Quote from: Kuiper on June 07, 2024, 05:47:33 PMI read this as basically the Men's Soccer Committee setting dials as far toward SoS as they could (since 25/75 is the default between Wins/SoS). Still have to have 10 wins to receive the NPI adjustment, which means it has a high bar for when wins only can help. (I could be wrong on my interpretation though)

The balance of WP and SOS could be any combination of numbers adding up to 100, though it it is worth pointing out that men's soccer is leaning more heavily into the SOS number than the rest of the fall sports. Football is at the opposite end of the spectrum electing a 40/60 WP/SOS split.

As for the ten wins number, the NPI system is designed such that teams don't benefit from a loss that raises their NPI or a win that lowers their NPI. Such games are dropped from a team's resume but that can present some wonky situations where an undefeated team at the top would essentially be left with a one-game resume (their best win, since every other lesser win would lower their NPI). The ten minimum ensures that every team is evaluated based on a resume made up of at least ten games.

That's getting a bit into the weeds of NPI but hopefully makes sense and is helpful.

There's not a ton to quibble about here.  I was initially wary at seeing the SOS dial being turned up higher but the QWB could offset that a bit for a team with a lower SOS.  Could also equalize teams that don't have a lot of wins/have lots of ties but game the system with a high SOS.

Minimum wins also seems like a good addition.

Ultimately I don't think a lot will change, like every year there will be 1 or 2 teams that get in that many will disagree about.

What happens if you don't win 10 games, ala 2022 Williams?

Quote from: Ejay on June 07, 2024, 08:01:24 PMWhat happens if you don't win 10 games, ala 2022 Williams?

This too I am wondering about.  I can't decipher if the minimum win threshold is for teams with a high Win% but low SOS getting to the table or for teams trying to game the system with a hard schedule but not winning the games.  I'd much prefer the latter but the former would still be a good adjustment.

I sent the following questions to the NCAA:

Does a team need 10 wins to be considered for at an large bid.

When calculating QWB is it the NPI at date of match or end of season

Will post answers when I receive a response.

Quote from: ziggy on June 07, 2024, 07:54:28 PM

Quote from: Kuiper on June 07, 2024, 05:47:33 PMI read this as basically the Men's Soccer Committee setting dials as far toward SoS as they could (since 25/75 is the default between Wins/SoS). Still have to have 10 wins to receive the NPI adjustment, which means it has a high bar for when wins only can help. (I could be wrong on my interpretation though)

The balance of WP and SOS could be any combination of numbers adding up to 100, though it it is worth pointing out that men's soccer is leaning more heavily into the SOS number than the rest of the fall sports. Football is at the opposite end of the spectrum electing a 40/60 WP/SOS split.

As for the ten wins number, the NPI system is designed such that teams don't benefit from a loss that raises their NPI or a win that lowers their NPI. Such games are dropped from a team's resume but that can present some wonky situations where an undefeated team at the top would essentially be left with a one-game resume (their best win, since every other lesser win would lower their NPI). The ten minimum ensures that every team is evaluated based on a resume made up of at least ten games.

That's getting a bit into the weeds of NPI but hopefully makes sense and is helpful.

Thanks.  That's how I read it too.  I kind of view these things from the prism of the weaker conferences where the conference has a larger conference schedule and, as a consequence, few non-conference game slots.  They may do that because there are no other DIII programs close by or because it just makes for easier scheduling, but an individual men's soccer team is required to abide by it even if they would like to play more challenging games.  That is perhaps the most sympathetic case for not penalizing a team on the SoS side for "bad" wins.

Take, for example, the Northwest Conference.  They play a double round robin because the closest DIII non-conference team, UC Santa Cruz, is a 10-12 hour drive away and the next closest might be Concordia Minnesota or Colorado College, which are both probably a 16-17 hour drive from the NWC's easternmost member.  Pacific Lutheran won the AQ with a record of 13-1-2 in conference and 14-2-5 overall.  They do not have a conference tournament and they play everyone home and away, so this wasn't a "fluke" AQ by a school that just got hot at the right time.  Whitworth finished second with a conference record of 11-2-4 and an overall record of 13-4-4, still a great W-L record, obviously, but their conference record didn't produce a great SoS and although they traveled almost 1,000 miles to Santa Cruz to play UC Santa Cruz and Cal Lutheran, they lost both games.  Whitworth's 11 wins would make the eligible for NPI adjustments.  Nevertheless, because they lost both games to PLU, they probably wouldn't have any strong SoS wins and therefore still wouldn't have gotten a Pool C bid.  If they had won one of those games, however, and beat some of their non-conference opponents (especially if they were stronger and they had not scheduled the NAIA teams), losing out to PLU might not have been disqualifying.  That suggests that this might strike the right balance of not penalizing Whitworth for having so many low SoS wins, while incentivizing them to maximize their smaller opportunities to gain high SoS results.

In the SCIAC, Redlands finished second to Occidental in the conference tournament and therefore lost out on the AQ.  Redlands had an overall record of 12-3-3 and 8-2-2 in conference.  Only 1 of its 4 non-conference wins, however, were against an NCAA opponent (UC Santa Cruz), so it would only have 9 countable wins and would not be eligible for NPI adjustments.  That doesn't really bother me, though, since Redlands lost three straight games to Occidental (including the SCIAC final) and had some weak ties to Pomona-Pitzer and La Verne, but other than paying a ton of money it might not have had to fly across the country or paying another DIII to fly in and play them, it didn't have any other options for better NCAA non-conference games.  It also didn't lose to a single DIII opponent other than Oxy.  So, it basically gets punished for being an island conference and not having the budget to overcome that.  Again, I'm not complaining because if they had won one of their two winnable ties, they would have had the necessary 10 wins.  Plus, you still need some good SoS to get Pool C and they didn't have it.  Nevertheless, it illustrates the margins are still very thin for island conference members, especially if they lack resources to travel.

Coach Bianco confirms that the 10 wins minimum criteria is poorly written.  It's actually 10 games minimum for NPI purposes according to him, which means it includes losses in the 10 games.  Here's a Twitter exchange where he discusses it and more about the NPI system with football and basketball D3 guys

https://x.com/BMBianco/status/1799429004522397850

Bianco:

QuoteThis is very poorly written. It's not a "min win", but as I understand it, the min. # of games that must be counted in the algorithm. All L's count. But also, SOS they say is an avg of your opponent's NPI ranking. How can SOS be the avg of something you're using SOS to determine?

Logan Hansen: 

QuoteAll losses count, even those that would raise a team's NPI? So a winless team would have all their games included in the calculation?

On your last point, the calculation iterates repeatedly. SOS, then NPI, then SOS, and so on, until the values stop changing.

Bianco:

QuoteYes, all loses count as I understand it. So if you play a "bad" team and lose, you'll be punished. What this wants to do (right or wrong) is not punish you for beating a "bad" team. In their words, "a game you should win". This counts for league and non-league opponents

Hansen: 

QuoteThat would be another change, then. Previously, losing "a game you should lose" wouldn't count, like if a winless team lost to an undefeated team.

Zac Snyder:

QuoteThis seems like an OK way to handle the problem of a one-game (or low sample size) resume for a winless team without needing to set a min loss number like they did for min wins. Any team "benefitting" from a loss will be well below any line for tournament consideration.

Bianco:

QuoteI guess the way I have looked at this is has more to do with how much do we want to evaluate a team's (any sport) full body of work? Last year in soccer, the original NPI #1 team had 80% of their 20 game schedule removed/"adjusted". Even raising that min # to 10 still removes 50%

Hansen:

QuoteYeah, I agree with you. There are plenty of metrics out there that wouldn't "penalize" a team for beating bad teams, while still evaluating the full body of work.

Bianco:

QuoteAnd I think the balancing act is trying to find a way that does penalize teams for avoiding playing good teams in favor of bad. By removing games, teams aren't as incentivized to play a rigorous schedule. I would understand adjustments for league game, but this applies to all.

[End quotations and just my own perspective] I think the last point is probably the key for most people. Coaches/teams can't control who they play in conference and they likely don't have a lot of input on conference affiliation, especially since it affects all sports.  They do, however, have control over their non-conference schedule for the most part and there's no particular reason to incentivize teams to avoid strong teams.

Only caveat I would make to that is that "avoiding playing good teams in favor of bad" implies some measure of intent.  It's easy for a coach from an Ohio school or many other places in the upper Northeast quadrant of the United States to find good opponents in close enough proximity to play them economically, especially if they work out out a home-and-away series that effectively splits the costs over two years.  So, you might infer intent if they choose bad teams rather than good.  It's much harder to make that inference for a school in the SCIAC or the NWC, for example, where you have to fly at significant expense to find any non-conference D3 opponent or subsidize the trip of a D3 to come to you.  Even in more densely populated D3 areas, it may be too difficult to find a reasonably close strong opponent to play your team if your team has gotten just good enough to be dangerous, but not good enough to provide an adequate SoS/NPI benefit.

Here is the explanation from the Men's Soccer Committee for why it set the dials where it did.  Given that the committee was opposed to the NPI system as currently developed, and to applying it this year in any event, it's not surprising that the dials were set in a way that prioritizes SoS as much as possible.  That preference theoretically protects the status quo as much as they could for teams that created schedules this year that were designed to maximize SoS (and there are many that fall into this category if you are following the 2024 schedules thread!).  I say "theoretically" because many insiders have suggested that the dials don't change things a ton.

https://ncaaorg.s3.amazonaws.com/championships/sports/soccer/d3/men/2023-24D3MSO_NPI.pdf

QuoteWinning Percentage/Strength of Schedule.

The committee set the NPI dial to 15/85 with the goal of balancing the weight of a team's schedule
strength with its performance within that schedule.

• Home/Away – Win/Loss Weights.

The committee didn't think there was a need for a home/away advantage to be applied at this time
within the NPI formula. However, they noted that they would examine any potential need for this in the
future with feedback from the membership in regards to scheduling and if there is truly any related
advantage in playing at home or on the road.

• Quality Win Base.

The committee set this to the default number of 54.0. The default was near or equal to the amount of
ranked teams the committee has been ranking over the past few years. It also aligned with a good
percentage of the membership sponsoring men's soccer that the committee felt was deserving of a bonus.
This would also encourage teams to schedule better teams in order to earn quality win bonuses as part
of the NPI formula.

• Quality Win Base Multiplier.

The committee increased and set the QWB multiplier to .750 in order to address ties within the soccer
formulas that are different than a majority of other sports. By increasing the multiplier to .750 it made
the value of quality ties equal to or more valuable than average/below average wins. This again
encourages teams to play quality schedules and earn wins and/or ties against good opponents in order
to earn a spot in the NCAA tournament.

• Overtime Weight.

There is no overtime in the regular season and so the committee did not apply an overtime win/loss
impact.

• Minimum Wins.

The committee set the dial to a minimum of 10 wins. They arrived at this number by setting the dial to
zero in order to see what teams were keeping when there was no minimum. A majority of teams were
keeping 10 or more games and so it seemed like a good number to bring everyone up to 10 that was
keeping less. Also, the average number of games being is around 17-18, the committee felt 10 allowed
teams to drop a few games, but still require they keep a majority of its schedule knowing losses and ties
are automatically kept.

I thought the intent to make quality ties worth more than average or below average wins was interesting.  It fits in the theme of scheduling strong opponents, but I wonder if it also (perhaps unintentionally) will incentivize teams to play conservatively in those games against tough opponents.  It might depend upon the difference in the effect on NPI of a quality tie v. a quality loss.

Meanwhile, in the other football, they've apparently set the dials such that SOS is basically irrelevant and it's all about not losing ... to anyone.  Analysis from Logan Hansen (which starts with "WP explains >96% of a team's NPI"): 

Stat-filled detailed analysis:  https://twitter.com/LogHanRatings/status/1800203411092324624
TLDR version:  https://twitter.com/LogHanRatings/status/1800209963715109131

Quote from: Ron Boerger on June 10, 2024, 07:56:43 PMMeanwhile, in the other football, they've apparently set the dials such that SOS is basically irrelevant and it's all about not losing ... to anyone.  Analysis from Logan Hansen (which starts with "WP explains >96% of a team's NPI"): 

Stat-filled detailed analysis:  https://twitter.com/LogHanRatings/status/1800203411092324624
TLDR version:  https://twitter.com/LogHanRatings/status/1800209963715109131

Not enough games or crossover games to make SoS make sense. When you have 1 or 0 OOC games in your conference, your SoS is going to be near .500 regardless of who you play or how good you are.

NCAA answers in italics below

1-Does a team need to have a minimum of 10 wins to be considered for an at large bid (e.g. High NPI but only 8 wins)?

The minimum wins does not require a team to reach that minimum to be considered. This dial is just specific to any teams that would be dropping games leaving them with under 10 countable games, those few teams would be required to keep 10 wins

One note, specific to soccer, with ties – those equal half wins. So a team that is 8-2-4 would keep the eight wins, and then four ties is equal to four half wins or two wins an that would equal 10. So even though it is listed as wins, ties do play a roll in hitting that minimum for those looking to drop a win against a bad opponent that hurts its NPI.

2-If a team receives an ACQ (Automatic Conference Qualifier) and does not have 10 wins and a low NPI I assume the ACQ overrides?

AQ will override the NPI in basically all factors as their AQ is their selection. But specific to this question, the 10 wins are not a requirement to be selected and so no matter the situation, they are eligible.

3-If a team receives an ACQ (Automatic Conference Qualifier) and does not have a high NPI I assume the ACQ overrides a low NPI?

This is similar in that AQ's are identified and it doesn't matter their rank, they are selected as an AQ. Then the remaining teams left for the Pool B/C are then selected in order of NPI from the teams remaining after AQ's are identified and selected.

4-When a QWB is calculated, is it the NPI ranking of teams day of match or end of season?

The QWB and all numbers will change after every result. So if you got a 1.3 bonus day of game, but then that team loses the next 5 games, that bonus likely will turn into a lower number or may even go away if that team drops far enough. Same goes for if you got the 1.3 and then that team won all of its remaining games, that bonus would increase for you. So really the only calculation that matters is the one at the end of the season as all the numbers and all the calculations will change after every result including your opponents results.

A D3FB stats guy developed a tool that supposedly allows you to check how the NPI system would have worked in prior seasons with various adjustments to the dials.  Presumably, you could also set the dials at the settings selected by this year's sport committee and then know exactly how they would do, or adjust the dials to see if that would matter.

He has only entered the data from prior seasons for two sports thus far (football and men's basketball for some seasons), but I thought I would post this here so you could check men's soccer if and when he enters the data for our sport.

Here is the site with the tool

https://loghan.shinyapps.io/NPI_Calculator_V1/

I have all the fixtures from the past couple of seasons in csvs.  Ideally, this guy would just throw whatever R script he is using on a github and opensource it.  Or we can wait

Here is a quote from a football guy that kind of sums up the coaches' view of NPI

https://x.com/LogHanRatings/status/1803482350040367467

QuoteDIII football only has at most 10 games to evaluate. Why it makes sense to ignore half those games will never make sense to me. If you're a "good team in a weak conference," schedule good teams in non-conference and win! You're letting them off the hook for avoiding the risk...of additional losses, and putting it all on teams who DO play a lot of good teams. Teams in strong conferences are saddled with maximal risk and minimal benefit, and teams in weak conferences get minimal risk and maximum benefit. The risk-reward balance is out of whack.

BUT, here's a paraphrased quote interview with a very experienced men's soccer coach, Jeb Burch from Centre that I think accurately summarizes the College President/AD view of NPI

https://www.youtube.com/watch?v=Lc6Qdy_-ZXo

~This is all about the cost of travel.  Many schools simply can't afford the travel in search of SoS that is required for tournament selection under the old system~

A lot of coaches who are at wealthier schools or who are at schools in parts of the country where you can't throw a stick without hitting another strong DIII program have trouble understanding this latter perspective, but I think Burch is dead-on right.  And I'm not sure the Presidents/AD's are wrong when I read about how many schools are in financial difficulty and expect it to get much, much worse before it gets better because of the demographic cliff.

I agree...its has occurred in SUNYAC... Brockport & Geneseo to Empire 8 cited travel time & costs. Same for New Paltz leaving SUNYAC in two years to TBD.
SUNY Canton & Morrisville into SUNYAC less travel.