NCHC Prediction 2016-17

nchc_2017_pre

Welcome back, NaCoHoCo fans. It’s been a long offseason – partying for some of us (North Dakota, Denver), pangs for others (Omaha, Miami), pleading for others still (Arizona State, Mankato). But the days are getting shorter, the lattes are getting pumpkin spicier and the rev of distant zamboni motors grows and grows. Let’s talk college hockey, eh?

It’s destined to be another exciting year in the National Collegiate Hockey Conference. Sure, it’s tough to beat a national championship season. How about another one? And sure, the conference lost a lot of talent to graduation and pro contracts, but there are incredibly talented goaltenders, forwards and defenders returning for another year of play. How will it all shake out?

The writers and experts have had their say, and once again, I’m taking a stab using shot data from last season to predict how the conference race might unfold this year.

We haven’t always agreed, these hockey writers and me (me being “data”, I mean). In the last couple of years I’ve been doing this, the data has uncovered some interesting trends that bucked conventional wisdom. Two years ago, this statistical model helped predict an insurgent UNO Maverick team (that ultimately made the Frozen Four). Last year, it pegged Denver as the conference favorite – wrong, but the Pios did make the Frozen Four.

Sometimes data finds the trends that we don’t normally see. Other times, it simply confirms what everyone already knows. This year, as we’re about to see, is one of those years.

I’ve collected individual-level data on all NCHC players from 2015-16, primarily goals, shots, shot%, save% and a derived possession-share (individual shots/all shots). That data is readily available thanks to better tracking by the NCHC and more in-depth shot statistics compiled by College Hockey News. (P.S. to NCHC’s marketing team: I love the new website and data page – huge improvements. Someone’s been reading?)

As is tradition, we’re going to adjust sh%, sv% and possession for each team based on what we know about roster changes, particularly about who is returning and who has left. To do this, I have to make some assumptions about players and teams. I’ll try to keep these as safe as possible:

Continue reading

2016 NCHC Prediction

NCHC_model2016

Here it comes – Year Three of the NCHC. Once again, the eight-team conference looks to be the toughest and most talented in the nation. Most polls put 4-5 teams in the national Top 10, and, having sent 75% of the conference to the NCAA Tournament in March, the NCHC looks poised to return at least half the conference this year.

As the 2016 NCAA hockey season gets underway, it’s time to predict the final 2016 standings of the National Collegiate Hockey Conference. Per usual, I will be doing this using only actual statistical data based on each team’s past performance.

This only gets tougher as nearly every team in the conference has proven a national contender over the last few years. Already the teams are so talented that the marginal differences between each team are so slim – any team could beat any other team on any given night (yes, even CC).

Last year when I did this, the statistical method of predicting did slightly better than all of the NCHC hockey journalists. There’s no guarantee I will do as well this year though, so the better my odds I made a few adjustments to the model to try and get an even more accurate prediction.

I’ve collected individual-level data on all NCHC players from 2014-15, primarily goals, shots, shot%, save% and a derived possession-share (individual shots/all shots). That data was much more readily available thanks to better tracking by the NCHC and more in-depth shot statistics compiled by College Hockey News.

As we did last year, we’re going to adjust sh%, sv% and possession for each team based on what we know about roster changes, particularly about who is returning and who has left. To do this, I have to make some assumptions about players and teams. I’ll try to keep these as safe as possible:

Continue reading

NCHC Model: Adjustments 2016

2015_CC_NCHC

Welcome to a long thorough analysis of my analysis – the NCHC model for predicting final standings. Before last season, I created a model to predict the NCHC final standings. In this post, I’m re-examining my methods and assessing their validity. Honestly, I thought I did OK last year, considering I did better than most of the expert writers and media. However, my model didn’t exactly nail the final standings, so there’s room for improvement. Today I’ll take a look at the components of the model and whether they worked as intended. Later this week, I’ll bring it all together for a 2016 prediction.

Be warned – this post is methods heavy. If you have an interest in NCAA hockey analytics, read on. If not, turn back now, and come back later in the week for my stats-based NCHC predictions. Still here? Ok, here we go.

First of all, kudos to the National Collegiate Hockey Conference for providing full-season and in-conference shot statistics for 2014-15. It’s a good start and a huge improvement over last year.

If you’d like to get familiar with the theory behind this model, I suggest reading my post from last year in which I created the NCHC model. Also, before we get started, the usual disclaimers: this analysis uses NCHC data taken from NCHC official records, and only considers intra-conference play during the regular season. Non-conference games, NCHC tournament games, and NCAA tournament games are not included.

2013-14 vs. 2014-15

Team '14 Sh% '15 Sh% Δ '14 Sv% '15 Sv% Δ '14 Poss. '15 Poss.  Δ
Colorado College 7.74% 6.81% -0.93% 89.39% 88.65% -0.74% 48.73% 42.10% -6.63%
Denver 9.08% 10.49% 1.41% 92.87% 90.65% -2.22% 44.81% 50.90% 6.09%
Miami 7.60% 8.99% 1.39% 88.90% 90.90% 2.00% 50.55% 55.20% 4.65%
Minnesota-Duluth 8.83% 8.95% 0.12% 89.83% 91.26% 1.43% 53.17% 53.00% -0.17%
Omaha 9.87% 10.15% 0.28% 89.22% 92.63% 3.41% 56.49% 46.50% -9.99%
North Dakota 11.06% 9.61% -1.45% 91.41% 92.75% 1.34% 48.38% 49.70% -1.36%
St. Cloud State 12.46% 9.18% -3.28% 91.14% 91.37% 0.23% 49.15% 53.70% 4.55%
Western Michigan 10.06% 7.38% -2.68% 90.11% 89.51% -0.60% 48.45% 49.80% 1.35%
NCHC Total 9.58% 9.02% -0.56% 90.42% 90.98% 0.56% 50.00% 50.00% 0.00%

The adage says the best predictor of future performance is past performance. Looking at this table, that holds. St. Cloud lost some shooting prowess, and Omaha had better goaltending, but not much jumps out otherwise. Performance was fairly steady across the three categories from ’14 to ’15 – except possession. Wild swings there, eh? Omaha lost nearly 10 percentage points in possession share. Meanwhile, Denver and Miami saw great improvements, which was a big reason for their good finishes. I’ll take a look at what might be driving possession swings a little later. For now, when comparing ’14 and ’15 performance, you can group the eight teams into four categories:

Continue reading

2015 NCHC Model Check-in

massa

Now that each NCHC team has played at least one league series, let’s look again at the predictive model we developed. In September, before the season, I made the following predictions:

2015 NCHC Prediction

Team '15 Shot% '15 Save%  '15 Poss% '15 Points* '15 Finish
North Dakota 11.26%  92.03% 49.50% 48 1
Nebraska Omaha 10.82% 90.45% 49.50% 41 2
Minnesota Duluth 9.05% 89.35% 54.50% 36 3
St. Cloud State 10.02% 90.21% 49.00% 36 3
Denver 9.44% 91.45% 47.50% 36 3
Miami 7.81% 90.90% 52.75% 34 6
Western Michigan 8.68% 90.45% 48.50% 30 7
Colorado College 8.31% 89.72% 48.75% 27 8
ALL NCHC 9.43% 90.57% 50.00% 288
*All 14-15 expected points +-4.85

Six week into the season, how are these predictions holding up? To be honest, it’s way too early to assess, but it’s not too early to estimate when evaluation becomes appropriate.

But just for fun, let’s look briefly at how each team has performed since the start of the season, especially in terms of the model’s three input metrics. Note, though,  this looks at all games, not just the NCHC. This is just to get a better general idea about how the teams are playing, as of Nov. 9, 2014:

hlogo-CC  Colorado College

Record Goals For Goals Against Shots For Shots Against Shot% Save%  Poss%
2-6-0 15 35 236 268 6.36% 86.94% 46.8%

Well, probably a bad place to start, for us and for CC. Low on shot percentage, save percentage, and possession share. They’re doing even worse than I predicted in all three areas so far. We knew this was going to be a rebuilding season, but it could be a long rebuilding season for Tigers fans, especially in NCHC play. I predicted CC would finish at the bottom of the conference, and I still feel pretty secure in that.

hlogo_DEN  Denver

Record Goals For Goals Against Shots For Shots Against Shot% Save%  Poss%
5-3-0 22 21 267 226 8.24% 90.71% 54.2%
Denver is lagging a bit in shot and save percentage so far, but they’re more than making up for that in a possession turnaround. All that returning experience is paying off. I still think they’ll make a run at home ice, but they can’t keep giving up 7- and 10-goal weekends like they have to Duluth and Western Michigan, respectively.

Continue reading

A 2015 NCHC Prediction

My 2015 NCHC prediction says North Dakota is the team to beat.

If you’re just tuning in, we’re working toward making a 2015 NCHC prediction. So far, we’ve developed a model that explains the primary inputs into how teams earn points. Then, we tested the model on previous seasons and other leagues. Everything is holding up, and we have a good idea of the strengths and weaknesses of the model. Now we get to the truly hard part – generating and estimation of how well we expect teams to perform this year on three measures – shooting percentage, save percentage, and possession time.

I’ve collected individual-level data on all NCHC players from 2013-14, including goals, shots, sh%, sv% and a derived possession-share (individual shots/all shots). Before I got any further, I just want to say: DENVER and COLORADO COLLEGE, GET YOUR ACT TOGETHER. All other schools provide individual-level conference data on goals, shots, blocks, and a host of other measures. CC links to an outside website, and Denver offers one PDF on season data. I really wish the league office would get some standards around this. But I’ll move on.

We’re going to adjust sh%, sv% and possession for each team based on what we know about roster changes, particularly about who is returning and who has left. To do this, I have to make some assumptions about players and teams. I’ll try to keep these as safe as possible:

Continue reading

2015 NCHC Prediction: Testing the Model

How will Miami and Duluth fare in a 2015 NCHC prediction?

In Part One earlier this week, I attempted to devise a model to predict last season’s NCHC standings, using puck possession (shots) and bounces (shot and save percentage) as the input variables. Not only did the formula reasonably predict the regular-season standings, but it appeared to express a strong fit to the data. Once again, that model is:

 \text{POINTS} = b + b_{1}(\frac{g_{f}}{sh_{f}})+b_{2}(\frac{sh_{a}-g_{a}}{sh_{a}})+b_{3}(\frac{sh_{f}}{sh_{a}})+e

However, I only have one season of NCHC data to work with. It’s possible this is an anomaly, or it’s possible I fashioned a model to fit the standings that we already knew. What I really need is another couple of data sets (aka a few more NCHC seasons) to test this exact model. I don’t have that for the NCHC because there has only been one season so far, but I have something close.

Six of the eight NCHC teams played at least three seasons in the WHCA, so I applied this framework to those three previous seasons, building a regression model that considers all three seasons at once to get a model that is significant beyond the .0001 level. The results:

Continue reading