According to Warren Nolan's ELO ...

[tatedog]

The Stanford regional is pretty ridiculous. Stanford is #7 and Fresno State is #9. How in the world does a 3 seed end up in the top 10 ELO?

 

[MSUDawg25]

What is ELO? I checked the site and I’m still not sure.

 

[T-TownDawgg]

Warren Nolan was never a member of Electric Light Orchestra

 

[wsjmsu75]

What is ELO? I checked the site and I’m still not sure.

I was wondering the same thing. Anyone care to enlighten us?

 

[tatedog]

What is ELO? I checked the site and I’m still not sure.

https://spreadsheetsolving.com/sports_league_power_ratings/

In a nutshell, it's a power rating system.

 

[UpTheMiddlex3Punt]

Probably because RPI has more influence over the pairings than Elo rating.

One major difference between RPI and Elo is that RPI is sequence-independent. Suppose you play team X during a season and beat them. If team X had a win over a top opponent, then lost to you, then lost to a really bad opponent, your Elo rating goes up much more than if team X had a loss to a bad opponent, then lost to you, then had a win over a top opponent. RPI does not care about this sequence of events. Elo does because points are adjusted after each matchup or at some regular interval, so perhaps the sequence may not matter over a three game series, but across weeks it does matter.

The Elo formulas are pretty simple and well defined with a few tweakable parameters, namely parameters that determine how much your score responds to recent events. A big problem is what to start everybody at. It is ridiculous to start all teams at the same rating since you clearly have some teams better than others, but you have no idea how good a team will be before the season starts. Beating a clearly overrated team early in the season will benefit you tremendously in the Elo but with RPI you will be rewarded the same as all the other teams that beat the overrated team. Nolan does not disclose many of the parameters for his model, which is fine. The Elo rating might be a better predictor of upcoming performance, but it will not earn you a better seed.

 

[wsjmsu75]

Let me put this more simply. What does ELO in this context stand for?

 

[DancingRabbit]

Don't bring me down

 

[tatedog]

Elo is actually a person's name, not an acronym. Sorry if my post misled you on that--it's kinda like Pascal, which many people carelessly turn into an acronym like I did here. The band ELO, however, is an acronym and I blame them for most of this confusion.

 

[T-TownDawgg]

Tate, your dry yet effective explanation of this snafu made me laugh

 

[tatedog]

Thanks. After 46 years I'm glad to have finally helped somebody with something.

 

[EminenceBunt]

ELO is my favorite ranking system in college baseball, much better than polls or RPI. Look at it this way:

RPI = a reflection of a teams results with respect to the toughness of their schedule played, aka, how good they have been

Polls= A reflection of a teams results, and also a teams public preconception mixed with the eye test. Aka, how good they were expected to be and how good they have been.

ELO = a reflection of a team’s pure ability in a vacuum. Aka, how good they are. Simply putC it’s a power ranking.

Not scientific by any means, but this is the way I see it. ELO has had us in the top 5 for months, well before the polls(results and conception based) and wayyy before RPI(results and schedule based).

If ELO has Fresno at 7, that tells me that they are, at the worst, playing like a top 12-15 team in the country right now. I think we got screwed up it’s this regional pairing, but like they say, you gotta beat the best to be the best.

 

[messageboardsuperhero]

Fresno was criminally underseeded. Stanford got screwed with their draw. If they come to Starkville next weekend, that will be a significant challenge.