Adjusted Stats Leaders

[Coachk]

Comm Based, your wrong, the girl in question is 32 years old now. My point is that this has been going on since the beginning of Girls High School hockey and is nothing new. I am also not talking about a players overall talent, which you seem to be sensitive about. Simply, that the stats of these players need to be adjusted to reflect the level of competition. I have coached in the Twin Cities, Iron Range and Rochester and I have witnessed the skill level and team compositions. It is hard to put together a good team with combined schools, multiple communities, travel requirements, cost and competing with other sports. Lucky for the teams you listed, they have been able to avoid these issues lately, but that is just a small subset of a larger group.

[Maverick2000]

Coachk...I'm glad you knew hockey from the 90's too bad it's 2014,,,,,never to much is correct with the south kids. The two girls from century both could play D1 hockey as have offers. The younger one is a national camp player and last I looked there was no southern mn only national camp. Too look even further at the south red wing has one mn gopher player from last season as well as three current d1 kid two whom made national camp. Another south player at Mankato east made national camp and is going to North Dakota and a new ulm player is a national camp kid as well.
Then you have dodge county. They have two national camp kids and 4 kids with d1 offers. Rasmussen is leading state in goals and has scored against BSM where DC lost 3-2 tight game. She has goals vs EP in a 4-3 game against them and can keep going but again since she's from the south I'm sure she is not good enough to make your special list. Having watched these south kids at national camp, prospects, as well as top AAA teams like ice cats, whitecaps, and reebok you may want to expand some on your special list as these players have proven they are more then worthy since again it's not the 90's.......

[Coachk]

Why on this forum, do the current people involved in girls hockey always seem to try and put down anyone that has been doing this for many years. I don't understand, as I was just trying to explain that adjustments to stats need to be made. You have pointed out many talented players and I have no doubt they are deserving of there success. That bunch is a small portion of the larger amount of girls playing outstate. I understand you are sensitive to comments made about out state players especially Southern Minnesota, I have been there. I was the one that put together the first girls team in Rochester, that went to Kasson and spoke to the board on how girls hockey could succeed. I understand. Again I will try and say, The competion as a whole is not the same as the Cities teams. No malice intended, just the facts with demographics, distance, cost and available ice. I realize this is not the 90s, but I would think you would have alittle more respect for the people who drove the introduction of Girls hockey and took their lumps, because most didn't want it to succeed. I have no dog in this fight, just trying to provide affirmation to those who provided the stats.

[Nimrod]

Sounds like we are all in agreement. There are very good players through out MN but due to population density and probably money concentration, there are more in the Metro area. Makes sense. I think we all agree too that if a kid has 50 goals playing against mostly teams that are in the lower 1/3 of the state in talent, its tough to compare their stats against someone playing in the all world Lake Conference (just a friendly poke guys).

[Bandy]

Comment on the original post: in general I think it's a bad idea to throw out data. If, hypothetically, someone has 80 points and plays for the team with the 31st toughest SOS she gets overlooked, while someone with 40 points playing for the team with the 29th toughest SOS makes the list. I'd say your top 30 SOS filter isn't an adjustment for SOS, but simply a top-ten scorer list that includes only teams with the 30 toughest SOS.

Two brilliantly crafted (imho) algorithms -- KRAPPI and KRAPGI -- were posted toward the end of last season. These algorithms don't punish someone as much for playing a weaker schedule. It's arguable just how much you should handicap a player who plays a weaker schedule, or inflate a player who plays a tougher schedule.

I'd also say that neither approach accounts for the effect of playing talent-loaded team. If you have a line that's loaded with talent, then all 3 forwards should be racking up lots of points. They're better at using each other than teams that have one stud playing with relatively weak linemates. Are BSM's Reilly and Panic playing on the same line?

Good thread (except for the arguing). In the end, it's all fun and numbers. What matters more is the more subjective process used to make the national camps, and the college recruiting process.

[sinbin]

Bandy, thanks for refreshing my memory as to what I alluded to in an earlier post. I seem to remember that these were pretty decent models to adjust for SOS (again, these are only models, folks, so only a representation of the real world). I must be getting old, since I certainly should have remembered the KRAPPI model.

[MNHockeyFan]

Comment on the original post: in general I think it's a bad idea to throw out data. If, hypothetically, someone has 80 points and plays for the team with the 31st toughest SOS she gets overlooked, while someone with 40 points playing for the team with the 29th toughest SOS makes the list. I'd say your top 30 SOS filter isn't an adjustment for SOS, but simply a top-ten scorer list that includes only teams with the 30 toughest SOS.

Two brilliantly crafted (imho) algorithms -- KRAPPI and KRAPGI -- were posted toward the end of last season. These algorithms don't punish someone as much for playing a weaker schedule. It's arguable just how much you should handicap a player who plays a weaker schedule, or inflate a player who plays a tougher schedule.

I'd also say that neither approach accounts for the effect of playing talent-loaded team. If you have a line that's loaded with talent, then all 3 forwards should be racking up lots of points. They're better at using each other than teams that have one stud playing with relatively weak linemates. Are BSM's Reilly and Panic playing on the same line?

Good thread (except for the arguing). In the end, it's all fun and numbers. What matters more is the more subjective process used to make the national camps, and the college recruiting process.

Bandy, can't really disagree with anything you're saying here. As I mentioned above, if anyone has a better/fairer way to look at this, by all means have at it. I don't recall the discussion last year about "KRAPPI and KRAPGI" but they must have assumptions built in about how and how much you reward/penalize a player's stats depending on how strong/weak the opponents are. What assumptions are made and whether the proper weighting is put on each player's SOS, how is one to tell?

And then there's the question of how strong/weak a player's linemates are...I don't see how it would be possible to accurately take that into consideration. Although I would hazard to guess that most of the state's top scorers do have at least one very good linemate to work with, whether her linemate(s) be more of the playmaker or goal scorer type. It does seem that most coaches (including BSM's by the way) do rely on a top first line to do most of their team's scoring for them, and most also have a designated top power play unit where a lot of the points are scored.

I certainly don't claim to have all the answers and am just happy that my original post created some good discussion.

[CommunityBased]

CoachK, I wasn't wrong about the 32 year old...I was not speaking about her, I was speaking of the CURRENT Rochester girl who is the states leading scorer. I included a link to last years national camp roster. Sorry if that wasn't clear.

I only posted to show an opposing view of your example of a Rochester girl from the 90's who didn't belong on the ice with metro girls. You used this example to justify the exclusion of out-state girls. I was showing you that the current top out-state girls have been tested at the highest level (HP Program). They have been identified as one of the top 10-15 players in their age group.

You also indicated that when out-state girls play AAA they only do well do to their more talented teammates. Wouldn't that same argument apply to girls playing on the top metro high school teams? Would the Rochester girl score more playing for a top metro team and playing with better players, would a top metro girl struggle playing with less talent at Century? Who knows?

I still agree that out-state numbers are inflated due to weaker competition. Same thing happens on the boys side. So what. The most talented kids will be identified and not by stats so don't sweat it. Just enjoy the best high school hockey in the world.

Also, I do appreciate the support you listed in helping to develop southern MN girls hockey. We need more of that if we are going to keep growing girls hockey...especially in the North and South. Thanks.

[Coachk]

Comm Based, understood. Back then it wasn't about exclusuion, more about not being sure what stats to trust. I have coached a few of the current girls over the years and there is alot of talent. I have said that for years, we just needed to have consistant competition to get stronger. You seek that and then you hear the next season, one program desolved or combined etc.. The hardest part was convincing the players that they are every bit as good as the cities teams. Keep the faith and as long as they are playing and getting better, thats all you can hope for.

[sinbin]

And then there's the question of how strong/weak a player's linemates are...I don't see how it would be possible to accurately take that into consideration. Although I would hazard to guess that most of the state's top scorers do have at least one very good linemate to work with, whether her linemate(s) be more of the playmaker or goal scorer type. It does seem that most coaches (including BSM's by the way) do rely on a top first line to do most of their team's scoring for them, and most also have a designated top power play unit where a lot of the points are scored.

Good point, MHF. I don't have stats to back it up, but have seen numerous instances of a couple scenarios. (1) Two very strong players - but then a drop-off. Do you have them each lead their own line or combine them on one strong line? I've seen a lot of experimentation, but it almost always seems to come back to putting them together on the same line. I assume the thinking is, "our top line will outduel our opponents' top lines and if the 2nd and 3rd lines can break even, we'll do OK". (2) Now that you have that strong 1st line of two very strong players, but a weaker 3rd link, it usually is a case of a rising tide lifting that 3rd ship. I've seen numerous instances again where the 3rd player has some decent or even semi-gaudy statistics and when the 2 top players graduate and leave her alone, she disappears from the stat sheets. Of course, that's somewhat a function of her new linemates, too. Many other scenarios exists (e.g., one strong player, but no support - does she try to do it all while her linemates watch, or does she force them to be better and make passes to them and chalk up more assists), etc., etc.

Do agree that linemates matter for stats, but at the end of the day, the scouts still usually are smart enough to see through that and get it right.

[luckyEPDad]

I don't think SOS is the correct statistic for adjusting scoring. Some teams win because they score a lot of goals. Some teams win because they have a stingy defense. SOS also is lacking in that the number is essentially based on ranking and is not a relative measure. I think most would agree that the top 10-20 teams are close to interchangeable when it comes to how difficult they are to score against.

A proper weighting would be based on how difficult it was to score those goals, and that should be based on goals against statistics. Compute the average goals against per game for the entire state. Compute a weight that is the state average GA divided by the team average GA. Weigh individual game goals using the weight for the opposing team.

Using the most recent KRACH rankings I computed a GAA of the top 10 teams and teams 110-119 (I decided to treat Worthington's GAA of 16.17 as an outlier). The average GAA of those 20 teams is 3.19. Using this number I can compute a weight for goals scored against each team.

Hopkins GAA = 1.14, GAAW = 3.19/1.14 = 2.8
Long Prairie Grey Eagles GAA = 3.0, GAAW = 3.19/3.0 = 1.06
Lake of the Woods GAA = 7, GAAW = 3.19/7 = 0.45

Now you adjust individual player goals by the GAAW of the team the goal was scored against. A goal against Hopkins counts for 2.8 points. A goal against LPGE is worth 1 point, and a goal against LOTW is worth half a point.

What do you think?

[Bandy]

LuckyEP: using GAA to adjust goals (or points) seems like a good approach to get at scoring prowess.

MNHock: KRAPPI & KRAPGI used season-aggregated (average) points (or goals) per game, and average opponents' SOS. One could adjust points or goals for individual opponent's SOS, but last year's edition did not go to that level of detail. Personally, I think getting to that level of granularity is overkill -- especially in a 25-game season. If we had a 250-game season, then maybe. Great scorers can be held pointless on any given day.

re. points per game -- the KRAPPI Development Group only included the top 210 points scorers in the state. So, there could hypothetically be a hotshot JV player played only one varsity game, scored 10 points, and would thus have an insanely high points per game. The KRAPPI Development Group would have excluded such a case (rightly so, IMO). There's a substantial difference between this kind of filtering, and only looking at the top-30 SOS. Using points or goals per game (rather than total points or goals) for top scorers is good because you don't punish the national U18 team players for missing a few games.

[Hockeywannabe]

I do not understand how SOS makes sense if you still play very weak teams and run up goals.

Here is a "top 10" scorer in SOS.

Team + Goals

Maple Grove = 0
BSM = 0
Minnetonka = 1
Hopkins = 0
Elk River = 0
Wayzata = 0
Edina = 0
Eden Prairie = 0
Centennial = 0

This is one of the states "top scorers" is SOS, how does SOS show ability to score against good competition when it does not factor in this information? SOS tells us they have played tough teams but does not tell you weather that individual has scored in those tough games.

[Bleedinred]

This all comes off as tortured use of formulas to tell a story that a few would like to have told. It discards a lot of quality out state athletes because they don't happen to reside in section 6AA.

[wolfman]

Great post Bleedinred guy.....

[observer]

I do not understand how SOS makes sense if you still play very weak teams and run up goals.

Here is a "top 10" scorer in SOS.

Team + Goals

Maple Grove = 0
BSM = 0
Minnetonka = 1
Hopkins = 0
Elk River = 0
Wayzata = 0
Edina = 0
Eden Prairie = 0
Centennial = 0

This is one of the states "top scorers" is SOS, how does SOS show ability to score against good competition when it does not factor in this information? SOS tells us they have played tough teams but does not tell you weather that individual has scored in those tough games.

Funny stuff. Exactly as Bleedinred says. Stupid exercise. Give me a couple of top linemates any day. Much easier for the top girls on top teams surrounded by other top players. Then they don't even score against top competition. End of discussion.

[Nimrod]

Having other top players on your line can work against you too. If they end up shooting and scoring most of the time and the puck never comes back to you, pretty tough to score. Also, the deeper your team the less puck time and ice time you have. I think the stats are subject to so many variables, including line mates, competition, play time, whether the coach sits top lines in blow outs or not, etc. that they are way over valued. If the point is valuing players' talents, I would go with what the D I coaches indicate when selecting their players. The rest of of us have agendas while they are putting their jobs on the line when evaluating and selecting players.

[Bandy]

This all comes off as tortured use of formulas to tell a story that a few would like to have told. It discards a lot of quality out state athletes because they don't happen to reside in section 6AA.

That was my main criticism of only looking at the top 30 SOS teams. If you look back to last year's thread (Points leaders -- accounting for Strength of Schedule), you'll see KRAPPI and KRAPGI included quite a number of players from teams that play weaker schedules. It's hard to argue that the state's top-3 points leaders don't have scoring prowess, notwithstanding weaker schedules.

Some of us get our kicks by torturing data.

[Hockeywannabe]

I am not as familiar with the outstate girls as i am with the girls from South Suburban Conference. I would be interested to hear who some of the experts think their Top 10 scorers are, stats aside. In other words, who would you not want coming down the ice against your team with the game on the line?

[Maverick2000]

Agree that the SOS is bs....above hockey wannabe shows a top 10 kid who does not score in big games that is in this special scores club. Here are two dodge county kids not eligible for this list but what they have done vs good teams.....
Team. Rasmussen. Dodds
Benilde. 1 + 1. 0 + 1
South St. Paul. 2 + 4. 2 + 3
Centennial. 2 + 2. 2 +1
EdenPraire. 2+ 0. 1 + 1
Achiever 1+0. 1 + 1
Warroad. 1 +2. 1+1
Orono. 3 + 1. 1+4
St. Paul united. 2+2. 3+o
Jefferson 1+o. 0 + 1
Totino grace. 4 + 1. 1 + 1

10 games. 19 + 13= 32 pts. 12 + 14 = 26 pts

Looks like south kids and teams don't do too bad but of course are not allowed to be in the special top 30 SOS club......

[MNHockeyFan]

Agree that the SOS is bs....above hockey wannabe shows a top 10 kid who does not score in big games that is in this special scores club. Here are two dodge county kids not eligible for this list but what they have done vs good teams.....
Team. Rasmussen. Dodds
Benilde. 1 + 1. 0 + 1
South St. Paul. 2 + 4. 2 + 3
Centennial. 2 + 2. 2 +1
EdenPraire. 2+ 0. 1 + 1
Achiever 1+0. 1 + 1
Warroad. 1 +2. 1+1
Orono. 3 + 1. 1+4
St. Paul united. 2+2. 3+o
Jefferson 1+o. 0 + 1
Totino grace. 4 + 1. 1 + 1

10 games. 19 + 13= 32 pts. 12 + 14 = 26 pts

Looks like south kids and teams don't do too bad but of course are not allowed to be in the special top 30 SOS club......

Absolutely no offense, Maverick2000, but the players in your post are perfect examples of what I was originally getting at: that it's SO MUCH harder to score points against the State's top teams. In your post you cherry pick a bit in choosing your list of teams. If you include only the games DC has played against the Top 30 (KRACH combined) and especially the Top 20 the picture is very different. Here's what their stats show.

DC has played 21 games overall, with 6 against teams in Krach's Top 30 and 5 against teams in the Top 20. Not that many against the best competition, which is why DC's SOS rank is pulled down to #45 currently.

Against all teams Rasmussen has accumulated:
40 goals + 22 assists for 62 points, which are per game averages of:
1.9 goals + 1.1 assists for 3.0 PPG

But against Top 30 competition (6 games) the numbers drop to:
5 goals + 2 assists for 7 points, which are per game averages of:
0.8 goals + 0.3 assists for 1.2 PPG (40% of all games PPG)

And against Top 20 competition (5 games) the numbers drop to:
4 goals + 0 assists for 4 points, which are per game averages of:
0.8 goals + 0.0 assists for 0.8 PPG (27% of all games PPG)

Dodds numbers (all games) are:
25 goals + 26 assists = 51 points. Per game averages of:
1.2 goals + 1.2 assists = 2.4 PPG

Top 30 competition (6 games):
3 goals + 3 assists for 6 points. Per game averages of:
0.5 goals + 0.5 assists = 1.0 PPG (42% of all games PPG)

Top 20 competition (5 games):
1 goal + 2 assists for 3 points. Per game averages of:
0.2 goals + 0.4 assists = 0.6 PPG (25% of all games PPG)

PLEASE understand that in no way am I picking on these players. There is NO DOUBT that they are both tremendous players and should be included in any conversation about the best players in the State. My only point here is to show how typical it is for players' point production to drop off (considerably) as the level of competition rises. That is true for ALL PLAYERS! The only difference is that the competition level is more consistently tough for some players as opposed to others. So I completely disagree with the notion that "SOS is bs" and I think the above stats - which for other players would show a similar drop-off - helps prove the point.

[powerplayer]

You all have WAY too much time on your hands!!!

[sinbin]

I think a Girls High School Fantasy Hockey League is imminent.

[pepperpot]

http://www.youtube.com/watch?v=Ur0g-R7ZODY

[MNHockeyFan]

You all have WAY too much time on your hands!!!

Being retired helps some!