Case Study (Williams)

Remember to vote on who will win the Championship!

I had planned to hold off on these Case Studies but I find I have some time on my hands. I don’t think there are a number of huge areas of concern but I like doing these to gain insights into what the RACs were thinking and what the NCAA was allowing. I also think these are extremely valuable to coaches when they do their scheduling. If a coach wants an NCAA Tournament at-large bid, what are the RACs looking for and how do you satisfy that with your schedule.

The first team I want to look at is Williams out of Region I. When I was doing my projection work, I struggled with the Region I rankings and ultimately put Amherst in front of Williams. The RAC didn’t agree and had Williams ahead. This, of course, mattered because Williams was selected for an at-large bid and Amherst was not. Here are the base numbers (using the criteria) for Williams:

Williams 14-8 (0.636), 0.590 SOS, 8-5 Ranked Record

From the surface, you’d say the record is a problem but the SOS is fine and the Ranked Record is outstanding. They do have 3 bad losses, which is a concern, but they are a potential at-large team with these numbers and what makes that true is that Ranked Record.

Here were the Regionally Ranked teams [with regional ranking] that Williams beat: Middlebury (twice) [5], Amherst [4], Bowdoin [6], Clarkson [7], Springfield [5], Babson [2] and Colby [7]. Outside of Babson, that’s not a particularly impressive list of teams. Only two are ranked in the top 4 in their region and none of those teams would be considered to be a top national team (by whatever measure you would want to use).

Here were the teams Williams lost to: Brandeis [unranked], Hamilton [unranked], JWU (RI) [unranked], Wesleyan [2], Tufts (twice) [1], Cortland [2] and MIT [1]. No problems losing to those ranked teams but those are three troubling unranked losses. Brandeis was 13-15, Hamilton 8-11 and JWU 20-11.

So, I have some concerns. This team has shown that they can play at a Babson level but can also sink to a much lower level. They are an average team according to the criteria to be very frank about it. Where they greatly benefited was that their entire conference made up their region’s regionally ranked teams. By finishing 3^rd in conference, they were guaranteed to have a good Ranked Record. Now, the criteria shouldn’t care how you created the criteria but context is important when you move from Ranked Record to Ranked Results. The first is looking at the win/loss numbers and the second is looking at the quality of the teams played.

The first comparison I wanted to research was how did they compare against Amherst.

Amherst 14-7 (0.667), 0.601 SOS, 6-6 Ranked Record

Amherst has a better record and a better SOS but that SOS advantage falls within the tolerance used to compare the statistic. Amherst has a worse Ranked Record but only one bad loss. If you factor those losses in then this is very close.

In their direct comparison, Williams has a Head-to-Head advantage but their Common Opponent Records are tied. That means the scorecard is 1-1-2 with Ranked Results on deck to break the tie.

Here are Amherst’s Ranked Wins: DeSales [7], Middlebury [5], Bowdoin (twice) [6], Emerson [3] and Colby [7].

And, here are their losses: Wheaton (MA) [unranked], NYU [1], Williams [3], Wesleyan (twice) [2], Springfield [5] and Tufts [1]. Wheaton finished the year 11-15.

These two teams play in the same conference and don’t really travel so not surprising these results are very close. Amherst has shown they can play at an Emerson level (less than a Babson level) but can also sink much lower. Very difficult comparison but I would have to agree with the RAC and put Williams ahead of Amherst primarily for that Babson win. Here’s the rub, however; from the surface, I would think that Amherst (better record and SOS) would have had a better chance at an at-large bid. So, it’s a little strange that Williams got in but Amherst didn’t. Maybe Williams was the last team selected or, as we’ll come back to later, maybe the Selection Committee is evaluating Ranked Results wrong.

The next point of conflict with Williams was their selection for an at-large bid. We have no idea at what point they went during the selection but we can look at the teams left standing at the end of the day and do some comparisons. Here were the teams under consideration at the end that didn’t get a bid:

• Emerson

• Brockport

• Stockton

• Region V Unknown

• Washington & Lee

• Case Western Reserve

• Region VIII Unknown

• Bethel

• Cal Lutheran

We have two unknowns because those regions got at-large bids or automatic qualifiers for all of their regionally ranked teams. I believe Region VIII probably had UW-Platteville and I want to use them, plus Case Western and Bethel in the next comparisons.

Williams: 0.636, 0.590, 8-5UW-Platteville: 0.719, 0.579, 4-6

There are no matches between them and no Common Opponent matches. The SOS falls within tolerance. You might wonder that maybe the tolerance is less this year. Well, if that was true and this was considered a Williams win then Amherst (who had the same margin above Williams) would have been ranked above Williams. We have to trust in the consistency. The scorecard has this Platteville 1-0 without looking at the Ranked Results. We’ve already looked at Williams Ranked Results so here is Platteville’s. Wins against Coe [3], St Olaf [7], UW-Stevens Point [3] and UW-Eau Claire [ranked previous week].

Their losses were to UW-Eau Claire [ranked previous week], Millikin [7], Aurora [unranked], UW-Oshkosh [4], UW-Whitewater (twice) [2], Carthage [unranked], UW-River Falls [unranked] and Cornell [4]. I honestly would consider their unranked losses to be against better teams than what Williams played for most of their entire schedule.

UW-Platteville proved that they can play at a level of Stevens Point and Coe and really didn’t sink to any low level throughout the year. They had a higher peak and a higher floor. There is absolutely no doubt in my mind that Ranked Results has to be given to UW-Platteville. It’s not even close or worthy of more than a minute of discussion. That makes the final scorecard in favor of Platteville 2-0 with three criteria tied.

Williams: 0.636, 0.590, 8-5Case Western: 0.786, 0.561, 3-6

There are no matches between these two teams but CWRU does hold a Common Opponent advantage. That puts the scorecard in favor of CWRU 2-1-1 without factoring in Ranked Results. You know how this is going to end, right? CWRU’s wins are Mount Union [previously ranked], Carnegie Mellon [2] and WashU [1]. Their losses are Smith [6], Emory (twice) [1], Chicago (twice) [6] and NYU [1]. They have no bad losses. So, when it comes to Ranked Results do you favor a team with fewer matches played against Ranked teams but beat WashU and CMU or one who played more and who’s best win is Babson? Not to mention the second team also has 3 bad losses. Like the previous comparison, there is no doubt that CWRU’s Ranked Results are better than Williams. You could argue that the wins only came in conference (UAA) but isn’t that what Williams did? You could favor more Ranked matches but that’s a function of a region that only ranks team in a single conference. It doesn’t matter how the criteria gets made. Final scorecard here is CWRU 3-1-1.

Williams: 0.636, 0.590, 8-5Bethel: 0.615, 0.597, 4-7

There are no matches between them and no Common Opponent matches. The scorecard has this for Williams 1-0-3 with Ranked Results to go. I’m not going to dive into Bethel’s record like I did above because we know how I’m going to interpret it. Still, just know that Bethel beat UW-Whitewater, Cornell, St Olaf and Augsburg. Whitewater is what makes this list but this is probably the closest comparisons to Williams that we have here. Is it enough to tie up the scorecard and then push it in Bethel’s favor? I think so, but it’s not something I can probably justify with just data.

Back to selection and we don’t know when Williams was selected. I believe it’s very likely that Platteville, Case Western and Williams were on the board at the same time. At the very least two of them were at the same time. The Selection Committee decided on Williams and, to me using the criteria, it’s not even close and in the other direction.

So, we have to ask ourselves why? I wrote an article on the Selection Process where we learned that Ranked Results were more results orientated in the past and that the RACs have moved towards raw numbers. In other words, if we are moving towards raw numbers then we consider Ranked Wins as being equal to other Ranked Wins. Beating CMS is the same as beating DeSales. My opinion is that anyone who believes this is an idiot and should not serve in a selection capacity. It’s not what the criteria states, it’s not what the NCAA has advised in the past and it’s just obviously wrong. I guess we have to re-look at the scorecards if the committee evaluated Ranked Results as Ranked Record.

Williams/Platteville goes from Platteville 2-0 to a tie 1-1. Williams/CWRU goes from CWRU 3-1-1 to a tie at 2-2-1. At that point, anything is possible. Williams has a better SOS than both (although statistically tied with Platteville) and that might have been the tipping point. I do have to point out that towards the end of the selection, it’s often not about just two teams. Is it possible that four or five teams were there in the end and there was no clear winner? I suppose, but in these comparisons, Williams wouldn’t be the team to bring forward after clear losses to two other teams and one close loss (in my opinion). Remember, I didn’t check the others, they could have lost those comparisons, too. 

So where does this take us? If I’m a coach with NCAA Tournament aspirations, I need to know why Williams was chosen. Did we see a dramatic shift in how Ranked Results are evaluated? Remember that we made the regions somewhat the same size and limited the rankings to 7 teams (new this year). Did that spur the Selection Committee into thinking that all of the rankings are equal?

On the assumption that the Ranked Results criterion is now evaluated like a record, what this means to scheduling is that you can avoid the elite teams. They don’t benefit you outside of SOS and you can get that same benefit with mid-level regionally ranked teams while also helping you in Ranked Results (and Record). Instead of UT-Dallas, for instance, going to the East/West Challenge, they should fly up to the Northeast and play 3 or 4 teams from the NEWMAC or NESCAC or SUNYAC. La Verne and UC Santa Cruz shouldn’t fly to Trinity. Now you have to win these matches, of course, but taking down Springfield, Smith, Coast Guard, Colby, Middlebury, Amherst, Williams, New Paltz, Cortland and Brockport are a lot easier than CMS, Emory, Trinity and WashU. If the Selection Committee considers all of these teams equal then the response is really quite simple.

Wrapping this up, I can’t believe that the Selection Committee has moved to Ranked Record. I simply don’t think they are that stupid. We are told that selection comes before travel so let’s not go down that rabbit hole. There were mistakes made in order to get Williams into the tournament and the onus falls on the RAC Chairs that were hurt. If I was the Region VII or Region VIII RAC Chair, I would have quit before I let Williams get in before a more deserving team from my region. Failing that, I would be broadcasting what happened to the offending teams and who is to blame. These are really easy mistakes being made in the comparisons and if you are not got to stand up to this nonsense then move on and get out.  

(Later Update) If you were following the tournament, you know that Williams made the Regional Final. This article is about how they should be evaluated by the selection criteria and not about the quality of team they are. Just because they cooked a really good meal, doesn’t mean they should be rewarded for the NCAA stealing the ingredients.