In part 2 of this question describe how RPI relates to rankings in a Hoops Dynasty world.

Oh..oh...wait! I know this one! Both begin with the letter R and have the letter I in them.

But seriously, no one has a clue how rankings are determined and therefore no idea how (or if) RPI relates to it. The good news is that rankings are being revised as part of the overall engine change that is currently in beta testing. Whatever ranking system we have now will not be here much longer.

weena, i had allways assumed you were a guy. but just read "Time Traveler" by HG Wells. and, one of the main characters from the future is .... Weena! but, this weena is, it would seem, female.

what gives?

Formula

The exact formula for determining the RPI of a team at any given time is as follows.

RPI = (WP * .25) + (OWP * .50) + (OOWP * .25)

Where WP is Winning Percentage, OWP is Opponents' Winning Percentage and OOWP is Opponents' Opponents' Winning Percentage.

As mentioned, the WP is weighted based on the locations of the games being played by the team (the other two; OWP and OOWP are not). The exact value of a home win or away loss is .6, whereas the value of an away win or home loss is 1.4. Games played at neutral locations count as 1. When calculating this number, games against subdivision opponents are ignored whether the game is won or lost. As an example, if a team loses to Duke at home, beats them away, and then loses to Cincinnati away, their record would be 1-2. Considering the weighted aspect of the WP, their winning percentage is 1.4 / (1.4 + 1.4 + .6) = .4117

The OWP is calculated in a similar way, ignoring location (every game is worth 1 point) and games against subdivision teams. It does have a few added restrictions, though: first, it ignores the game/games played against the team in question. It also requires that the Opponents being examined have their (unweighted) WP's created before averaging them together. Continuing from above, Duke has played one other game and lost, while Cincinnati has played two other teams and won. Thus the OWP of the team is (0/1 + 2/2)/ number of opponents. OWP = .500

The OOWP is the most elusive number for most people to calculate, but there are a few tricks for finding it. The OOWP is another way of saying your Opponent's OWP; e.g., it would ignore Duke and Cincinnati in the examples above, but would not ignore the team for which you are calculating the RPI. To reiterate, the team in question is part of the team's OOWP (Opponents' Opponents' Winning Percentage). Ignoring this aspect is one of the most common mistakes made in calculating RPI, for, mathematically, it shouldn't be there, and the most re-occurring opponent of your opponents is yourself. Ignoring the team in question causes a large margin of error. Also, you cannot just average the OWP of your direct opponents, as doing so ignores the fact that every team has played a different number of teams; this would unfairly weight the opponents of teams who have played less. Continuing the example above, a team has played Duke twice and Cincinnati once. Duke has played one other game and lost, while Cincinnati has played two other games and won.

Next, for simplicity, assume none of the unnamed teams has played any other games. The OOWP will have these teams this main times in its formula:

(team in question WP) + (team in question WP) + (unknown Duke Opponent WP) + (team in question WP) + (unknown Cincinnati opponent 1 WP) + (unknown Cincinnati Opponent 2 WP) / (number of teams listed).

Plugging in the records and calculating the WP produces the following:

(0/1)+(0/1)+(0/0)+(1/2)+(0/0)+(0/0).

Duke or Cincinnati must be ignored, as they are each other's root opponent. The team in question's record changes for the same reason. Ignoring Duke, this team is 0/1; ignoring Cincinnati, the team is 1/2. The resulting OOWP is .5/6 = .0833

(WP * .25) + (OWP * .50) + (OOWP * .25)

Plugging this in gives you

(.4117 * .25) + (.500 * .50) + (.0833 * .25) = .3737 = RPI

it seems fairly straightforward, but every time i try to work through the steps, it makes my head spin

here is HD's definition of RPI, stright from the FAQ:



A: The RPI (Rating Percentage Index) is a measure of strength of schedule and how a team does against that schedule. It does not consider margin of victory. It is used by the NCAA as one of their factors in deciding which teams to invite to the NCAA tournament and where to seed them. The basic formula is 25% team winning percentage (WP), 50% opponents' average winning percentage (OWP), and 25% opponents' opponents' average winning percentage (OOWP). When calculating winning percentage (WP), the RPI weights a road win as 1.4 x a normal win, a home loss as 1.4 x a normal loss, a road loss as 0.6 x a normal loss, a home win as 0.6 x a normal win and neutral site games are weighted at 1.0.

couldnt find any official HD explanation of the ranking system.

as W says, rankings are being revised.

One thing that is clear is this, if you take two teams with similar RPI, where TeamA is 20-2 with a very weak SOS and TeamB is 16-6 with a very strong SOS, teamA will almost allways be ranked higher.

my original assumption (which i think is logical and I hope they adpot in the new formula) was that Prestige or some sort of value that assesses the team's recent history would be included. and maybe it is in there in some respect. But to me, it seems that the previous season record is given quite a lot of weigth (especially early in the season) but that seasons prior to that seem to have very little if any impact.

not sure if any of that helps

Quote: Originally posted by oldave on 1/26/2010weena,   i had allways assumed you were a guy.  but just read "Time Traveler" by HG Wells.   and,  one of the main characters from the future is .... Weena!  but,  this weena is, it would seem,  female.what gives?

Your original assumption is correct, oldave. I am 100% male. Wells was always a little bit odd anyway and the misuse of a fine male name proves it. There is a German band that uses the Weena name but I am not German either. My roots are from the British isles so any non-American uses of the name you may find in a Google search are not me either. Also, I am not a type of soil or a street in Rotterdam either although both use the name.

I have posted various accounts of the origin of my ubiquitous ID in various forums on WIS and one day I am sure I will likely post yet another version. There used to be a guy on WIS who used the ID "Princess". At least I don't have to try and explain that one.

excelllent explanation as usual, dubya.