It has been a while since I told you what the metrics below really mean:
[1] The difference between any two team’s ratings is the number of points I would expect the higher one, on average, to defeat the lower one playing a set to 25 points. Home court is a value pretty close to about 1.0 which in practical terms would make a couple of equal teams on a neutral court deemed to be 50-50 turn into the home team about 60-40. Of course, it isn’t always a 10% boost for the home favorite because as the two teams gravitate farther apart, the extra point for being home has a relative lesser effect than 10%. Once you know the probability for any team to win a set based on this difference, it isn’t too difficult to figure out being the first to win 3. That is essentially the probability to win a match seen for every contest coming up in week #5 at the end of this post.
[2] The Probability for a Natty Bid is the sum of two distinct metrics. First is the chance for any team to win their conference and secure an automatic. Then I have to add on a conditional probability for every excellent team to get an at-large, under the assumption they didn’t win their conference. That second one is too big a pain to do properly with little “bang for the buck,” so I just ballpark it and make sure they add up to 5.0 so that the sum of blue metrics below adds up to precisely 19.0 – The number of teams who will be in the tournament.
[3] The average team has 10.0 points, so for a landscape of 129 teams it will add up to about 1,290 points making it so the median (65th best team) should be pretty close to that 10.0 value. Just looked down to see 10.41, a little larger than it was a month ago. That makes sense because what starts out as a mound shaped symmetric distribution spread about 10.0 with variability at s= 4.0 shifts to one skewed, moving closer to a variability s=4.6ish by the end of 14 weeks. What this basically means is at the end of the season the highest value will be pretty close to 20.0 while the lowest will be close to 0. If they played, and I hope they wouldn’t, the forecasted set score would be 25-5, a 20-point differential. Sure, the best in the land could likely win by a larger margin than that, but they wouldn’t, even if they could. It would be a chance for their backups to their backups to get some photo ops and populate the stat sheet.
Since I started doing this a few years ago there has been a distinct shift for teams not running up scores when games are in the PS region of the Win-Matrix. In 2023 more than 25% of forecasted blow-outs in January (blow-out defined as p(w) by favorite > 95%) had prohibitive favorites winning by at least 36 points in a sweep. i.e. By average margins beyond 25-13 per set. The incidence of that is less than one-third that (8%) in the Januarys of the 2024 & 2025 season. (January is the month representing near 100% non-conference play indicating the type of matches scheduled when teams have more control over who they play and when.) For those not reading my Win-Matrix post from Friday, you’ll have to go back to see what that PS region means. LOL – Shameless!
Lots of jockeying in the top 15 with the Frog-Jump Invite and the Cougar Classic this weekend. Throw in some other hotly contested matches in the Midwest and it was a volleyball junkie’s dream weekend, and not just because nobody watches the NFL Pro Bowl!
The first fixin’ is Conferences+. You can get a taste of the order of conference strength on the left side using a weighted formula from the top 6 teams’ rating metric. Its color shows projected bid winners and those left out on the bubble’s edge. I throw in some “Bubbleicious” match scores from last week that were between teams less than 50% to win their conference and among the top 25. (Most potential to become bubble relevant if they can get it going soon, but not so much as to win a conference title.) Scan right from each conference name to see the 5 best ranked teams according to the T100.
The 2nd fixin’ is my attempt at putting together a bracket based on the 19 teams chosen above. Admittedly I am running blind in this exercise not knowing where the NPI stops and where the committee begins. The T100 should mimic the NPI (I hope), so it is where/how the committee interjects itself that has me a little befuddled. My latest quandary, “Do I offer a regional host spot to a team that doesn’t win its conference championship, even if the T100 (NPI?) ranked order were to tell me to?” The NCAA reveres conference champs enough to invite at least a half dozen who won’t be among the best 20 in the land, something in principle for which I don’t necessarily have an objection. Might it follow for this respect to transcend hosting a regional except in the instance an Independent is worthy? Don’t know for sure about that. If hosting a regional makes no difference as pertaining to having won a conference championship the week before, then I breathlessly anticipate the day none of 4 regional hosts is their conference champion, and the uproar which would ensue over how the NPI could do such a thing. LOL
Below is the 3rd fixin’ and the 3rd installment of matchups for the week. I have been casually checking in on them to see about 5 in 6 have accurately identified the winner posted a week in advance. Remember, the real time probability for a Saturday or Sunday match certainly shifts based on performances earlier in the week. Those shifts take the win forecast accuracy closer to 7 in 8 in real time. (Only 5 or 6 of the 140+ matches this week would need to flip-flop for this to happen – No doubt they will.) For those who checked out the Win-Matrix post on Friday, you saw more than 90% above the main diagonal indicating the favorites had won those matches. Exactly, after I run next week’s T100, I suspect more than 90% of the previous week’s games will be above the main diagonal, too, even if only 87% were correct. That’s how this thing works. It learns from its mistakes and recalibrates to make it retrospectively a better order than it was the previous week. Doing this for 14 weeks keeps it snuggling closer and closer to the absolute truth for whatever that ranked order is at any given moment. And even then, there is a little thing known as natural variability which will keep it from ever being perfect in predicting winners. If any of those couple sentences intrigue you, last Friday’s post about Win-Matrices is still out there waiting for a read. (Another shameless plug! LOL)
And I got it done before Monday, Ramius! How about that raise? LOL