I have teased the following data a little for a week or so and I finally had a chance to make the bar chart with some updated data. I did a quick analysis of the current (as of Saturday Evening) rankings from the RPI and from Kenpom. If you take the average differences for all teams in each conference and compare them, this is what you get:
As you can see, the Big Ten is, by far, the conference that is under-valued the most by RPI compared to the more rigorous and statistically sound Kenpom efficiency-based metric. On average, Kenpom ranks Big Ten teams almost 30 slots higher than the RPI. I should also note that although I included only the top 16 conferences, the Big Ten is dead last by this measure when all 36 conferences are considered.
Interestingly, the Big East, Big 12, ACC, SEC, and A-10 are all also under-valued, but by a significantly smaller margin. Among the more notable conferences, only the Pac 12 and AAC are over-valued by the RPI.
Now, I will admit that my title is intentionally provocative. It is not like there is some evil mastermind out there who created a formula with the sole intention to screw over the Big Ten. The RPI formula is quite public and ridiculously simple. But, the fact of the matter is, for whatever reason this year, the RPI math is very much working against the Big Ten.
The clearest example of how this plays out is by comparing MSU to Cincinnati. Based on the RPI, Cincinnati has 5 quad 1 victories:
Buffalo (neutral, RPI = 33)
@ UCLA (50)
@ Temple (40)
@ UCF (62)
Houston (home, 21)
MSU has only 3:
UNC (7)
Purdue (14)
@ Maryland (64)
If we instead switch to Kenpom's numbers, the situation is reversed. Cinci now only has 3 quad 1 wins:
@ UCLA (49)
@ SMU (70)
Houston (home, 30)
MSU has 5 quad 1 wins
UNC (10)
Purdue (5)
@ Maryland (39)
@ Indiana (74)
Penn State (28)
As others have pointed out, maybe the problem isn't the RPI or Kenpom, maybe it is just the committee itself (and there lack of.. intelligence). After all, it is pretty hard to think about any sane person looking at the wins for either team and deciding that Cinci has the better resume. MSU has (at worst) two Top 15 wins and Cinci has (at best) only one Top 25 win. It isn't brain surgery.
If I get a chance, I will do a similar analysis for the past few years and see what type of bias may be present. For now, the evidence seems pretty clear that the Big Ten is getting unfairly screwed over.
As you can see, the Big Ten is, by far, the conference that is under-valued the most by RPI compared to the more rigorous and statistically sound Kenpom efficiency-based metric. On average, Kenpom ranks Big Ten teams almost 30 slots higher than the RPI. I should also note that although I included only the top 16 conferences, the Big Ten is dead last by this measure when all 36 conferences are considered.
Interestingly, the Big East, Big 12, ACC, SEC, and A-10 are all also under-valued, but by a significantly smaller margin. Among the more notable conferences, only the Pac 12 and AAC are over-valued by the RPI.
Now, I will admit that my title is intentionally provocative. It is not like there is some evil mastermind out there who created a formula with the sole intention to screw over the Big Ten. The RPI formula is quite public and ridiculously simple. But, the fact of the matter is, for whatever reason this year, the RPI math is very much working against the Big Ten.
The clearest example of how this plays out is by comparing MSU to Cincinnati. Based on the RPI, Cincinnati has 5 quad 1 victories:
Buffalo (neutral, RPI = 33)
@ UCLA (50)
@ Temple (40)
@ UCF (62)
Houston (home, 21)
MSU has only 3:
UNC (7)
Purdue (14)
@ Maryland (64)
If we instead switch to Kenpom's numbers, the situation is reversed. Cinci now only has 3 quad 1 wins:
@ UCLA (49)
@ SMU (70)
Houston (home, 30)
MSU has 5 quad 1 wins
UNC (10)
Purdue (5)
@ Maryland (39)
@ Indiana (74)
Penn State (28)
As others have pointed out, maybe the problem isn't the RPI or Kenpom, maybe it is just the committee itself (and there lack of.. intelligence). After all, it is pretty hard to think about any sane person looking at the wins for either team and deciding that Cinci has the better resume. MSU has (at worst) two Top 15 wins and Cinci has (at best) only one Top 25 win. It isn't brain surgery.
If I get a chance, I will do a similar analysis for the past few years and see what type of bias may be present. For now, the evidence seems pretty clear that the Big Ten is getting unfairly screwed over.
