Earlier this week I unveiled my college hockey ranking system and today I'd like to have some fun by producing odds for each team in the NCAA Tournament.
What I've done is taken the Log5 formula and my Pyth% from my previous article. Using these and the tournamnet bracket I'm able to get projected odds for any matchup in the NCAA Tournament. If you want more knowledge on the Log5 method, this link gives you far more information than you could possibly want. This practice is somewhat similar to what Ken Pomeroy, uber-famous college basketballs statistical guru, presented for the college basketball NCAA Tournament. The main difference being that Ken's run a simulation which I unfortunately don't possess the skills to do the same. Still, we can get odds for fun for the tournament.
|Seed||Team||Quarters||Semis||Finals||Championship||1 in …|
|9||St. Cloud State||0.347||0.060||0.032||0.009||111|
The data in the table represents each team's chance of making it to each round and is sorted by the teams most likely to win the national title, using only my college hockey rankings data.
The West Region
Minnesota Golden Gophers - Minnesota is the number one team in my system in addition to the NCAA and have, as the reward for being the top overall seed, the opportunity to play Robert Morris, the winner of the Atlantic Hockey Association's automatic bid. My system gives Minnesota a 96 percent chance of getting by the Colonials. They're good favorites against Notre Dame or St. Cloud State, but the combination of the draw and my system's rankings likely over-inflate Minnesota's chances. My system ranks the four teams in the Midwest region fairly low relative to the other regions. This gives Minnesota the potentially easiest route to the finals.
Robert Morris Colonials - RMU picked up the AHA autobid and are rewarded with the task of attempting to upset the best team in the nation. It probably doesn't help with last years first round exit still fresh in many Gophers minds -- both players and coaches -- so Robert Morris has their work cut out for them. Bad news for the Colonials: they've had 181 penalty kill situations faced, fourth most in the country. Worse news: If you look at power plays and penalty kills as a ratio -- in other words, power play chances divided by penalty kill chances -- Minnesota is second in the country. They draw about 28 percent more power plays then their opponents. That mean's Minnesota's top-20 power play rate in the nation becomes much more effective: they get a lot of chances, relative to their opponents, to work with the man advantage.
Notre Dame Fighting Irish - The Irish finished seventh in their inaugural Hockey East season, posting a 9-9-2 conference record. Their goal margin indicated that was about right. But a run to the Hockey East semifinals -- including a three-game series win against Boston College, the second-best team in my rankings -- helped dispel the notion that the Irish aren't a team to be reckoned with. Hockey East was a bloodbath all year long, easily the best conference in the nation. Notre Dame models themselves after some of the better puck-possession teams in the NHL, and it's worked. If we use Shots For Percentage, just the total shots on goal for a team divided by the total shots on goal in games they play in -- in other words, shots on goal / (shots on goal + opponent shots on goal) -- Notre Dame ranks fourth, controlling around 55 percent of the shots on net in their games. When you adjust it for schedule, it drops to sixth. Needless to say, though, Jeff Jackson's squad is going for a game type that has worked wonders for teams like the St. Louis Blues this year. (Ask the Toronto Maple Leafs about playing dominant puck-possession squads.)
St. Cloud State Huskies - The Huskies are a team my system hasn't ever ranked as high as their win totals, poll rankings, or KRACH rankings put them. The latest USCHO.com poll has SCSU ranked eighth in the nation and my system puts them closer to 20. The main reason is that their +32 goal margin shrinks to +19 after regressing their shooting rates. They've scored 132 goals this year on the back of a 11.6 percent shooting percentage. When regressed, based on their 1135 shots on goal, they sit around a 10.2 percent -- which is still the second-highest regressed shooting percentage in my database, just after Boston College's 10.7 percent -- and that drops their goal total down to 116. Regressing their opponent's shooting percentages shave off three goals allowed, dropping them to 97 goals allowed from their raw total of 100. Coupled with a solid-but-not-spectacular schedule faced, and the Huskies come out as a good team but not great.
The Midwest Region
Ferris State Bulldogs - Ferris is the first team in the NCAA Tournament that my system is very down on. They sit 31st in my rankings, a far cry from their sixth place spot in the USCHO.com poll and their fifth place spot in the KRACH rankings. The main reason for this is my system rates their schedule as below-average, facing the 44th toughest schedule in the country. The Bulldogs have faced the second-most penalty kill situations this year, being short handed 192 times this year. Conversely, they've had the 12th most power play chances this year at 170. What's interesting about Ferris is that their penalty killing has been so good -- sixth nationally -- that on a per-game basis, their power play is out producing their opponents extra man units. Ferris averages about 0.71 power play goals per game while they surrender about 0.63. Still, it's important to keep in mind their schedule they've faced, when looking at the extra-man data. Even the KRACH rankings have Ferris playing just the 47th toughest slate in the nation. If they take penalties in the NCAA tournament at their full-season rate, there's a good chance FSU could get bounced early.
Colgate Raiders - Colgate gets the 12th seed after finishing second in the ECAC standings. I currently have the Raiders ranked 20th nationally and their second place finish might've been rather tenuous. They outscored ECAC opponents by only 13 goals, fourth in the ECAC. Overall, their +5 goal margin puts them middle of the pack nationally. That said, my system gives them the edge over Ferris -- slightly, basically a 60-40 game -- for the third time in four games this season and to take on the winner of North Dakota-Wisconsin. Special teams should dictate the winner between Ferris and the 'Gate, as both teams are pretty even statistically. Their main differences lie with Ferris' great penalty killing and Colgate's ability to draw penalties. The Raiders have had 188 power plays this year, third-most nationally. Colgate also have a great ratio of penalties drawn to penalties taken -- 11th best, by my data, drawing about 11 percent more power plays than they let up. The Bulldogs from Big Rapids, as previously mentioned, tend to take a lot of penalties. This is Colgate's key to a first round win: keep Ferris State down a man as much as possible and hope you get a pair of power play goals.
Wisconsin Badgers - Like Ferris, and to a lesser extent St. Cloud State, Wisconsin has hovered lower in my ranking system than the polls or the KRACH rankings. I have them 27th, even after winning the Big Ten title. Wisconsin in my system rates as a solid team that played a fairly nondescript schedule. That's really what drags them down. If I were hand-ranking teams, Wisconsin would be much higher than 27th; I think they're dangerous and Joel Rumpel is a fantastic goalie. The Badgers great penalty kill has masked the fact that they take a good deal more penalties than their opponents, which is always a dangerous game to play. Still, it's that edge that the Badgers walk that makes their veteran club as good as they are.
North Dakota - North Dakota, like this whole region, sit lower in my rankings than they do in the polls. They're 13th in the latest USCHO.com poll but just 22nd in my rankings system which, like Wisconsin, is mostly schedule-related. Finishing third place in the inaugural NCHC Frozen Faceoff and making its 12th consecutive NCAA Tournament, North Dakota is certainly a team no one wants to face with seemingly every defenseman on the roster an NHL draft pick, they're an underrated offensive team. They're averaging 3.05 goals per game which is a solid number coupled with that blue line.
The East Region
Union Dutchmen - Union sits 12th in my ranking system, while currently topping the USCHO.com poll. Even the KRACH rankings have moved Union over Boston College for the No. 2 spot. I like Union, subjectively, and my system thinks they're very, very good. They sit lower in my rankings for a couple reasons: they've faced an average schedule and regressing hurts their overall goal margin. They've scored 140 and allowed 79, good for a +61 goal margin, third-best nationally. When regressed, their goal margin is +39, which is actually second best in my system. It's that their schedule has been average, so there's no residual boost there that a Boston College (+37 regressed goal margin, ninth-toughest schedule) or a Minnesota State (+37 regressed goal margin, 22nd toughest schedule) receive. That still means Union's very, very good -- great, maybe -- but in this system leaves them closer to the back-half of the top ten then the top three.
Vermont Catamounts - My system gives the Catamounts the third-best odds of winning the tournament, about 10%. This is probably the second thing people will notice about my NCAA Tournament odds -- depending on how dominant you think Minnesota is, their odds of winning it all are certainly the first eye-popping thing about them. Vermont's goal margins are unspectacular -- +19 raw, +15 regressed -- but they've played a tough, tough schedule. I've got their schedule ranked as the third-toughest in the nation, akin to playing basically a top 20 team every single night. It's this fact that takes Vermont from an above-average team to one that's sat inside my top five for a while now. Using shots as a proxy for possession, they rank 10th in my database in puck possession. Without adjustment, they rank 16th in shots on goal against, allowing just under 28 per game. When you really study that ranking with their schedule in mind, it's very impressive. They're behind some teams like St. Lawrence, Air Force, Michigan Tech and Bowling Green which aren't the strongest of clubs or teams that played soft schedules. In that light, Vermont starts to soar.
Quinnipiac Bobcats - the Bobcats are a fantastic hockey club, sitting eighth in my rankings and seventh in the latest USCHO.com poll. (Hey, they agree on a team!) What makes the Bobcats so hard to defeat is they kill your special teams. They draw the most power plays in the country (211) and even though they take a lot themselves -- 172, 12th most nationally -- the ratio that they draw power plays versus penalty kills is the fourth best mark in my database. With this in mind, they then pummel you with shots. Their shot margin is tops in the nation by over two full shots, where they average 12 shots more than their opponents. Even after adjusting for what I have as an average-at-best schedule, Quinnipiac controls 61.5 percent of all shots on goal in their games. The second best team sits at just 56.5 percent.
Providence Friars - Providence is going to have to avoid the penalty box against Quinnipiac, This is my second-most anticipated game because it matches two very good teams in the first round. I can see either team riding their strengths to a Frozen Four without squinting too hard. Providence is 10th in the nation in shots per game and do a decent enough job suppressing them. When they do relinquish a shot, Jon Gillies is a great backstop to have. If Gillies takes over, Providence will be a tough out in the NCAA Tournament.
The Northeast Region
Boston College Eagles - The Eagles were surely disappointed being ousted from the Hockey East tournament by Notre Dame, but they still get the no. 2 overall seed. BC will score their goals, that's a given. They average over four per game, no one else above 3.7. They dominate on the PK, teams scoring at only a 9.5 percent clip, tops in the country. BC could be better at evening up the penalty situations -- their 0.85 power-play-to-penalty-kill ratio is fifth worst in the nation -- but they've masked it with their fantastic penalty kill work. Thatcher Demko has been so good between the pipes, you forget he's just a freshman. In about 40 percent of the available minutes, Brian Billett has been every bit as good as Demko, though. BC's got a very solid duo between the pipes. BC's goal margin is tops in the country and after regression it sits third and that comes against the ninth hardest schedule in my database. BC is fantastic and poised to go deep in the tournament, though they appear to be in the tougher half of the bracket.
Denver Pioneers - The Pios, winners of the inaugural NCHC Frozen Faceoff, sneaked into the back of the USCHO.com poll, clocking in at No. 17. They rank just 25th in my system, though. My regression knocks their goals scored down to 106 from 110 but their goals against is where they drop off. They're regressed goals allowed is 102, ten more than their raw goals allowed total. This puts them at just +4 regressed goal margin. They do play a bit-above-average schedule, but that puts them closer to t he middle of the pack than a NCAA Tournament team.
UMass Lowell River Hawks - The River Hawks have the fourth-best odds to win it all. Jon Gillies is great, but if I had to choose a single goalie in this tournament to pin my hopes on, it's Lowell's Connor Hellebuyck. The sophomore followed up his fantastic freshman campaign, during which he posted a 0.952 save percentage, with a 0.943 save percentage this year. It's just been a matter of keeping him on the ice, but Hellebuyck has posted better rates the last two years than any other goalie in the nation. Lowell aren't possession monsters in my database, but they rate so highly because they're just very good and did it against a very good schedule. They draw more penalties than they take, their power play clicks at the 12th best mark in the nation, though their penalty killing leaves a lot to be desired. Their 81.3 percent penalty kill doesn't come into play that much, as they've faced just 139 penalty kills, 15th fewest in the country. Minimizing your time in the box and having arguably the best goalie in the nation will go a long way. Oh, they lead the nation in goals allowed per game at just 1.85. Lowell is dangerous.
Minnesota State Mavericks - I mentioned earlier that Providence-Quinnipiac is my second most anticipated game. The one I'm looking forward to this Lowell-Minnesota State contest. Why? Well, the Mavericks are the second best team in shot margin per game, outshooting opponents by over 10 shots per game. Even after i adjust for schedule, they're second in the country, controlling over 56 percent of the shots in their games. They achieve this by the same model as Quinnipiac: draw a ton of power plays (second most nationally) and don't take a ton yourself (165 penalty kills this year, 22nd in the country). Unlike other teams in my system, their shooting rates are close to the national average, so the regression doesn't hit them hard. Their goal margin is +36, seventh nationally, and their regressed goal margin is +37, fourth best in my database. Their schedule only rates as the 46th toughest, and that's what keeps Minnesota State them in the back-half of my top 20 and not bordering the top then or higher. I'll be interested to see if they dominate the shot totals as their wont to do, and if they do, how many saves Hellebuyck will have to make for UMass-Lowell. Minnesota State leads the nation in shots on goal per game, slightly ahead of Robert Morris.
There's various ways you can quantify the toughest region, I'll go with the easier way: averaging the ratings of each team in the region using my system. The East region features teams with an average Pyth% of 0.747 which is on par with the seventh through ninth place in my rankings. The Northeast is the second toughest at 0.709, roughly equal to the 11-13th best teams. The West is next with an average Pyth% of 0.631, on par with teams in the back-half of my top 20. Finally, the Midwest features two teams that are better subjectively than my system makes them out to be, thus giving the Midwest an average Pyth% of 0.564 which is equal to 24-26th in my system
I think the tournament committee got the seeding pretty much spot on, subjectively. My system thinks Vermont is woefully underseeded while Ferris and Wisconsin are too high. My dark horse for the tournament, though, would be UMass-Lowell. I like them to make a good run.
The Frozen Four
My system gives the best odds of a Frozen Four consisting off Minnesota (75.9 percent chance of reaching the Frozen Four), Colgate (32 percent), Vermont (41.2 percent) and Boston College (55.8 percent). The most likely game in the finals would be Minnesota vs. Boston College, which still only has around a 25 percent chance of happening. I should mention that when I tracked this data last year, though I didn't do it publicly, my system did give Yale the lowest odds of winning the tournament among teams that weren't from the AHA. That's why they play the games.