Ken Pomeroy - The RPI Is Dead

[hipsterdoofus]

http://kenpom.com/blog/index.php/weblog/the_rpi_is_dead/

If Ken Pomeroy is not updating the Ratings Percentage Index, then maybe it is starting to lose its relevance.

[X Factor]

I prefer realtimerpi.com. It updates every 5 minutes so there can be a lot of movement throughout the day.

Currently, X is 105th. :)


RealTimeRPI (http://realtimerpi.com/rpi_Men.html)

[Masterofreality]

Hell, now we're down to 117th!!

Damn that strength of schedule. The sky is falling, the sky is falling!!

[chico]

I hope Billy found this out from a loved one instead of having to hear about it first on a message board.

[Swifty]

Dayton is 100 and we are 136th. I guess they really have passed us as a program.

[Snipe]

MHet is going to be pissed.

[blobfan]

Dayton is 100 and we are 136th. I guess they really have passed us as a program.

Someone's going to have to explain that one to me. Wofford is at 224 and IPFW is at 206. How does that lead to Dayton outranking us? I guess Wofford has a stronger SOS.

No wonder people constantly discount the RPI.

[Billy]

MHet is going to be pissed.

Maybe this simply provides MHettel with an opportunity to fill the gap this creates in the market space.

Let's be honest, the RPI is becoming marginalized with each passing season. I've been preaching this for the last three years on this message board...and I think this is just more evidence of it. The message board battle I have fought on this subject has been less about math (despite the way some have tried to spin it)...and more about it's application and relevance.

The college basketball fan is becoming more sophisticated, and this "tool" is too antiquated to fulfill what it's trying to accomplish.

Good for Pomeroy for seeing the light. He has a pretty good rating system himself that he should use his focus his efforts on fine tuning.

[MHettel]

I already knew about this from earlier in the week.

In fact, I don't think there is any question that the RPI is less usefull now than it's been in the past, but I think thats mostly due to the fact that teams and conferences have really figured out how to manipulate it so well.

I do question why Ken Pom did this. He says he wasnts to focus more on his own metrics that he develops, but frankly, they confuse me. I don't know what they intend to show. If he doesn't have the RPI there then I probably wouldn't go to his site to even have a chance to see his other stats.

I actually use the games file he posts to upload all the results into my RPI engine.

I'll keep it going this year, and do my usual updates, although they probably will mean less now than ever before.

I also added a cool feature where it seeds an entire tournament and bracket. Then i use seed vs. seed history to simulate the entire tourney. I'm now backtesting some of my criteria to see if I can get close to seedings that essentially match last years real seedings. If so, I'll use those criteria throughout the season.

[waggy]

There are so many places on the net to get RPI information already, that it doesn't make sense for Pomeroy to bother with it. Time and money. That's my take.

[Swifty]

Wow how stupid. Toledo has two losses but is 31 in the RPI. We have two wins and are 58 in the rpi.

[PMI]

W-L is only about 25% of it I believe. I think about 75% is opponents' records and opponents' opponents' records. But it's not time for this.