in the net.......not sure how many guys USM loses next year but if they stay healthy and find a little more O could be interesting

hockeyfan77
03-09-2011, 04:28 PM
Ha...26 on the game sheet...and 3 went in the net.......not sure how many guys USM loses next year but if they stay healthy and find a little more O could be interesting

Actually in your latest game it was 20 and you scored two. :)

Dyce
03-09-2011, 04:41 PM
Too many outliers to justify the use of the t-test with a sample of size less than 30. Using a non parametric test like the one I used is the only statistically responsible procedure, and the results suggest a trend, but are not significant.

I think it's best to defer to the professor on this one. What did you teach again... ;)

norm1909
03-09-2011, 05:30 PM
Too many outliers to justify the use of the t-test with a sample of size less than 30. Using a non parametric test like the one I used is the only statistically responsible procedure, and the results suggest a trend, but are not significant.

t-test :confused: This is hockey not golf :p :rolleyes: ;)

PSUChamps2001
03-09-2011, 05:37 PM
<table>
<tr><td> TEAM </TD><TD> </TD><TD> MIN </TD><TD> AVG MIN </TD><TD> DIFFERENCE </TD></TR>
<tr><td> St.Anselm </TD><TD> </TD><TD> 22 </TD><TD> 14 </TD><TD> 8 </TD></TR>
<tr><td> NEC </TD><TD> </TD><TD> 16 </TD><TD> 15 </TD><TD> 1 </TD></TR>
<tr><td> Becker </TD><TD> </TD><TD> 10 </TD><TD> 17 </TD><TD> -7 </TD></TR>
<tr><td> Brockport </TD><TD> </TD><TD> 10 </TD><TD> 15 </TD><TD> -5 </TD></TR>
<tr><td> Colby </TD><TD> </TD><TD> 8 </TD><TD> 13 </TD><TD> -5 </TD></TR>
<tr><td> Bowdoin </TD><TD> </TD><TD> 23 </TD><TD> 18 </TD><TD> 5 </TD></TR>
<tr><td> Skidmore </TD><TD> </TD><TD> 14 </TD><TD> 13 </TD><TD> 1 </TD></TR>
<tr><td> Norwich </TD><TD> </TD><TD> 10 </TD><TD> 9 </TD><TD> 1 </TD></TR>
<tr><td> Morrisville </TD><TD> </TD><TD> 72 </TD><TD> 21 </TD><TD> 51 </TD></TR>
<tr><td> J&W </TD><TD> </TD><TD> 59 </TD><TD> 28 </TD><TD> 31 </TD></TR>
<tr><td> Trinity </TD><TD> </TD><TD> 10 </TD><TD> 14 </TD><TD> -4 </TD></TR>
<tr><td> Wesylan </TD><TD> </TD><TD> 10 </TD><TD> 17 </TD><TD> -7 </TD></TR>
<tr><td> Potsdam </TD><TD> </TD><TD> 33 </TD><TD> 19.5 </TD><TD> 13.5 </TD></TR>
<tr><td> UMASS Boston </TD><TD> </TD><TD> 8 </TD><TD> 13 </TD><TD> -5 </TD></TR>
<tr><td> Babson </TD><TD> </TD><TD> 14 </TD><TD> 13 </TD><TD> 1 </TD></TR>
<tr><td> Hamilton </TD><TD> </TD><TD> 39 </TD><TD> 15 </TD><TD> 24 </TD></TR>
<tr><td> Amherst </TD><TD> </TD><TD> 8 </TD><TD> 12 </TD><TD> -4 </TD></TR>
<tr><td> Tufts </TD><TD> </TD><TD> 26 </TD><TD> 18 </TD><TD> 8 </TD></TR>
<tr><td> Conn College </TD><TD> </TD><TD> 14 </TD><TD> 17 </TD><TD> -3 </TD></TR>
<tr><td> S,Maine </TD><TD> </TD><TD> 25 </TD><TD> 15 </TD><TD> 10 </TD></TR>
<tr><td> UNE </TD><TD> </TD><TD> 69 </TD><TD> 22.3 </TD><TD> 46.7 </TD></TR>
<tr><td> Norwich </TD><TD> </TD><TD> 25 </TD><TD> 9 </TD><TD> 16 </TD></TR>
<tr><td> St.Michaels </TD><TD> </TD><TD> 14 </TD><TD> 17 </TD><TD> -3 </TD></TR>
<tr><td> Williams </TD><TD> </TD><TD> 8 </TD><TD> 13 </TD><TD> -5 </TD></TR>
<tr><td> Middlebury </TD><TD> </TD><TD> 24 </TD><TD> 11 </TD><TD> 13 </TD></TR>
<tr><td> NEC </TD><TD> </TD><TD> 18 </TD><TD> 15 </TD><TD> 3 </TD></TR>
<tr><td> Babson </TD><TD> </TD><TD> 12 </TD><TD> 13 </TD><TD> -1 </TD></TR> </TABLE>

1985
03-09-2011, 05:38 PM
Actually in your latest game it was 20 and you scored two. :)

Last game was a 4-3 loss to Babson, not sure what you are looking at and the shots were 26

collegehcky2
03-09-2011, 07:03 PM
Too many outliers to justify the use of the t-test with a sample of size less than 30. Using a non parametric test like the one I used is the only statistically responsible procedure, and the results suggest a trend, but are not significant.

Can you regress it across other teams? Or is that what you did?