Same formula, different name.
Division III calling its power rating formula a “Strength of Schedule Index” was certainly misleading. It didn’t measure how strong the schedule was; it measured how a team did against that schedule. And even though we put a disclaimer on our listing of the index in large, bold type, it didn’t sink in with Division III fans.
But instead of fixing the formula, the championships committee will change the name. This year it will be called the Quality of Wins Index. It still won’t really accurately contribute toward discovering how good a team is, however, and since this is a major factor used in selecting and seeding NCAA playoff teams, it would be helpful if it did.
Here’s the problem — a team gets certain points for each in-region game based on the opponent’s record, the location of the game (home/away) and the opponent’s regional record. However, it fails to measure exactly how an opponent got to that regional record. Did it go 7-3 by playing in a weak conference or did it go 4-6 in a strong league? There are many times a 4-6 team from one conference is better than a 7-3 team in another, and in the instances that our sample 7-3 team is willing to play a tough team, it shows. But a team is rewarded with 14 points for beating any 7-3 team at home, 15 points for doing it at home, regardless of how the opponent got to 7-3.
Strength of schedule ratings used to measure opponents’ opponents’ record. They need to start doing it again. It can be done very simply — take the current SOSI (or QOWI) and multiply it by the combined regional winning percentage of the team’s opponents’ opponents to get a new metric with which to measure teams.
For example, let’s use two teams with identical SOSI/QOW numbers from last year: Ithaca and Mt. St. Joseph. Both had a 10.300 SOSI/QOW.
Ithaca was 8-2, 8-2 in region. They played the following schedule:
Sept. 11 Buffalo State W
Sept. 18 Hartwick W
Sept. 25 at St. John Fisher L
Oct. 2 Lycoming W
Oct. 9 at Norwich W
Oct. 16 at Brockport State L
Oct. 23 Utica W
Oct. 30 Springfield W
Nov. 6 at Alfred W
Nov. 13 at Cortland State W
Those 10 teams had a combined opponents’ regional record of 430-325, .5695 winning percentage. Multiply that by the 10.300 QOW and get 5.869. Let’s call that Contextual Quality of Wins (CQOW).
Here’s Mt. St. Joseph, 10-0, 10-0 in-region:
Sept. 4 Wilmington W
Sept. 11 Rose-Hulman W
Sept. 18 at Kalamazoo W
Oct. 2 at Hanover W
Oct. 9 at Franklin W
Oct. 16 Manchester W
Oct. 23 at Defiance W
Oct. 30 Bluffton W
Nov. 6 Anderson W
Nov. 13 at Thomas More W
Those 10 teams had a combined regional record of 393-346, .5318 winning percentage. Multiply that by the 10.300 QOW and get a CQOW of 5.478.
This is how we could get more useful information into the Division III playoff selection and seeding process. If both Ithaca and Mt. St. Joseph were up for an at-large bid, Ithaca should get it. This analysis shows how it can be done objectively. And it can be done with the information the NCAA already has.
I think the whole concept of regionality also throws the playoff system out of whack. I know why they do it, but I don’t like it. I do think that your idea does have some merit and would be a vast improvement.
Oh, indeed. But regionality is a basic Division III tenet that I don’t think we’re going to be able to affect much change on.
Does the new name apply to basketball as well?
Yes, it does. I’ll do a similar post on D3hoops but I haven’t had a chance to do the math — and with 25 games in basketball there’s quite a bit.
Respectfully Pat, I think it would be quite informative for you to post your Strength of Winning modification as well as the current “SOSI” (or new QOWI), beginning the first of November as academic assessment of the relative strengths of the 2 Indices.
I’d like to. If we can calculate it on a regular basis I’ll use it.
I’m just trying to hash this out in my head (and be a bit of a devil’s advocate in the process).
Hypothecially, let’s say in the MSJ-Ithaca scenario provided, MSJ was part of a conference that did not have an automatic qualifier. As a 10-0 team that could potentially lose a playoff spot to an 8-2 team, wouldn’t MSJ be getting cheated a bit? They’re Quality of Wins Index might be weaker than other teams, but MSJ did beat every team they came against.
Schedules are made years in advance, so they are being penalized for playing a weaker sked even though they played it as flawlessly as the schedule would allow them. They might not have had a chance to prove themselves, but does that mean they aren’t a good enough team?
I point this out only to suggest/reiterate that the new system is not yet perfect either.
Conversely, why would we punish Ithaca because they played a harder schedule than Mt. St. Joseph?
The fact remains you would be penalizing a team for attaining perfection. As stated in an earlier post, they played the games as scheduled and had no way of knowing the quality of the opponent several years in advance of playing. There just seems to be something wrong with leaving an undefeated team home for the playoffs.
Would the playoff alignment this year take care of this specific situation as shown from your example and allow both teams into the playoffs?
I think that the 32-team format allows too many at-large (Pool B and Pool C) bids for an undefeated team to be left out. Realistically, I cannot foresee a 10-0 Pool B team not getting a bid. The only 10-0 Pool C candidate that I can foresee is in the MAC or the NCAC (and the NCAC having 10-0/7-0 co-champs is a real stretch).
Pat’s example showed how much value playing in a strong conference can mean in his scenario, especially if it is applied to seeding.
I see the real value as rewarding teams that schedule strong non-conference games at the 9-1 and 8-2 records.
“The fact remains you would be penalizing a team for attaining perfection.”
No. We would be penalizing them for not challenging themselves.
Or, we wouldn’t be rewarding them for playing a weak schedule (a la Lakeville 1995, Thomas More in whatever year that was …)
What this is intended to show is exactly how much a TRUE strength of schedule plays a part in what record a team achieves. What would Mt. St. Joseph had done against Ithaca’s schedule? Probably less than 8-2 from the looks of it. Again, MSJ is covered by the automatic bid, but were they a better team than Ithaca? No, I don’t think so. SOSI/QOWI was inconclusive but those are numbers taken out of context. Contextual SOSI/QOWI is a lot more telling.
I suppose that I look at it a bit like punishment for a perfect-record team because schedulers really don’t have much idea how an opposing team is going to perform two or three years down the road when the schedules are made.
Did anyone think that Hartwick would stumble from an 8-2 team in ’02 down to a 1-8 team this past season?
Or on the opposite front, that Thiel would go several years with only two or three wins a season, and then break out with a commendale 7-3 record in 2004?
I probably wouldn’t contest this matter at all if a team with MSJ’s quality index had even just one loss. Maybe the index would serve the teams better if it measured the margin of victory in each game, so that way a team that has a weak sked and is undefeated can boost its rating by beating teams by 20 or 30 or 40+ points.
Actually, I don’t condone that, because running up the score is not what Div-III is about. But it seems to be a way for a team with a weak schedule to be able to prove itself in comparison to the 8-2 teams out there that might have a better index rating.
I have a problem judging the strength of an undefeated team’s schedule, because no one’s beaten them, it’s hard to say if they would have been beaten had they had a harder schedule. Any undefeated team should make the playoffs, and with the 32-team format, they will.
Remember there are essentially 11 at-large bids now, since Pool B leftovers will be thrown in with Pool C after the first few are selected. Unless there were several undefeated teams in non-AQ leagues, or in leagues that have AQs but it’s possible conference opponents did not play during the season (since the NEFC has a title game, I think this is only possible in the MAC), the teams would likely get in, even if playoffs were based on a ratings formula.
Don’t forget, the rating is only part of the formula. There is still some subjectiveness at the hands of the committee. More than likely, with 32 teams, ratings would be separating 8-2 teams from 7-3 teams, or better yet, matching a team with nine games and one loss against a 9-1 team.
That said, there is an issue with scheduling. Rowan can hardly get a Division III non-conference opponent besides CNU, while bad teams fill their openings quickly. The AQ allows teams to take chances in their non-conference scheduling, and some still aren’t doing it.
But we’re glad to see the committee rewarded CNU this year with a playoff bid, since they scheduled aggressively and went 8-2, as opposed to taking a 9-1 team that got one more win by playing weaker teams.
True you can only play who you schedule, but that doesn’t make strength of schedule a less legitimate way to make sure we get the correct 32 teams in the show.