Interconference Play via Ranks, Rivalries & Ratings – Intraconference Sims
INTERCONFERENCE – BY RANKS
The graphic below shows a gold strip representing the national rank (from 1 to 124 top to bottom) for every team in a conference (or Independent) as it is reported by Inside Hitter, Massey, and T100. The conferences are arranged in order of their T100 average rating, going left to right.
INTERCONFERENCE – BY HEAD TO HEAD (RIVALRY)
In an attempt to see which conferences are competing against each other, and the degree to which its teams are being successful by winning matches, the following matrix offers every conference’s win% (color coded for success rate) against every other. The single value below its win rate is the number of interconference matches played through last Sunday, 3/3. (The NACC and the CCIW have played 44 matches to date with the CCIW winning 70% of these – The most across the spectrum.) Blacked out are conferences whose teams have not played, and grayed out are both conference matches or those conferences whose teams have played just once. I considered graying out those with just two “head to heads” until noticing each was 50% from winning and losing 1 match apiece.
INTERCONFERENCE – BY RATING (T100)
Rather than create 13 dot plots across the whole landscape by using ratings, I think it is more useful to see them all, 3 conferences at a time, by way of the region to which they belong. The independents are dispersed across all 4 regions which is why there is a 4th dot plot in each region to compare. The “best” 6 teams for every region according to T100 ratings have been labeled. It is these 24 teams who will be eligible for Pool B (3) & Pool C Bids (2) decided by a NCAA National Championship Committee. (Based on their version of the “best 6” per region, of course.)
INTRACONFERENCE – BY SIMULATION
Using the T100 ratings as the basis to express the win probability for all conference matches remaining on the schedule, and then playing these matches 100 times in a simulation, provides the average place each team finished, the average number of wins they achieved (and losses), but most importantly, the frequency of the place each ended up across all 100 of the seasons. There are 549 conference matches this year. Roughly 30% have been already played, so we know those are either a 100% or 0% win probability depending on who won. The graphic shown below is an example for just one conference’s distribution’s graphic (UVC) showing how often a team finished in every place of its conference standings. Below this graphic is a link to every conference’s simulation results so you can pick and choose the ones you are interested in seeing.
Each solid represents a frequency polygon for the distribution of a team’s finishing place in 100 sims. Those with very high peaks are more certain to finish in these places. For example, Elmira ending up 8th in roughly 85 simulations suggests an 85% likelihood for that to happen. The wide nature of the shorter solids reflects an uncertainty for these teams’ forecasts. For example, MIT looks to have its peak at 4th place in its distribution, but only about one-third of the way up. i.e. it took exactly 4th place in 36 of 100 sims. Its frequency polygon ranges from 1st place (4 times) to 7th place (2 times). Its few conference winning seasons hidden behind the Fisher, Vassar, and NYU distribution, and more often than not, resulted from a tie-break scenario. This being a consequence for the current state of Vassar already at 5W-1L having just 1 conference match against New Paltz remaining to be played.*
If Vassar had beaten Fisher last Saturday on 3/2, its orange peak would have been very near 99, because New Paltz winning 6 in a row would have been the only scenario for Vassar not winning the regular season title under that condition. (p = .9%) Win-out scenarios are still available for NYU, MIT, and New Paltz to get to 6-1 and guarantee a regular season conference championship in its present state, however. This because each of these 3 with one loss still has St. John Fisher on their schedule, the only remaining team in the UVC not having a loss.
For the above graphic, I superimposed a wire frame of same with 70% transparency to see the solids being shadowed behind. I didn’t do this for the conferences linked below because I pasted the frequency data table from which it was created so it can be seen.
Not exactly sure where Conference Corner will go from here, since I emptied the chamber on my best ideas for all 12 above. I was going to do 3 different conferences per week, but by the time we’d get to the last 3 in week #4, the season will be so close to being done. I, like all of you, would be chompin’ at the bit for conference seeding and tournament projections by then. With almost a third of the conference matches complete, it seems the right time for simulating all of them right now.
Sometimes I come across really outlandish facts doing research which I believe are worthy of mention. Like when I quoted Crocodile Dundee and Top Gun last week, only to find out they were #1 & #2 at the box in the same year with the same production company. Those that read my posts recognize this from my writing style, I think. Below is one, even related to the volleyball topic of the day – Conference Scheduling. Do Coaches do it? AD’s do it? Commissioner’s do it? Are there any rules governing those who do it? Now I am on to a NIKE vibe – Just Do It!
* The very fact Vassar played in 6 of the first 9 UVC conference matches and is scheduled for 1 of 19 remaining to finish the season is a very peculiar thing on March 2nd, with more than a month remaining in the season. To this point, almost a week later (March 7th), across all conferences, 70% of conference matches are yet to be played. There is no other team of the 106 belonging to conferences who has even played half of its conference schedule, yet. This while Vassar has played its 4th, 5th, and 6th of 7. (The significance of 4th, 5th, and 6th, is that each represents more than half when none of the other 105 teams has even played half?)