By Austin Tymins
After this weekend’s exciting round of 16 in the NCAA Men’s Lacrosse Championship, I have decided to give a preview of the upcoming quarterfinals while referencing a few new quantitative measures of team ability. In a previous post, I introduced a Pythagorean expectation model for college lacrosse. The first Pythagorean models were developed for use in baseball but have now been applied to a variety of sports. Instead of focusing on simple winning percentage, the Pythagorean model is reliant on scoring margin which has been proven to be more predictive. I also extended the approach to find team-specific exponents using the Smyth/Patriot method (Pythagenpat) which also extends the predictive ability of the model.
In my post published yesterday, I developed an iterative Simple Rating System for college lacrosse. This system allows me to disaggregate strength of schedule from margin of victory and thus allows a clearer picture of team skill. I have listed the results from both the Pythagorean and SRS approaches for completeness and comparison, but the reader should give much more weight to the SRS results. I will use take my previous research a step further to determine which team is favored in each matchup and by how much. This will rely on the Log5 method and the characteristics of the SRS.
Maryland (1) vs. Syracuse (8)
Maryland: Pyth WinPct 74.0%, SRS Rating 4.22
Syracuse: Pyth WinPct 79.3%, SRS Rating 4.60
Favorite: Syracuse 57.3% win (Pyth) and 52.4% (SRS)
Maryland entered the tournament as the one-seed after a very successful season in the challenging (but probably overrated based on SOS calculations) BIG 10 conference. However, both the Pythagorean and SRS slightly favor Syracuse in this matchup. The game will be most interesting when Syracuse’s 7th ranked offense takes on Maryland’s 6th ranked defense. On the other end, Syracuse is no slouch defensively as they rank as the 13th best scoring defense allowing only 8.50 goals per game.
The game could slow down considerably when Maryland has the ball as its highly efficient offense has the 4th lowest turnover rate in the country and the 5th best clear percentage. Maryland’s efficiency with the ball has been a critical factor in taking pressure off of their defense and 17th ranked goalie Kyle Bernlohr.
This game could very well be won with the faceoff or in man-down situations. Syracuse has the 3rd best faceoff percentage in the country at 62.6% while Maryland is only 22nd with 53.7%. A Log5 faceoff approximation suggests that Syracuse will therefore win 59.1% of faceoffs in this game—equivalent to 4.35 extra possessions. In man down, Maryland has the 3rd best man down defense at 75.0% while Syracuse has the 5th man down offense at 50.0%. Therefore, something will have to give when Maryland gets in the box.
Brown (5) vs. Navy
Brown: Pyth WinPct 90.5%, SRS Rating 8.04
Navy: Pyth WinPct 72.1%, SRS Rating 2.76
Favorite: Brown 85.7% win (Pyth) and 88.5% (SRS)
I truly can’t say enough about this Brown team. The fact that they are the 5 seed behind even Yale is an absolute farce and speaks to the need in lacrosse for better analytics. Brown led college lacrosse in winning percentage (88.2%) and scoring margin (8.18 per game, next closest 4.80). Even after adjusting for their slightly weaker schedule, Brown is far and away the nation’s best collegiate lacrosse team. The recency bias as a result of Brown’s first round exit against Harvard (woohoo!) in the Ivy League Championship is also very apparent. Though I haven’t yet calculated historical SRS ratings, I would venture to guess that this year’s Brown team has the potential to end the season as one of the all-time greats.
Brown has the most potent offensive attack in lacrosse this season averaging 16.76 goals per game. For comparison, Denver University, the next closest, is averaging approximately 3 goals fewer per game. Brown’s offensive attack is evenly spread as well with 3 of the country’s top 10 goal scorers in Dylan Molloy, Kylor Bellistri, and Henry Blynn. Without the ball, Brown has a top 20 defense and the nation’s top goalie in Jack Kelly who has saved 62.2% of shots, a full 3.5% better than the next closest goalie.
The faceoff aspect will also favor Brown as they enter the game third in the country in faceoff win percentage at 66.8%. Navy is no slouch either at 11th in the country and 58.9% however. Log5 suggests Brown will therefore win 58.4% of faceoffs and will receive 4.03 extra possessions in this game.
Brown’s undoing may be their inability to protect the ball at times, illustrated by their turnover rate which is 10th highest of 70 teams. Compounding this problem, Navy’s defense is the 2nd best in the country only allowing 7.13 goals per game. Even with these caveats, Brown’s victory and advancement to the semifinals seems all but assured. The NCAA has even scheduled the quarterfinal game at Brown Stadium in Providence, RI giving them the only home field advantage of the quarterfinal matchups.
Notre Dame (3) vs. North Carolina
Notre Dame: Pyth WinPct 75.7%, SRS Rating 4.56
North Carolina: Pyth WinPct 66.4%, SRS Rating 3.22
Favorite: Notre Dame 61.2% win (Pyth) and 58.3% (SRS)
This game is the second most lopsided matchup in the quarterfinals even though it’s basically 60/40 (Brown is just that good). Notre Dame has the 4th best scoring defense in the country only allowing 7.64 goals per game. They are also extremely adept at forcing turnovers at the 6th best rate in the country. In the cage, Notre Dame’s goalie Shane Doss has the 6th highest save percentage in the country.
Conversely, North Carolina is an average defensive team but can really make some noise on offense. They average 12.67 goals per game which is good enough to be 8th in the country. North Carolina will also have an edge on faceoffs as Notre Dame only ranks as an average faceoff team winning only 50.6% of its faceoffs while North Carolina is quite good and the 13th best in the country at 56.9%. Log5 predicts North Carolina will win 56.3% of faceoffs in this game and 3.03 extra possessions.
Keep an eye out for situations in which North Carolina goes man down. Notre Dame converts on man down opportunities 47.1% of the time which is good enough to place 8th in the country. North Carolina however is 7th best in the country at killing man down chances at 71.7%. For whatever reason, North Carolina only seems to play elite defense when entering forced switching situations on man down (relative to its peers).
Loyola (Md.) (7) vs. Towson
Loyola: Pyth WinPct 71.2%, SRS Rating 3.11
Towson: Pyth WinPct 79.8%, SRS Rating 3.00
Favorite: Towson 61.6% win (Pyth), Loyola 50.8% (SRS)
In this matchup, the Pythagorean win percentage favors Towson while SRS favors Loyola. Upon closer inspection, we can see that this difference occurs because SRS takes strength of schedule into account and these teams diverge in that respect. Towson played a schedule half a goal easier than the NCAA average while Loyola played one more difficult than average by half a goal.
Loyola and Towson are nearly identical in offensive output ranking 21st and 22nd respectively. However, these teams are two of the most elite defensively in college lacrosse. Towson gives up only 7.11 goals per game—a figure low enough to claim top spot defensively in college lacrosse. Loyola meanwhile has given up 8.06 per game which is still good enough to be the 9th ranked defense. While both teams are elite defenses, it appears they prefer to clamp down on defenses in longer possessions rather than by causing turnovers.
On the other side of the ball, Towson is the least turnover-prone team in the country giving the ball up only 10.39 times per game. Unfortunately for Loyola, faceoffs are not necessarily their specialty. Loyola wins 50.6% of faceoffs while Towson is actually below average at 49.1%. Loyola is expected to win 51.4% of faceoffs against Towson and receive .72 extra possessions.
These teams are so close in overall ability and and have nearly identical strengths and weaknesses. If you have to skip one game in the quarterfinals as a fan, I would suggest missing this one as it has the potential to be the lowest scoring and least exciting. In addition, these are two of the three weakest teams remaining in the field, and it is unlikely they make a deep run into the later rounds of the tournament.
Tournament Conclusion
I decided to take my SRS numbers and the distribution of scoring margins to run simulations on the tournament. The probability results from 10,000 Monte Carlo simulations are listed below:
Make Semi |
Make Final |
Win Final |
|
Brown |
88.49% |
63.21% |
47.77% |
Notre Dame |
58.32% |
34.58% |
12.92% |
Syracuse |
52.39% |
17.23% |
9.67% |
Maryland |
47.61% |
14.66% |
7.89% |
Loyola |
50.80% |
22.58% |
6.58% |
North Carolina |
41.68% |
21.26% |
6.33% |
Towson |
49.20% |
21.55% |
6.16% |
Navy |
11.51% |
4.58% |
2.05% |
As we can see, Brown has a nearly 50% chance of walking away with the NCAA Championship. Notre Dame’s chances are supported by the fact that they are playing in the weaker half of the bracket. Additionally, Syracuse has the second highest SRS in the field after Brown but has a lower probability of making the final than the entire other half of the bracket, including: Notre Dame, North Carolina, Loyola, and Towson.
Note: The quarterfinal games will take place on May 21 and 22 on ESPN2 and ESPNU.