Subtle Changes & SOS

Loyal readers know I love scheduling. I love the art of it. I love the process. I’m sure the coaches who actually have to do the work are less excited about it then I am, but seeing how programs put together a schedule that hopefully gives them a chance at an at-large bid is a joy to me. An offshoot of scheduling and one of the more important selection criteria is strength of schedule or SOS. For this article, I wanted to dive into SOS a bit more and how it can be positively and negatively impacted by scheduling. I also thought it would be a good idea to pick on someone’s 2023 schedule and see how subtle changes would have impacted their SOS. Let’s see…who is going to go under the microscope? I could do Chapman but I just met Coach Mary Cahill in person and she asked me nicely not to pick on her so I won’t do the Panthers. Oh! A jungle cat is pretty cool! Okay, let’s switch over and do the Leopards of La Verne!

I like La Verne as an example because they were a very good team this past season, made the tournament as an at-large and then had a great run to the regional final. What maybe is forgotten is that their SOS was only 0.541. I just did my at-large criteria article and they had the lowest SOS of any of the at-large teams (by a good amount). The next lowest was St Olaf at 0.560. We know that it’s harder to put together a 0.600 level SOS in the West because we are so isolated but that is still a pretty low number for a program that will be looking to go back to the NCAA Tournament next year.

(How hard is it to put together a 0.600 SOS in the West? Only UC Santa Cruz had one in the 0.600s this past season but it took two flights and two additional tournaments they hosted to accomplish that number. Next highest was Colorado College at 0.591 but they also needed multiple flights. CMS was at 0.574 with Whittier and Pomona-Pitzer right behind. These were accomplished with a maximum of just one flight.)

To help me with this article, I’m fortunate that there are other volleyball crazy people out there that like to reach out to me. One warning is that if you do reach out, you instantly become a good friend and are on the hook to help me whenever I ask. One unfortunate soul that fell into this trap was Mike Stob who is a retired professor who taught (among other things) mathematics and statistics at Calvin University. I asked him awhile back some “what if’s” about La Verne’s 2023 schedule and impacts to their SOS:

• What if La Verne played St John Fisher and Nebraska Wesleyan instead of Southwestern and Carthage in the Pacific Coast Classic during week 1?

• What if they played Juniata at the Cal Lutheran tournament instead of UC Santa Cruz for a second time that year?

• What if La Verne had been able to add in a late season tournament? (To be super annoying, I listed three different tournaments that actually occurred in 2023.)

The purpose of my questions were to gauge how subtle changes to a schedule could impact a team’s SOS and, consequently, their at-large chances. In my first question, I’m replacing two opponents that might have been better teams for two other teams that were regionally ranked (Nebraska Wesleyan still counted even though they were dropped in the final rankings). In my second question, I was wondering how playing the only undefeated team in the nation changes their SOS. In my third question, I was wondering how much of an impact is there to SOS when adding a good tournament to an already packed schedule.

Keep in mind, I’m not trying to follow any La Verne financial guidelines the Athletic Director has handed down or even NCAA rules regarding the number of dates a team can play. I’m just doing “what if’s” to see how SOS is impacted.

Doing the first change alone raises the La Verne SOS from 0.541 to 0.545. What was interesting to me in this one is how luck in scheduling plays a role, too. I’m not saying that Southwestern and Carthage were for sure better than St John Fisher and Nebraska Wesleyan, but if La Verne had these two on their schedule instead, they get a small SOS bump and would have added two more regional wins (most likely) to their at-large resumé.

Doing the second change (replacing UCSC with Juniata) takes the 0.541 SOS up to 0.557. That’s still not great as they still would have been lower than St Olaf but interesting to see that even adding the team with the best record in the nation only helps so much. (It should be noted that we can’t make these changes in a vacuum. Playing a different match or adding a match causes a new result, which obviously effects SOS for both teams and ripples out into the scheduling multiverse.)

As for the three tournaments, I specifically mentioned:

• The UCSC Tournament where they could have played Christopher Newport and Mary Washington.

• The Calvin Tournament where they could have played Hope, Calvin and Millikin.

• The Millikin Tournament where they could have played Millikin, WashU and Hendrix.

Mike came back with the first tournament potentially moving the SOS to 0.558, the second moving it to 0.563 and the third moving it to 0.551. (Again, everything is an approximation because someone is going to have to win and lose these new matches and that impacts SOS numbers.)

The biggest gain to the La Verne SOS tweaking their schedule was to add in Juniata and play that second tournament. Their new SOS would have been 0.577.

This is a fun exercise (at least to me) but what I came away with was that small changes aren’t really that helpful. One aspect of SOS that I’ve written about in the past is how important a team’s conference is to the calculation. This past season we saw the better SOS numbers in Region X go to the SCIAC and SCAC. The SCAC did this by having the best out-of-conference record of the four Region X conferences. The SCIAC did this by having a number of good tournaments that brought in quality programs. Both of these methods are important because a conference that struggles makes it so much harder for a good team from that conference to overcome that drag on their SOS.

Looking at La Verne again, and their SOS number was right at the average for the conference (and below CMS, Whittier, Pomona-Pitzer and Cal Lutheran). Why was this? The SCIAC went 33-41 out of conference or 23-40 without La Verne’s contribution. La Verne also didn’t take a trip this past season and were a bit unlucky on the teams they drew at their three tournaments. This was enough to drag their SOS down. Remember that with conference play (only), each team is going to be around 0.500 SOS because the final conference record of all teams will be 0.500. In fact, the better teams (like La Verne) will be below 0.500 because we have to subtract out their record because you don’t play yourself. Now, the selection criteria uses not only the record of your opponent for SOS but a portion also includes the record of your opponent’s opponents. That’s why out-of-conference records are so important. Whew. I hope I haven’t lost anyone. All of that to say, in summary, is that a strong conference is probably the best thing any team can do for their SOS. Or, in other words, something that is totally out of a single team’s control potentially controls their at-large chances.

It was at this point writing back and forth to Mike that he put on his teaching hat. (Or is it a professor’s hat? It wasn’t the Sorting hat. Hufflepuff for the win!) Here is what he wrote:

“Here’s a good way to think about it.  Suppose that a team has 25 opponents and a .500 SOS.  Then each opponent contributes .020 to their SOS.  The most a team could contribute is 1/25 or .040.   So, swapping out one of their average teams for a great one gets them at most .020.  In reality, the most that any one team could contribute to strength of schedule is about .9/25 = .036.  (Even an undefeated opponent doesn’t have all undefeated opponents.  This year the maximum Opponents’ Winning Percentage (OWP) was .722. (Chicago).  Only two teams were above .700 (the other being Whitewater).“

(One note about Mike is that he’s obviously old school going with the two spaces after a period. It took a lot of years for me to break that habit but I still love to see it as it takes me back to my junior high school typing class.)

Taking Mike’s example, let’s apply it to La Verne who actually had a 25-match schedule prior to the tournaments. 16 of those 25 matches are conference so they are sitting at a below 0.500 SOS. If their remaining 9 matches had been against 0.500 teams then they obviously would have had a poor SOS. If those nine teams could each contribute Mike’s max of 0.036 (or 0.016 more than our average nine teams) then the La Verne SOS would be around 0.644. That’s a fantastic SOS but not something that they could ever achieve in reality. Best case, maybe, each team contributes 0.030 and that would take La Verne’s SOS to 0.590 in this hypothetical schedule.

Looking forward to 2024 and La Verne again. Achieving that 0.590 SOS looks to be pretty impossible without a big improvement by the SCIAC as a whole. Without that, they would need to schedule high quality tournaments and that probably means at least one flight. La Verne is actually in a good position because there are some good SoCal tournaments each year but you just never now about the quality of the teams coming. With the way the SCIAC schedules, La Verne is not allowed to add a fourth tournament but that’s what I’d like to see. Adding a top-notch tournament with a slight improvement to the SCIAC probably gets La Verne to 0.570 and that’s probably good enough for an at-large bid as long as they continue to win and beat ranked teams.

Let’s expand this a bit and move away from the La Verne example. The ASC, NWC and SCIAC all have the same scheduling issue although the ASC’s (unfortunately) gets a bit of a reprieve as more teams leave. The issue is that their conference schedule takes too many dates to play. Each play home and home which creates the double whammy of having to play poor teams multiple times. For each poor team a program plays twice, it drags down their SOS while taking almost 10% of their allowed dates to get all of their matches in for a season. Again, another example of how a below average or poor conference can negatively impact your SOS. One solution to this issue, which we’ll never see in the West, is to go to a single round of conference play. This would minimize the conference impact while opening up dates to play quality programs. In the West, unfortunately, we are on an island so there just aren’t any DIII teams sitting around in California, Oregon or Washington to play. It can be done but it comes with flights to the point where you are basically mimicking that UC Santa Cruz schedule. One solution to the open dates is to get creative in conference where only one round of matches count towards conference but you allow the top half to play a second match against each other while the bottom half plays a second match against each other. You’ll notice that in the MIAA, they have moved to a single match conference schedule but Calvin and Hope still play an extra match against each other. This is done for rivalry purposes, but it does give both teams the side benefit of boosting their SOS.

Assuming these conferences can’t move to a single match conference schedule, the best they can do is to condense conference matches such that multiple matches are played on the same day. If you do this twice in a season then you can open a weekend up for another tournament. The problem, again, then becomes money as flights are required.

I’m a little all over the place here but let me give you my last thought on all of this. What if our four conferences could free up a weekend and whatever teams could get the money agreed to meet up at one location (rotating between the four conferences). You are guaranteed at least 3 matches on a known date but you don’t know the teams until a week before. The tournament then matches the best teams with the best teams and we work on down the line. This would help with regional rankings and improve our best teams’ chances of getting an at-large by helping ranked results and SOS. This could also be done in conjunction with an out-of-region conference like the East/West Challenge is sort of done today if only one of our Region X conferences could do something like this. (Mike chimed in on this saying that the MIAA and NCAC have been doing this for the last couple of years. Each conference has nine teams so they split them up three-ways and roll the ball on the court.)

Still with me? In summary, the West has a problem getting programs with high SOS numbers because of the lack of opponents in driving distances of our conferences. To fill our dates, we play long conference schedules that actually hurt our at-large chances. While other regions can easily get over a 0.600 SOS, our region’s equivalent is probably closer to 0.570. With La Verne getting an at-large selection this past season, maybe this has finally sunk in with the Selection Committee. From our examples above, you can see that there isn’t a magic pill that will take a team’s SOS in the West from 0.540 up to 0.600 in any realistic way. The West also suffers from a lack of regionally ranked matches for the same reason. When you consider that national selection typically comes down to record, SOS and ranked results, the West is typically down 2-0 in criteria before anyone looks at the data. The easy solution is money. Make more trips while freeing up extra tournament dates. I think there are some creative options that aren’t being explored but in the end the battle for a travel budget is a tough one for these coaches.

Thanks again to Mike for the help on this article!