RPI algorithm differences illustrated

[GoneGolfin]

Recently I have been monitoring some significant differences in what the Adjusted RPI reports and what the Pseudo RPI reports (Boyd Nation's). In the past I have talked about the significance of not counting ties in the RPI (the Pseudo RPI does not count ties). Because of this, Gonzaga has been ranked ahead of Louisiana each day this week (in the Pseudo RPI) ... when they should not have been (as illustrated in the Adjusted RPI).

Going into today's action, we had Louisiana with an Adjusted RPI of .5500, good for a rank of #55. Meanwhile, Gonzaga has an Adjusted RPI of .5484, good for a rank of #58.

However, the Pseudo RPI has Louisiana at .550 #56 and Gonzaga at .551 and #55.

Why the disparity between the RPI algorithms w/respect to Gonzaga? It comes down to a single tie against Notre Dame that is accounted for properly in the Adjusted RPI and is dropped as if the game was never played in the Pseudo RPI.

Because the Pseudo RPI does not count ties, Gonzaga's WP component improves from .6351 to .6389. However, the real impact is in the OWP (as always). Dropping the tied Notre Dame game from Gonzaga's schedule results in an OWP increase from .5030 to .5055 (remember that OWP is half the RPI formula). This is because Notre Dame has a poor WP (.4118) resulting in a rather negative impact on Gonzaga's OWP. Not even taking into account the increase in OOWP (which is negligible), we come to a Base RPI calculation of .5482 (an increase of .0022). When factoring in the RPI bonus, we have a Pseudo RPI of .5506 (which Boyd rounds to .551 for reporting purposes only ... not ranking purposes) ... vs. the correct .5584 that the Adjusted RPI reports.

Hence, this .0022 difference is worth three spots in the RPI rankings for Gonzaga.

All of this is corroborated in the official NCAA RPI released this past Tuesday (through last weekend's games). In the official release, Gonzaga was #56, immediately behind Louisiana at #55. This was also the case in the Adjusted RPI (.5502 vs. .5500). However, the Pseudo RPI had Louisiana at #56 and Gonzaga at #55.

Brian

[CajunRed]

Brian...53 RPI ranking as of this morning. If we would win either the SBC regular season or the tourney, we are in a regional. How high of a RPI ranking do we need if we DON'T win either of the above?


igeaux.mobi

[RaginDave]

_ Brian...53 RPI ranking as of this morning. If we would win either the SBC regular season or the tourney, we are in a regional. How high of a RPI ranking do we need if we DON'T win either of the above?


igeaux.mobi _

Boyd's World is stating that we need to finish 9-2 to have an RPI inside 45 which I think would guarentee an at large regional bid.

[CajunRed]

Man...I hope you and Boyd's are correct!

Boyd's World is stating that we need to finish 9-2 to have an RPI inside 45 which I think would guarentee an at large regional bid.

igeaux.mobi

[J-Town Cajun]

I would think finishing 2nd in the regular season SBC race and winning a couple in the tourney along with an RPI in the top 45 will get us an at large, any less than that is a long shot.

[GoneGolfin]

_ Brian...53 RPI ranking as of this morning. If we would win either the SBC regular season or the tourney, we are in a regional. How high of a RPI ranking do we need if we DON'T win either of the above?

How high depends on a slew of other variables. The Cajuns could finish with a Top 40 RPI ranking and still miss the field. They can finish with a Top 60 RPI ranking and make the field. Plenty of other variables need to be filled in.

Based on average past history, they would not be in the field at this time. But are not too far away. The Cajuns are currently third in the pecking order in the SBC (Troy, FIU, UL, FAU, WKU) with UL, FAU, and WKU tightly packed.

Much will be determined how the Cajuns perform against Troy and Rice ... as this represents four more games against the Base RPI Top 50. The Cajuns can really help themselves by winning at least 2/4.

Brian

[GoneGolfin]

_ Man...I hope you and Boyd's are correct!

You do realize that this changes daily and is dependent on many independent moving variables? In the end it may be the case that 7-4 gets us into the Top 45 RPI ... or it may be that 11-0 still leaves us outside the Top 45 RPI.

Brian

[raginsaints]

Mama always said math is the devil