polish
03-07-2006, 01:09 PM
Despite high-major protestations, the Missouri Valley Conference has not cracked any sort of mystical RPI code that has allowed six of its teams to reside in the top 45. Rather, the secret to the Valley's success is twofold. One part mostly anyone who follows hoops understands. The other, I'm not sure anyone does.
The bulk of the Valley's RPI surge has come the old-fashioned way: smart nonconference scheduling combined with lots of wins in those games. Anyone who follows college hoops and the RPI recognizes this recipe -- it's just usually followed by megaconferences with the word "Big" in them. The result: The Valley had a bunch of teams enter league play with gaudy records and strong RPIs, and the parity in the league helped preserve them.
The kicker that helped create the Valley's "RPI perfect storm," though, comes from the highly flawed adjustments the NCAA made to the RPI formula last season.
In an attempt to push teams out onto the road, the NCAA starting weighting home wins and road losses as 0.6 wins and road wins and home losses as 1.4 wins in Factor I of the RPI. The problem is that no one who was involved in creating the adjusted formula realized that not all 0.6s and 1.4s are created equal.
Check out this simple example:
• Team A (6-4 overall, 5-0 at home, 1-4 on the road)
Its adjusted record is 4.4-2.4 (.6470 win percentage, or .0470 more than its unadjusted 6-4 mark).
• Team B (6-4 overall, 3-2 at home, 3-2 on the road)
Its adjusted record is unchanged at 6-4 (.6000 win percentage).
Team B actually won two more road games than Team A, but is now significantly behind Team A in Factor I. Why? Because the new NCAA formula actually rewards teams for not losing at home much more than it rewards them for winning on the road.
How does this apply to this year's Missouri Valley (and Big Ten)? Since practically all of the home-and-home series between the top six teams were split, with the home teams winning both times, there was less impact on both teams' overall adjusted win percentages than if the split had occurred with both road teams winning.
Here's an example from this season:
Wichita State and Southern Illinois each won at home in their two meetings this season. Remove those two games from each team's record, and you have the following: Wichita State is 23-7 (.7667 winning percentage) and Southern Illinois is 21-9 (.7000 winning percentage).
Assume, for simplicity, these are the teams' adjusted records, too. Because both teams' marks are well above .500, a split is better for both teams if the home team wins each game (shown in the table below).
There's no place like homeWSU Adj. Win Pct.SIU Adj. Win Pct.Home teams win both23.6-7.6 (.7564)
(-.0103 lower)21.6-9.6 (.6923)
(-.0077 lower)Road teams win both24.4-8.4 (.7439)
(-.0238 lower)22.4-10.4 (.6829)
(-.0171 lower)
Why is this so? In simple terms, adding in 0.6s (for home wins and road losses) makes a smaller impact on the overall adjusted winning percentage than adding in 1.4s (for road wins and home losses) if you go 1-1 in those games. Again, the NCAA formula rewards teams (or in this case, punishes them less) for winning at home.
How much of an impact can this have? That 0.0135 difference above (in Wichita's profile) would have been the difference between 32nd to 38th in Sunday's RPI calculations -- a critical difference in the at-large selection process. Just some food for thought the next time the Valley comes up in conversation. Don't blame it for taking advantage of a badly flawed system.
Kinda Interesting
The bulk of the Valley's RPI surge has come the old-fashioned way: smart nonconference scheduling combined with lots of wins in those games. Anyone who follows college hoops and the RPI recognizes this recipe -- it's just usually followed by megaconferences with the word "Big" in them. The result: The Valley had a bunch of teams enter league play with gaudy records and strong RPIs, and the parity in the league helped preserve them.
The kicker that helped create the Valley's "RPI perfect storm," though, comes from the highly flawed adjustments the NCAA made to the RPI formula last season.
In an attempt to push teams out onto the road, the NCAA starting weighting home wins and road losses as 0.6 wins and road wins and home losses as 1.4 wins in Factor I of the RPI. The problem is that no one who was involved in creating the adjusted formula realized that not all 0.6s and 1.4s are created equal.
Check out this simple example:
• Team A (6-4 overall, 5-0 at home, 1-4 on the road)
Its adjusted record is 4.4-2.4 (.6470 win percentage, or .0470 more than its unadjusted 6-4 mark).
• Team B (6-4 overall, 3-2 at home, 3-2 on the road)
Its adjusted record is unchanged at 6-4 (.6000 win percentage).
Team B actually won two more road games than Team A, but is now significantly behind Team A in Factor I. Why? Because the new NCAA formula actually rewards teams for not losing at home much more than it rewards them for winning on the road.
How does this apply to this year's Missouri Valley (and Big Ten)? Since practically all of the home-and-home series between the top six teams were split, with the home teams winning both times, there was less impact on both teams' overall adjusted win percentages than if the split had occurred with both road teams winning.
Here's an example from this season:
Wichita State and Southern Illinois each won at home in their two meetings this season. Remove those two games from each team's record, and you have the following: Wichita State is 23-7 (.7667 winning percentage) and Southern Illinois is 21-9 (.7000 winning percentage).
Assume, for simplicity, these are the teams' adjusted records, too. Because both teams' marks are well above .500, a split is better for both teams if the home team wins each game (shown in the table below).
There's no place like homeWSU Adj. Win Pct.SIU Adj. Win Pct.Home teams win both23.6-7.6 (.7564)
(-.0103 lower)21.6-9.6 (.6923)
(-.0077 lower)Road teams win both24.4-8.4 (.7439)
(-.0238 lower)22.4-10.4 (.6829)
(-.0171 lower)
Why is this so? In simple terms, adding in 0.6s (for home wins and road losses) makes a smaller impact on the overall adjusted winning percentage than adding in 1.4s (for road wins and home losses) if you go 1-1 in those games. Again, the NCAA formula rewards teams (or in this case, punishes them less) for winning at home.
How much of an impact can this have? That 0.0135 difference above (in Wichita's profile) would have been the difference between 32nd to 38th in Sunday's RPI calculations -- a critical difference in the at-large selection process. Just some food for thought the next time the Valley comes up in conversation. Don't blame it for taking advantage of a badly flawed system.
Kinda Interesting