RPI Forecast Tool

[GT Hawkaholic]

I posted this in another thread but thought it was pretty cool. Buddy of mine found it so I can't take credit!

You can change past games and predict future games as well. I put in having us win every remaining home game and winning road games at Rutgers and at NW and it has us at 19-12 (11-7) with an RPI of 62.

http://www.rpiforecast.com/wizard/Iowa.html

Enjoy!!

 

[El Simbolo]

Wow, that ranking is volatile. Pre Purdue we were 131, but if we had beaten Nebraska we would have been 86 before paying Purdue?

 

[relishingwinning18]

Wow, that ranking is volatile. Pre Purdue we were 131, but if we had beaten Nebraska we would have been 86 before paying Purdue?

You have to wonder, if we were to get up for consideration/bubble team, if they would take into account the absence of Cook during the Omaha loss.

 

[El Simbolo]

Will they discount Duke's loss without the tripper? Giving teams some slack because of an injury makes sense to me, but that lose for Duke should count 100%

 

[El Simbolo]

Looks like our ceiling is RPI of 23 if we win out and finish 24-7. Would they really seed a 16-2 big ten champ that won 15 in a row as a 5-6 seed?

 

[El Simbolo]

You can change past games and predict future games as well. I put in having us win every remaining home game and winning road games at Rutgers and at NW and it has us at 19-12 (11-7) with an RPI of 62.

!!

I hope we win out, but 19-12 seems a best reasonable case scenario - let's get it started Sunday!

 

[El Simbolo]

Really hard to take serious a ranking system that would move us from projected 105 to projected 85 had we won a game we lost in double OT. That makes no sense

 

[StormHawk42]

RPI just isn't worth worrying about until February. Interesting yes, but you'll get heartburn worrying about it.

 

[GT Hawkaholic]

Really hard to take serious a ranking system that would move us from projected 105 to projected 85 had we won a game we lost in double OT. That makes no sense

That would have been a road win vs a top 50 RPI team. We just beat a top 50 RPI team at home and moved up 20 spots for reference.

 

[El Simbolo]

I understand that midway thru the season RPI is more volatile, but if I'm reading the site right it predicts that we'll finish at #105, but if the season played out the same as their prediction but we won in Lincoln it predicts like #85!at the end of the season. You'd think by seasons end with full 30 games played that scoring 1 more point in regulation of one game against a good but not great team wouldn't affect your ranking by 20 spots

 

[DavenportHawk8]

I understand that midway thru the season RPI is more volatile, but if I'm reading the site right it predicts that we'll finish at #105, but if the season played out the same as their prediction but we won in Lincoln it predicts like #85!at the end of the season. You'd think by seasons end with full 30 games played that scoring 1 more point in regulation of one game against a good but not great team wouldn't affect your ranking by 20 spots

It is also based on how your opponents do. So if ISU & Purdue can get hot and win games down the stretch and move up. Iowa's RPI will also move up. If teams like Delaware St, Kennesaw St, and UNI keep losing, its going to affect Iowa's RPI.

RPI is a good tool, but I think you are seeing teams "take advantage" of the formula and schedule according to it.

Nevada: 14-3 RPI 42 (Record vs top 50 0-1, 100-200 11-2, 200+ 3-0). So this team has played 1 top 50 team and lost. The rest of their games are teams 100-200 and they have won a fair share, but have 2 losses to teams RPI 100+ (same as Iowa). They played majority of their games in the "sweet spot" 100-200 teams.

New Mexico St: 14-2 RPI 75 (played 2 non div 1 opponents so they don't count) They have not played a team in the top 100. All their games are vs teams 100+. How can that put them in the top 100 in RPI, when they haven't even played a team in the top 100? Makes zero sense.

RPI is a good tool to use, but I think the committee doesn't rely solely on it. SOS and who you played in the non-conf I think carries some weight in the post season selection. I thought Iowa had 13 non-conf games and they played 6 "good teams" and 3 mediocre teams and then 4 REALLY BAD teams.

 

[hawkeyeguy69]

You have to wonder, if we were to get up for consideration/bubble team, if they would take into account the absence of Cook during the Omaha loss.

I think more importantly, they would look at how we finished. If you get rid of the 4-5 start and look at say an 15-7 (11-7) finish, then that would be something to consider. We would have to win 2 BTT games tho. So, 17-8 with a 21-13 record would get us some consideration. The RPI would bump up a little more than 62 too, with those BTT wins.

 

[DodgerHawki]

That would have been a road win vs a top 50 RPI team. We just beat a top 50 RPI team at home and moved up 20 spots for reference.

Bingo. RPI doesn't factor in margin of victory at all, and gives extra weight to road wins and home wins are counted less. The RPI is a helpful tool, but should never have become the only metric of importance for the committee. It was good to see that the committee will be considering using other metrics in future years.

If Iowa really wants to make huge RPI leaps, start winning road games against Top 50 RPI teams like Illinois, Northwestern, Minnesota, Maryland, Michigan State, Wisconsin. Even a road win against Rutgers would help out.

 

[El Simbolo]

If we simply hadn't played our three 300+ opponents are expected final RPI changes from 105 to 90. How much harder is it to schedule 200-300 teams instead of 300+?

If we had replaced our five worst teams (all +267) with 225-250, our expected only drops from 105 to 93. You guys must be right that the sweet spot is somewhere below 200. Do you think we could have gone 5-0 at home against teams between 150-200? Since Omaha is 156, predicting 4-2 against there theoretical five plus Omaha is reasonable. I wonder if that 4-2 would yield better RPI than we currently have?

 

[El Simbolo]

I tested 150-200. Going 4-1 against them would land us 14 better than our current 105 (91)

5-0 would be 28 better than 105 (77)

 

[El Simbolo]

If you add my 4-1 versus 150-200 to GTs Calc above it would land us at 18-13 instead of 19-12 and RPI of 55 instead of 62. Gotta quit doing this and go walk the dog...

 

[DavenportHawk8]

That tool will drive you crazy.

Just know the NCAA is looking at changing the RPI and the selection criteria. It won't happen until 2018, but I can guarantee you that they will not be putting much weight into it this year.

They will use it as a baseline, but if a team has an RPI of 50 and played literally no one in the non-conf they will most likely be left out. If a team with an RPI of 70, but played a tough non-conf, they will most likely be in. Iowa played (Seton Hall, UVA, ND, and ISU all top 50 teams). They are 2-5 vs top 50 (not bad). They are 2-2 100-200 (not great, but one of those L's will hopefully move up Memphis).

Iowa's chances:

12 wins: They are in the dance. No doubt they will have the wins and RPI won't matter.
11 wins: Bubble, will need at least 2 wins in DC
10 wins: Bubble outside looking in need at least 3 wins
9 or lower: Going to need to win it all in DC.

 

[Joes Place]

RPI just isn't worth worrying about until February. Interesting yes, but you'll get heartburn worrying about it.

RPI is really over-rated as a metric, which is why NCAA Selection Committee is downplaying it. As I'd posted in another thread, it is an "unbiased" metric of the whole season - it weights a first game W/L the same as a last game W/L, not taking anything into account the improvement a team makes during the season.

A 'weighted' RPI (which is basically what lots of other polls use) would split seasons into something like 'thirds', and weight the last 1/3 of the games much more than the first 1/3. Which is really how it should be: the performance in the last 1/3 of the season is more indicative of how a team will perform in the tournament, and how good the team actually is vs. averaging the performance out over the whole season. Finish in the top 5 of a power conference that is recognized as a competitive league, and you will play in the NCAAs. For Iowa, that probably means finishing 12-6 or maybe 11-7 in the B10. If Iowa is 'tied' for 5th with 3 or 4 other teams, the likelihood of being left out is much higher, because the B10 will probably get 6 or 7 teams in the NCAAs at most.

 

[L. Wade Childress]

Really hard to take serious a ranking system that would move us from projected 105 to projected 85 had we won a game we lost in double OT. That makes no sense

Road games are weighted 1.4 vs .6 in the own record part of the calculation.

The SOS component is not home/away weighted.

RPI is seemingly arbitrary but it gives a transparent target for every coach to manage towards each year so it's not all bad. It measures your actual resume, not how good you are.

 

[L. Wade Childress]

I'll believe the comments about using RPI less when I see it. High major at large berths almost directly correlate to RPI.

 

[SotaHawk87]

That tool will drive you crazy.

Just know the NCAA is looking at changing the RPI and the selection criteria. It won't happen until 2018, but I can guarantee you that they will not be putting much weight into it this year.

They will use it as a baseline, but if a team has an RPI of 50 and played literally no one in the non-conf they will most likely be left out. If a team with an RPI of 70, but played a tough non-conf, they will most likely be in. Iowa played (Seton Hall, UVA, ND, and ISU all top 50 teams). They are 2-5 vs top 50 (not bad). They are 2-2 100-200 (not great, but one of those L's will hopefully move up Memphis).

Iowa's chances:

12 wins: They are in the dance. No doubt they will have the wins and RPI won't matter.
11 wins: Bubble, will need at least 2 wins in DC
10 wins: Bubble outside looking in need at least 3 wins
9 or lower: Going to need to win it all in DC.

Nailed it

 

[El Simbolo]

After last night, not sure this thread matters so much anymore

 

[Mohawkeye]

After last night, not sure this thread matters so much anymore

I don't think last night has much effect at all on our chances. This is not a bad loss for the RPI. We need to win our home games and steal 2-3 on the road.

 

[Joes Place]

After last night, not sure this thread matters so much anymore

One loss isn't going to make that much impact. That was clearly one of the 'road games' that seemed at the outset to be 'most winnable', but who knows? Young teams come out and play inconsistently. That's what you live with.

Iowa needs to protect home court, and win at least 2 or 3 road games to finish in the upper third of the conference (or, win 3-4 road games if we lose one at home); that was true after the Purdue game, and hasn't changed today. Only, we have 1 less road game to pull an upset now.

 

[DavenportHawk8]

I don't feel too good about road games. 1 out of 3 games on the road we have been competitive. The other 2 we have been blown out of the water and showed no effort.

I think our best chances of a road win left are:

Rutgers: 75%
Illinois: 50%
Minny: 20%
Wisky: 10%
MD: 10%
MSU: 15%

So we do not have a lot of "winnable" games left. The way this year will go, I will bet we knock off at least 1 of these teams (Minny, Wisky, or MD). This team seems like they will get it together at some point. The problem is I could also see us losing @ Rutgers or even losing to PSU or OSU at home. So yeah its kind of an up/down battle.

 

[El Simbolo]

I realize that NU game doesn't affect our RPI too much, but I think it's a good example that we likely don't have the consistency to win enough games to make our RPI matter. Unless of course it's used to select NIT teams too