Pool C

[KnightSlappy]

Believe it not, the 2016-17 basketball season is just days away. But the season can't start without Hoopsville hitting the air!

Tune in tonight starting at 7pm as Dave talks to the two preseason numbers one teams, finds out how the offseason went for the two defending national championships, and touches bases with the men's and women's basketball committee chairs.

Guests include:
- Kevin Vande Streek, men's basketball committee chair and head coach for Calvin

On Hoopsville KVS mentioned that they have a method they're hoping to implement to correct the way the HAM is applied to the SOS. Do you have any idea what this is? Is the NCAA actually changing the way the multiplier is applied?

(I wish someone was around to tell us the multiplier was being applied incorrectly three or four years ago ).

[Dave 'd-mac' McHugh]

Believe it not, the 2016-17 basketball season is just days away. But the season can't start without Hoopsville hitting the air!

Tune in tonight starting at 7pm as Dave talks to the two preseason numbers one teams, finds out how the offseason went for the two defending national championships, and touches bases with the men's and women's basketball committee chairs.

Guests include:
- Kevin Vande Streek, men's basketball committee chair and head coach for Calvin

On Hoopsville KVS mentioned that they have a method they're hoping to implement to correct the way the HAM is applied to the SOS. Do you have any idea what this is? Is the NCAA actually changing the way the multiplier is applied?

(I wish someone was around to tell us the multiplier was being applied incorrectly three or four years ago ).

Not changing it... but trying to make sure the math is correct. Your argument, my friend. Your argument is what is spurring this. The idea that when they went to using the multiplier on just the overall record for your away games and your home games, that the multiplier was basically negated. They want the NCAA to make sure that isn't the case. And if it is the case, I believe they will try and have them go back to the way it was - using the multiplier for each individual record for each individual game - not a grand total of records.

I meant to contact you on whether my efforts had gotten you a chance to talk to NCAA stats. If not, I will try again ... because clearly the national committee has been listening to people (you, me, others).

[KnightSlappy]

I just leafed through the new Pre-Championship Manual for the first time this season. It looks like they did finally update the OWP/OOWP/SOS section (page 18-19) to reflect what they're actually doing with summing the opponents' wins and losses to come up with an overall OWP (and OOWP). They still don't fully explain it with the multiplier, however.

[AO]

Believe it not, the 2016-17 basketball season is just days away. But the season can't start without Hoopsville hitting the air!

Tune in tonight starting at 7pm as Dave talks to the two preseason numbers one teams, finds out how the offseason went for the two defending national championships, and touches bases with the men's and women's basketball committee chairs.

Guests include:
- Kevin Vande Streek, men's basketball committee chair and head coach for Calvin

On Hoopsville KVS mentioned that they have a method they're hoping to implement to correct the way the HAM is applied to the SOS. Do you have any idea what this is? Is the NCAA actually changing the way the multiplier is applied?

(I wish someone was around to tell us the multiplier was being applied incorrectly three or four years ago ).

Not changing it... but trying to make sure the math is correct. Your argument, my friend. Your argument is what is spurring this. The idea that when they went to using the multiplier on just the overall record for your away games and your home games, that the multiplier was basically negated. They want the NCAA to make sure that isn't the case. And if it is the case, I believe they will try and have them go back to the way it was - using the multiplier for each individual record for each individual game - not a grand total of records.

I meant to contact you on whether my efforts had gotten you a chance to talk to NCAA stats. If not, I will try again ... because clearly the national committee has been listening to people (you, me, others).

I'm thoroughly confused by your explanation Dave.  They were using the multiplier on each individual game, but they weren't getting their intended effect.  They were making road games more important to SOS, not tougher. 

If they want to fix the multiplier they have to apply it to the opponents W/L pct, not their total Wins and Losses.  They can then apply a weight to this adjusted W/L pct to give less credit to games against teams without a full D3 schedule.

Here is what that would look like:

Under the 2016 calc, the multiplier did nothing as OWP remained at .599 despite the team playing 4 of the 6 games on the road and one neutral court game.   When we apply the multiplier to the W/L pct, the OWP moves up to .655 which we believe would be the most accurate SOS.  If the NCAA wants to continue to give less weight to opponents without a full D3 schedule they can then apply the weighting to each game and give less credit for those games against teams E and F who had only 14 and 15 D3 games.  The OWP then drops to .628 to reflect the fact that we have less confidence in the results against E and F.
Check my work here.

[Dave 'd-mac' McHugh]

I don't have time to read through everything (at the D3hoops.com Classic calling games and running the production), but I thought I would try and clarify what I was saying.

The intended plan with the multiplier is to apply it to each individual team's winning percentage on who they play. Play Team A on the road, multiply by 1.25. Play Team B at home, multiply by 0.75, etc. HOWEVER, the NCAA stats guys adjusted that (despite what was written in the manual) and started taking every team a program played on the road, adding up and getting a total winning percentage and multiplying that number by 1.25. Then adding up all the home opponents, getting an overall winning percentage and multiplying by 0.75. As KnightSlappy has shown to me on many occasions (of these boards), it has a tendency to kill or negate the multipliers for a lot of teams (especially if they have a relatively balanced schedule of home and away as I figure).

The committee is making sure the NCAA stats guys understand this "quirk" with the math and prove the correct numbers and multiplication is being done so the multipliers have an impact like they were designed... not a watered down version that seems less likely to have a multipliers at all.

[AO]

I don't have time to read through everything (at the D3hoops.com Classic calling games and running the production), but I thought I would try and clarify what I was saying.

The intended plan with the multiplier is to apply it to each individual team's winning percentage on who they play. Play Team A on the road, multiply by 1.25. Play Team B at home, multiply by 0.75, etc. HOWEVER, the NCAA stats guys adjusted that (despite what was written in the manual) and started taking every team a program played on the road, adding up and getting a total winning percentage and multiplying that number by 1.25. Then adding up all the home opponents, getting an overall winning percentage and multiplying by 0.75. As KnightSlappy has shown to me on many occasions (of these boards), it has a tendency to kill or negate the multipliers for a lot of teams (especially if they have a relatively balanced schedule of home and away as I figure).

The committee is making sure the NCAA stats guys understand this "quirk" with the math and prove the correct numbers and multiplication is being done so the multipliers have an impact like they were designed... not a watered down version that seems less likely to have a multipliers at all.

Certainly no rush to reply.  Are you saying that the "stats guys method" is the current plan for 2016-17?  They didn't use that in 2015-2016 from what I can tell.  I assume if they do use this new method they would then weight the home, road and neutral percentages by the number of opponent games involved in each of those 3 categories? 

[Dave 'd-mac' McHugh]

I don't have time to read through everything (at the D3hoops.com Classic calling games and running the production), but I thought I would try and clarify what I was saying.

The intended plan with the multiplier is to apply it to each individual team's winning percentage on who they play. Play Team A on the road, multiply by 1.25. Play Team B at home, multiply by 0.75, etc. HOWEVER, the NCAA stats guys adjusted that (despite what was written in the manual) and started taking every team a program played on the road, adding up and getting a total winning percentage and multiplying that number by 1.25. Then adding up all the home opponents, getting an overall winning percentage and multiplying by 0.75. As KnightSlappy has shown to me on many occasions (of these boards), it has a tendency to kill or negate the multipliers for a lot of teams (especially if they have a relatively balanced schedule of home and away as I figure).

The committee is making sure the NCAA stats guys understand this "quirk" with the math and prove the correct numbers and multiplication is being done so the multipliers have an impact like they were designed... not a watered down version that seems less likely to have a multipliers at all.

Certainly no rush to reply.  Are you saying that the "stats guys method" is the current plan for 2016-17?  They didn't use that in 2015-2016 from what I can tell.  I assume if they do use this new method they would then weight the home, road and neutral percentages by the number of opponent games involved in each of those 3 categories?

As I understand it... when the multiplier was added in the first place, it was used for individual games/records... and stayed that way for a number of years. At some point, the NCAA stats guys felt, thought, whatever that the better way to do the math was the other way and made the switch. I really don't know when that was, but it was three or four seasons ago. KnightSlappy discovered it first and our attempts to get to the bottom of it were very confusing as the men's committee wasn't really sure how it all happened - they were just informed of the change and didn't think much of it since Stats was saying it would all work out. KnightSlappy discovered at the end of that season that it wasn't exactly working as it should (since his math was working the same as theirs). He and I worked to get this to the NCAA and it has taken some time, but now the committee has asked the Stats group to prove why their "new" math is better than the intended math. The basic argument is that the "new" math makes the multiplier basically pointless - which I agree. The men's committee has been told by many coaches that the multiplier is worth keeping in the system, so they want to make sure the math isn't making it pointless.

As for your question, they don't weigh it by the number of opponents... the current method is to take all of the teams you play at home and take that total record/winning percentage and multiply it by .75 - same with away games (for get neutral since that is 1.0 which makes no difference). What the committee rather have, wants to prove, or whatever is that EACH opponent have their record multiplied INDIVIDUALLY. That way the weight of the multiplier is put DIRECTLY on each opponent's WL percentage - not an average of all the opponents.

Think of it this way, if a team played ten games on the road and ten games at home, the multiplier is killed off. Ten games at 1.25 and ten games at 0.75 basically will equal 1.0. You have to have an uneven balance to get any weight in either direction. Now, if you go and put the multiplier on each opponent it doesn't matter how many you play on the road at home because the math is influenced by the direct opponent's record, not the sum total.

I hope that makes sense - my brain is a bit fried from basketball here in Vegas! LOL

[KnightSlappy]

The problem isn't that they're applying the multiplier to a group of games rather than the individual games (the associate property tells us it's the same thing) it's that they're applying the multiplier to the opponents' win value and total games played (the numerator and denominator). This has a scaling effect rather than a multiplicative effect (a home game with the 0.75 multiplier becomes less important in the total SOS, not 'easier').

This is because the stats group, based on our communication with them (through Dave), wanted to account for the fact that some opponents play a different number of D3 games (UMPI might play only 12) and didn't want them to carry the same weight in a team's SOS (I don't love this idea but it's not awful).

What AO is suggesting, I believe, is to apply the multiplier to the numerator only (i.e. the opponents' win totals) and leave the games played denominator alone. This would achieve the multiplicative effect of the HAM while maintaining differences in opponents' games played (if they wanted to keep that). They could then sum the adjusted win totals and raw games played totals to come up with the overall SOS.

The other simple solution is to generate a percentage for each game, apply the SOS, then average the percentages (the way the handbook used to spell it out).

[ElRetornodelEspencio]

Should just use Massey ratings.

[Dave 'd-mac' McHugh]

KnightSlappy - I see your point... though, it still sometimes goes over my head. As much as I loved math growing up in school... I got out of it in that depth and never got back into it.

Personally, how they were doing it before the "slight" change seemed to make the most sense to me. How they are doing it now seems to not get the desired affect. Go back to the previous math and we might be able to be happy with the numbers.

I would say, putting the multiplier on the WL% makes the most sense, though I totally understand why the NCAA would want to avoid having a team like UMPI who plays 12 D3 games having a different WL influence... but it is still a percentage in the grand scheme of things. It should work just fine.

[smedindy]

Is the NCAA afraid the UMPI could sneak into "C" consideration based on their 12-game schedule (much like Nebraska Wesleyan could have in the past if they chose to), or that the teams that play UMPI get a benefit if they don't adjust?

To me, if UMPI is 6-6 or 12-12 against D3 competition, it's the same thing.

[Ralph Turner]

Is the NCAA afraid the UMPI could sneak into "C" consideration based on their 12-game schedule (much like Nebraska Wesleyan could have in the past if they chose to), or that the teams that play UMPI get a benefit if they don't adjust?

To me, if UMPI is 6-6 or 12-12 against D3 competition, it's the same thing.

In 2004 University of Dallas went 13-12 and earned a Pool B bid. We just don't have enough Pool B's this year for there to be a bid.

[Gregory Sager]

Is the NCAA afraid the UMPI could sneak into "C" consideration based on their 12-game schedule (much like Nebraska Wesleyan could have in the past if they chose to), or that the teams that play UMPI get a benefit if they don't adjust?

To me, if UMPI is 6-6 or 12-12 against D3 competition, it's the same thing.

In 2004 University of Dallas went 13-12 and earned a Pool B bid.

Gave Sul Ross State a fight in the opening round, too, losing only by 71-67.

[KnightSlappy]

Is the NCAA afraid the UMPI could sneak into "C" consideration based on their 12-game schedule (much like Nebraska Wesleyan could have in the past if they chose to), or that the teams that play UMPI get a benefit if they don't adjust?

To me, if UMPI is 6-6 or 12-12 against D3 competition, it's the same thing.

I agree with you. NCAA doesn't seem to want to count it the same in an opponents' SOS. I think more because 0-12 doesn't feel the same as 0-24. Or, to get more extreme, a 4-0 (or something) from a previous Nebraska Wesleyan squad counting as 24-0 (or whatever).

I don't think it's about them getting into consideration (or not) themselves.

[Dave 'd-mac' McHugh]

To follow up and just make sure it didn't sneak past anyone... while there are two more bids in the NCAA tournament this year (for a full 64-team field)... there are still NO Pool B bids available (not enough teams eligible). Remember, all Pool B teams are in Pool C, but no way any team near .500 will get in as a Pool C.