e-parade

I'm looking forward to hopefully seeing a bunch of good teams visit the Mullins Center in the coming years...it would be nice to try and keep scheduling with Temple though, I always enjoyed watching those games in person.

jpschmack

Yeah, but we don't ALL need marquee OOC wins. VCU and Butler's schedule HAD to be tough, so dropping GW, Richmond, Xavier from their OOC schedules and joining the A-10 makes their overall schedule tougher.

Butler and VCU probably had a OOC schedule harder than necessary lined up becuase they had 12+ easy wins waiting for them in conference.

The difference between leagues like C-USA, WCC, Horizon and CAA and the MWC, MVC and A10 is simply: "Don't schedule for YOUR RPI. Schedule for OUR RPIs"


Our league is now deep enough to operate like a major conference. We should play for 10 easy wins OOC, then four "challenging" games. For the top half, against the toughest teams we have a shot to beat. For the bottom half, the toughest teams possible just to offset the perception of an easy OOC schedule.

If I need to illustrate this using last year's numbers for everyone and how they should schedule, I can and will.

jpschmack

You're right. And when one tries to say it simply, it sounds like the opposite of what's true. Which is why I usually end up having to write long explanations.

The problem with RPI… wait, hold on…

I think the better analogy would be "The NCAA is using on base percentage" and as a result, many teams aren't swinging at anything, crowding the plate, never sacrificing bunting or flying, and getting good OBPs simply by racking up walks and HBP against crappy pitchers.

In a nutshell, the main problem with the RPI is two fold:

#1 - It creates a statistical difference between results that don't matter.
#2 - Conference affiliation dictates the numbers based on factors that don't matter to a team's performance.

Examples for #1.
Two teams play in the same conference and go 10-6 with the exact same results against each team, splitting with each other.
OOC, those two teams play four similar opponents with the exact same results against each team.
In the 10 remaining games, Team A plays 10 teams rated 150-225 and wins each game by exactly 16 points.
In the 10 remaining games, Team B plays 10 teams rated 225-300 and wins each game by exactly 32 points.

Furthermore, let's say the two teams have the exact same roster. No talent differential what so ever. (Think of Create A School in EA Sports NCAA Football, where you can create Xavier and give them "LSU" as their roster. Xavier's roster and LSU's roster have the EXACT SAME TALENT NUMBERS, just different names).

Those teams are perfectly equal on Selection Sunday: 10-6 conference, same marquee wins. Split with each other. 10 cupcake wins. The only difference is the FLAVOR OF THE CUPCAKE.

If you're an NCAA team, you should go 10-0 vs teams 150+. But there's an RPI difference in those two teams. For no real "reason" that makes any sense.

Examples for #2
Because it's math, it can be manipulated if a conference works together after figuring it out. Like I'm talking about now.

Let's say you have 9 mediocre teams. Each plays 14 OOC games against horrible teams and goes 14-0. They play 16 conference games each. The top four teams will have great records, and have wins over teams with great records (each other), but they really all suck.


Now, obviously, it doesn't quite work like these EXTREME examples. But we can make it work for us.
By all scheduling 10 easy wins OOC, we can start 160-0 vs cupcakes.
In the other four games each, we play "the best teams we have a shot of beating beating" for the top half, and "We're gonna get creamed but it makes our SOS stronger" for the bottom teams. And we go 16-48 in those games.

We're 176-48 OOC as a conference. When we go .500 against ourselves, our top half can count all the games vs the top half as marquee games, and we can get a lot of teams in the dance.

This is what big conferences do. We can do this now. We just need to intelligently schedule those "teams we can beat" games to get our marquee OOC games and give some teams a little credit in December before they lose 2-4 times in conference.

xavierhoops

I see what you're saying but you would never convince the very top of the conference to go along with scheduling 10 easy OOC wins and only 4 tough games. Assuming at least 2 of those 4 are in some sort of holiday destination tournament, that only leaves 2 spots for marquee OOC games under your scheduling philosophy.

I'm most familiar with Xavier, so I'll use them as an example. Your league scheduling philosophy is drastically different from Xavier's (which has served X extremely well over the last 20 years). In fact, it's the exact opposite. Xavier scheduled 10 "challenging" games (pretending X didn't suck in Hawaii and played 2 good opponents) and 4 "easy wins" last year, and the easy wins included Oral Roberts and Miami. You could show whatever RPI numbers you wanted, but seeing as X has made 11/12 using it's current model, you'd never convince Mercurio to totally flip his scheduling philosophy from 10 challenging games down to 4.

bprichard

You don't really need the top teams to do it, though. X will get their wins regardless. It's the middle and bottom teams who need to get on board.

xavierhoops

Problem is, not every team in the top half of the A-10 is a marquee team and considered a "good win" for Selection Sunday. I'd say there are about 3, sometimes 4, A-10 teams per year that can really be counted as "marquee" games or considered really good wins.

duq81

You're missing his point. You're looking at things as they are. He's looking at things as they could be. We really weren't that far from what he's suggesting. Only URI, and GW truly pooped the bed ooc last year. The other schools either reached his goal, or were short by a few games. Even Fordham wasn't a disaster ooc.

xu95

I know a couple of people have already mentioned it, but I believe the majority of the ADs and coaches are against an 18 game schedule. The ones who are for it are the ones who have a hard time filling out their OOC schedule.

antboy

As I said, the rpi is flawed. But Team A in your example above has the better record. Everyone knows that NCAA teams "should" beat teams rated 150 and above. But upsets do happen, and slightly better teams are more likely to pull an upset than worse teams. If team A beats those higher rated teams, they've demonstrated better performance. The actual quality of these teams is an abstraction, something we can't really know for sure, but can only use the data provided (in this case, that season's game results).

Also, margin of victory, which some ratings use, is an awful measure, because of all the things that happen in a blowout. Some teams and coaches play the subs more, resulting in 20 point wins instead of 30 point wins. Free throws inflate close games into blowouts that don't reflect the real difference between the teams. Ratings should devalue margin of victory greatly if it is used at all, so that a 30 point loss is not really that much different than a 16 point loss.

This is why a conference's overall rpi should never be used to compare conferences. Only their OOC rpi average should be used. I haven't seen whether the committee uses only OOC rpi or overall rpi, but if it's the latter, they are using it incorrectly. This would negate the gaming of the system that you suggested. It would also help with the way the BCS conferences, they of the many home games and few, if any, true road games, inflate their records, which then inflates the overall conference rating.

I think the rpi is imperfect, I just haven't seen another rating that does enough to penalize teams for playing all their games at home. If the selection committee went back to emphasizing the individual team's rpi, which does account for home-game bias, it might encourage these teams (including Dayton from last year, who played 2 true road games OOC and lost them both) to go out and play more road games. But we all know why they won't do that.

X-man47

Xavier's coach is quoted on the XavierHoops board as saying it will be 16 games with one partner and 14 one-only games.

__________________
Xavier always goes to the NCAA tournament, and almost always makes some noise while there. Projecting anything less than that this season feels like folly.
Eamonn Brennan, ESPN (Summer Shootaround, 2012)

buckswope

http://espn.go.com/blog/collegebaske...sly-tough-a-10

Not suggesting Andy Katz is more informed than Xavier's coach, but Katz did imply a 16-game schedule yesterday when he said Butler would need to find 3 new OOC opponents and now he's saying the opposite. It's possible that he meant to say something else (note the 5am timestamp) and it'll be subsequently corrected - but it's also possible that the plan has changed in the last 24 hours.

RF1

To my knowledge, this is speculation based on the comments of a few coaches that made it into a published report.

xavierhoops

Everyone beating 10 cupcakes a year and 1, maybe 2, challenging games does not magically make the entire top half of the league (8 teams) all marquee games. (For schools in the middle and near the top it might... but doesn't for the schools at the very top who can get multiple marquee games on their OOC schedule)

Chalmers0

Brad Stevens was just on The Dan Dakich Show in Indy and they discussed a 16 game schedule.

Also said they were very unsure of what to do regarding the Xavier/Butler ESPN Tip-Off marathon game. Mentioned the possibility of that being the conference game but said obviously neither team would really want that to be the case.

MudFlap

Simple. X and Butler do a home and home and play everybody else once.