Wichita State's Opponent RPI and Record

[The Mad Hatter]

@Jamar Howard 4 President:, thanks for doing the math. Considering that these projections calculate a 72.12% chance of 4 losses or better, if we keep doing what we are supposed to do, top 10 in the RPI is a perfectly reasonable expectation.

[ShockTalk]

Mad Hatter, you are correct. KU currently leads the way with an expected RPI of 6.1 yet obviously by leading the way, they should be #1. Expected RPI is an average of 10,000 simulations, so an expected RPI of 1 would mean 1 in each of the 10,000 sims. (not going to happen) For the top 15-20 teams, this affects things a little. After that, the phenomenon just doesn't factor in much. Anyway, not sure if this is the right way to look at things, but as of today:

KU - 6.1 = #1
WSU - 14.8 = #8
UNC - 24.0 = #20

If that can be translated to into the forecast, we might see something like the following for WSU:

Final Record . Expected RPI . Probability
30-0 . . . . . . . 1 . . . . . . . 1.56%
29-1 . . . . . . . 2 . . . . . . . 8.22%
28-2 . . . . . . . 3 . . . . . .. 17.11%
27-3 . . . . . . . 5 . . . . . .. 23.42%
26-4 . . . . . . . 8 . . . . . .. 21.81%
25-5 . . . . . . . 18 . . . . .. 15.24%
24-6 . . . . . . . 25 . . . . . . 7.73%
23-7 . . . . . . . 34 . . . . . . 3.43%
22-8 . . . . . . . 44 . . . . . . 1.10%
21-9 . . . . . . . 55 . . . . . . 0.28%
20-10 . . . . . . 65 . . . . . . 0.08%
19-11 . . . . . . 72 . . . . . . 0.01%

If I'm figuring things right, WSU should finish with a top 10 RPI if they can keep their losses to 4 or fewer. Amazing!

Those losses would also need to be reasonably expected losses, not bad losses.

[rrshock]

Short of a number of unexpected quality wins, the Valley's RPI has already started to take shape. rpiforecast.com has the MVC's winning % increasing slightly over the rest of the OOC schedule, but has the MVC dropping to 11th in RPI. Probably due to a slightly easier 2nd half of the OOC which may also decrease the number of quality games to get those upsets. Your assumption also would require Valley teams to not have any more the bad loses that they are currently having.

You mean they can have more bad losses than they are currently having?

[The Mad Hatter]

Those losses would also need to be reasonably expected losses, not bad losses.

Actually I don't think that is true. RPI take the overall winning percentages into account, and isn't based on where those percentages come from. WSU's winning percentage would be the same regardless of where the losses came and our opponent winning percentage would be the same as well. The only bad losses that could hurt WSU's RPI would be our opponents failing to meet their projected win totals.

Bad losses hurt your resume just like good wins help your resume, but your overall RPI rank doesn't reflect that.

[jdmee]

I think that is correct as well

RPI = (WP * 0.25) + (OWP * 0.50) + (OOWP * 0.25)

if we lost to a 24-3 team and won vs a 3-24 team our WP = .5 and our OWP =.5 and the OOWP doesn't change
it would be the same if it was reversed.

The only factor is the 1.4/0.6 Home vs Away.

[MadaboutWu]

Those losses would also need to be reasonably expected losses, not bad losses.

I feel like we have been through this before, but is not the case that against whom we lose does not matter--only have many losses given the same cast of opponents. If some of our bad losses are worse than expected we would make up for it with better than expected wins. And the ratio of our record, our opponents' records, and our opponents' opponents' record does not change.

[MadaboutWu]

jdmee beat me to it. TMH or JMH4prez, is that right?

[ShockTalk]

You mean they can have more bad losses than they are currently having?

Funny guy. Unfortunately, whether they have more bad losses than currently or just one more bad loss is no longer funny.

[The Mad Hatter]

jdmee beat me to it. TMH or JMH4prez, is that right?

I believe so. In other words, the only bad losses are home losses as far as overall RPI is concerned (although even there, at least in conference play, if all your losses were at home and your expected losses were on the road, it would still even out in the end).

[RoyalShock]

Actually I don't think that is true. RPI take the overall winning percentages into account, and isn't based on where those percentages come from. WSU's winning percentage would be the same regardless of where the losses came and our opponent winning percentage would be the same as well. The only bad losses that could hurt WSU's RPI would be our opponents failing to meet their projected win totals.

Bad losses hurt your resume just like good wins help your resume, but your overall RPI rank doesn't reflect that.

But a "bad loss" is usually an unexpected loss, which does affect their record and does affect our RPI and subsequent rank.

[ShockTalk]

From a team's current standing, or projected standing, if that team has a loss that is unexpected (bad), they will also need an equally unexpected win (good) to counter it. We no longer have that kind of win available to us. In fact, the spread favors WSU in all remaining games. Our projected RPI is based upon reasonably expected losses where we were favored by the spread.

[jdmee]

If we are going to end up having 4 losses on the season, then for RPI purposes it doesn't matter who you lose to. If we have 4 "bad" losses, then our record is 26-4 Our OWP remains the same and the OOWP remains the same. This is the same as if our losses were 4 "good" losses.

Now when talked about at large bids and Seeding then it matters quite a lot on who we lose to. A bad loss is a reason to drop a seed line, quality wins will can raise a team a seed line. This is why the NCAA Tourney seed is not strictly based upon RPI. WSU will end the season with probably no real top quality wins that will let us raise a seed line come Tourney time. Any loss we have will probably be used to drop us in the seeding.


Now each out of conference loss a team in the MVC loses will directly lower our RPI. These losses drop our OWP.

[The Mad Hatter]

From a team's current standing, or projected standing, if that team has a loss that is unexpected (bad), they will also need an equally unexpected win (good) to counter it. We no longer have that kind of win available to us. In fact, the spread favors WSU in all remaining games. Our projected RPI is based upon reasonably expected losses where we were favored by the spread.

Not exactly, because we are actually favored in every game we play from here on out by Sagarin, so the 4 projected losses are not losses that are expected by those measures, but losses that are expected just because the statistical models suggest that it is unlikely you will win every game, even if you are favored to do so. In other words, the RPI Forecast model is already taking into account the losses that would be unexpected by projecting the games individually. If we went game by game, we would be expected to go undefeated and be #1 in the RPI. As long as we have 4 or fewer losses, it doesn't matter who they come to for us to be a top 10 RPI team.

As long as we take care of business and lose 4 or fewer games from here on out, the only thing that could hurt our overall ranking is our opponents winning fewer games than expected against non-WSU competition.

[ShockTalk]

I stepped away for a moment, so without reading any new posts, I'M WRONG!

If a projection has you only losing 3 more games and that is where you end up, you would have had to beat a team they projected you to lose to. They counter each other.

That said, don't think for a second that the polls or selection committee is not affected adversely.

[The Mad Hatter]

That said, don't think for a second that the polls or selection committee is not affected adversely.

No argument on that front. Bad losses/good wins are definitely a factor for polls and in the selection process (as they should be).