Nationals 2017 was the NCDA’s 199th ranked event since we began accurate records in the 2011 season. We had 57 matches at Nationals 2017, in a 286 ranked match season. There were 304 event items in all, and 32 active member teams this year. After the Championship, this is where the League stands at the end of the 2017 Season:
The Dust Settles After Nationals
Of the 57 matches played at Nationals, only 8 were upsets (85.96% success rate). The largest being the UWP def BGSU 2-1 upset, but it was still close with a 3.733 rating gap. 7 occured on Saturday and only 1 occured on Sunday, WKU def CSU with a tiny 1 point rating gap.
GVSU was the top rated team with a 56.623 Gonzalez rating going into Nationals, and was seeded #1 in the Tournament Bracket with a 57.213 rating.
By the time the 2017 Championship rolled around, it was #1 GVSU (57.684) vs #2 CMU (55.464) with a Rating Gap of only 2.220. The average Rating Gap for the Championship has been just 2.222 since 2011. To put that into comparison, the average rating gap for a historical upset is 2.677. Every Championship matchup has featured similarly rated teams, but there hasn’t been a technical upset…yet.
The past three championships have been fairly consistent statistically, with GVSU #1 vs CMU #2. The rating gap averages 2.236, ranging in the very narrow band of 2.185-2.303. The historical data shows that an upset is a lot more likely given this rating exchange, but an upset just hasn’t occurred yet. The probability appears to be closing.
By defeating #2 ranked CMU (55.464) in the 2017 Championship, GVSU finishes the season with the current highest Gonzalez rating of all time, 59.240. The past record (58.206) was set by GV at the end of the 2016 Season. This is pretty much the standard procedure given the past results. The National Champions usually set the highest rating record, but this is certainly helped even further by a couple of factors: GVSU has won 5 consecutive championships, and has largely gone undefeated in the double strength Nationals matches.
Also, Even More Observation
The Gonzalez system is great at what it does. It predicts any given team’s rating with a high degree of accuracy, exactly when we need it: the Tournament Bracket. The following table shows the National Championships and three different types of seeds those teams had, see the description after for more info:
|‘Ship||Win Team||Seed / Rank||Loss Team||Seed / Rank|
|Published Seed||Pre Ship||Pre Bracket||Published Seed||Pre Ship||Pre Bracket|
|2011||CMU 51.374||1||1||1||GVSU 48.268||3||2||5|
|2012||SVSU 51.287||3||1||2||UK 49.925||5||3||5|
|2013||GVSU 52.800||6||1||5||MSU 51.281||5||2||6|
|2014||GVSU 55.843||1||1||1||SVSU 53.987||2||2||2|
|2015||GVSU 56.667||1||1||1||CMU 54.363||2||2||2|
|2016||GVSU 57.140||1||1||1||CMU 54.956||2||2||2|
|2017||GVSU 57.684||1||1||1||CMU 55.464||2||2||2|
The “Published Seed” notes the seed given to that team under the current system used at the time.
- 2011, 2012, 2013 used different variations of the NHL/Champ system which had different values based on Saturday performance.
- 2014, 2015, 2016 used a combo algorithm which included the Gonzalez System as one of it’s parts.
- 2017 used the Gonzalez System exclusively.
The “Pre Ship” number notes what Gonzalez rank that team had going into the Championship match. Gonzalez standings are updated after each match, so while GVSU was #5 going into the bracket, by the Championship rolled their rating climbed to #1. Impressively, each match was 1 vs 2… with the exception of UK in 2012, who held the 3rd rating at the time. GVSU narrowly held #2 after their loss to SVSU in the final four, just a tiny 0.361 ahead of UK.
The “Pre Bracket” number indicates what Gonzalez rank that team had at the start of the Tournament Bracket play, what they began Sunday with. There were more “shakeups” in 2011, 2012, 2013 but the Gonzalez system is adaptive. The 2013 MSU team might have been rated #6 in Gonzalez going into Sunday, but they won against UWP, made a big upset on SVSU, and later upset Kent. That performance set them up as the #2 rating narrowly behind GVSU. That rating gap was a tiny 1.519, the shallowest rating gap of all the National Championships.
In 2014, we started using the Gonzalez system as part of our seeding for the tournament bracket. In the four Nationals since then, the top two seeds of the Bracket have been sent to the Championship match.
Standings going into 2018
At the end of every season, we revert some of a team’s rating (currently 75%) to the League Average. The League Average (Mean) takes all the team ratings of every team that played a ranked match this season. Using GVSU as an example:
End 2017 Rating: 59.240 * 0.75 = 44.430
End 2017 League Mean: 41.082 * 0.25 = 10.271
Begin 2018 Rating: 44.430 + 10.271 = 54.701
The following table shows the ratings going into the 2018 Season.
|Rank||Team||End 2017||Begin 2018|
Overall, we now number 1250 total ranked matches, with only 236 technical upsets. Our overall system success rate is 81.12%.
|2017 League Mean Rating*||41.082|
|2017 Total Ranked Matches||286|
|2017 Success Rate||76.57%|
|2017 Participating Teams*||31|
*Excluding UNT, who was the only active, dues paying Member Team not to play a ranked match this season.
That’ll do it for this season. It has been a pleasure being able to work with this data over the season, I’ve enjoyed this over the years.
Keep playing dodgeball.