Cornell Lax #1 RPI

According to Lax Power,

[www.laxpower.com]

Cornell is #1 in the nation in the RPI after 3 games. Although, the site issues a word of caution that the RPI does not mean much until all teams have played at least 5 games. Still, Cornell is definitely a Top 5 RPI team and should solidify itself as #1 in the country if it pulls off wins against Army and Duke. Although that trip to Durham could be rough (and I am not looking past Army either).

Meanwhile in other lax news, mighty Drexel which is ranked #20 nationally only because of its season opening 1 goal win over UVa, squeeked by Binghamton yesterday in an offensive explosion game 4-2.

Too bad Penn did not play Drexel this year. Seems like the Quakers are willing to play only the crappy local area schools.

At least as of 11pm tonight, Cornell is #5 in RPI at that Laxpower site, just two ahead of St. Joseph's and their impressive 0-4 record.

Al DeFlorio
At least as of 11pm tonight, Cornell is #5 in RPI at that Laxpower site, just two ahead of St. Joseph's and their impressive 0-4 record.

Well, that's RPI for you. St. Joe's four opponents (Delaware, Navy, Drexel, Duke) have a combined record of 15-1. That'll drop eventually - their remaining opponents have combined record of 7-23.

We all know how worthless RPI is this early into a season anyway. After all, at this point Rutgers and Mount St. Mary's are tied for #1; both have only played (and lost to) 5-0 Delaware.

RPI: rewarding losers since the 1970s.

Josh '99
We all know how worthless RPI is this early into a season anyway.

I think it's close to worthless throughout the season. Doesn't take into account whether your wins have come against the best teams on your schedule or the worst, and the primary weighting is given to whom you've played, not how you've done. Whalen's pegged it succinctly.

Al DeFlorio
Whalen's pegged it succinctly.

Who is this Whalen of whom you speak?

ETA: just noticed my signature doesn't actually say "Whelan" anywhere in it.

jtwcornell91
RPI: rewarding losers since the 1970s.

Good thing that I graduated in the the 1960s.

jtwcornell91

Al DeFlorio
Whalen's pegged it succinctly.

Who is this Whalen of whom you speak?

ETA: just noticed my signature doesn't actually say "Whelan" anywhere in it.

Sorry, John. I know which is right and consistently type it wrong.

Al DeFlorio

Josh '99
We all know how worthless RPI is this early into a season anyway.

I think it's close to worthless throughout the season. Doesn't take into account whether your wins have come against the best teams on your schedule or the worst, and the primary weighting is given to whom you've played, not how you've done. Whalen's pegged it succinctly.

Well, right. Worthless was maybe the wrong word. "Statistically insignificant?"

jtwcornell91
Who is this Whalen of whom you speak?

He's a friend of Beej's.

Beeeej

jtwcornell91
Who is this Whalen of whom you speak?

He's a friend of Beej's.

I know those guys. I think they both went to Darmouth

Al DeFlorio

Josh '99
We all know how worthless RPI is this early into a season anyway.

I think it's close to worthless throughout the season. Doesn't take into account whether your wins have come against the best teams on your schedule or the worst, and the primary weighting is given to whom you've played, not how you've done. Whalen's pegged it succinctly.

While literally true (50% of RPI is opponent's win %, typically, although I can't find the official lax formula), practically I think its a mischaracterization.

The reason is that, from team to team, Strength of Schedule will only vary so much. You'll play some good teams and some bad ones, a lot of conference games, and overall everyone will have a SoS that varies by a modest amount. For comparison purposes, the SoS column on USCHO has the hockey SoS varying from ~.53 to ~.45. So the max from the best to the worst if 0.10 or less. So take this times ~2/3rds and the max RPI difference between teams based on SoS is probably 0.7 or less.

On the other hand, win % can vary from 0 to 1 - or, more practically perhaps, lets say 0.15 to .75 - or a range of .60 or so. While only multiplied times only ~1/3rd, this is still .20, or about 3 times the impact from SoS differences, despite the lower weighting.

Now, where the correct balance lies is impossible to say, very subjective, and subject to a ton of personal bias. But despite the literal 'weighting', I don't feel its fair to say that the primary factor is who you've played. Winning matters most.

If I recall what I saw on Laxpower correctly, the RPI used for lacrosse weights a team's own winning percentage much more highly than other sports (I think 50%).

If a team's record is worth 50% of the RPI, it's hard to imagine how a team with an 0-4 record--like St. Joseph's--could have the seventh best RPI.

The problem with the "team's record" piece is that it doesn't consider whom you've beaten in achieving that record. So if you've played Duke, Virginia, and Hopkins and have an 8-3 record, RPI doesn't care whether the three losses were to those schools or whether you've beaten all three.

Al DeFlorio
If a team's record is worth 50% of the RPI, it's hard to imagine how a team with an 0-4 record--like St. Joseph's--could have the seventh best RPI.

Yeah - you're right about that. I was probably confused.

Josh '99

Al DeFlorio
If a team's record is worth 50% of the RPI, it's hard to imagine how a team with an 0-4 record--like St. Joseph's--could have the seventh best RPI.

Yeah - you're right about that. I was probably confused.

You weren't confused, the lax power page is stupid. It uses numbers talking about how they don't know the real numbers. Um, excuse me, then how did you calculate the rankings. I have no doubt the real numbers are well known (and if not officially, easy to reverse engineer over a bit of time). Rather the people who made the explanation page were too lazy to bother asking the script writers what they were. So they made a guess, which is a silly guess at that.

Al DeFlorio
The problem with the "team's record" piece is that it doesn't consider whom you've beaten in achieving that record. So if you've played Duke, Virginia, and Hopkins and have an 8-3 record, RPI doesn't care whether the three losses were to those schools or whether you've beaten all three.

That's fair, RPI intentionally ignores who you beat and who you lose to with your own schedule. So does KRACH, etc.

However, this is a difficult question. Lets say you beat Minnesota twice and lose to AIC twice, and that's your total record. And another team loses to Minnesota twice and beats AIC twice.

Who's better? Is it better to be able to beat the top team, even if you completely blow it when playing a cupcake (mmmmm, cupcake)? Or is it better to beat who you're supposed to even though you couldn't beat a top team. I'm not sure there is a right answer to that question. Or, if there is, its would require numerous other factors such as goal differential, perhaps injuries, tiredness, and who knows what else. It would become incredibly complex.

It reminds me of our tiebreak with Q this year. Since we won it on record against top 8, and we had the same overall record, really, we were rewarded for having a worse record against the bottom 4. Does that make any sense? Not really. Would is likewise make any sense to lose for doing better against the bottom portion but not the top portion? Not really.

At best all it seems to get you is a confidence interval - how variable the team has been throughout the year. But I'm not sure it says much about the overall ranking.

DeltaOne81However, this is a difficult question. Lets say you beat Minnesota twice and lose to AIC twice, and that's your total record. And another team loses to Minnesota twice and beats AIC twice.
[/quote
I understand your point, Fred, but that's an extreme example.

Let's look at this year's Cornell lacrosse schedule. If Cornell should beat Duke, Syracuse, and Princeton, but lose at Army and, say, at Yale, I'd be more impressed with the team's strength than if they were to beat Army and Yale but lose to Duke and either of the other two. Same bottom-line won-lost record (and, therefore, RPI), but, to me at least, an indication of a stronger team.

To come back to your example, a team that can beat Minnesota twice has demonstrated it's capable of beating anyone. The second worst team in the country could beat AIC twice while losing to Minnesota. Any high school team can lose to Minnesota--and a few, perhaps, might even be able to beat AIC.

 

___________________________
Al DeFlorio '65

DeltaOne81

Josh '99

Al DeFlorio
If a team's record is worth 50% of the RPI, it's hard to imagine how a team with an 0-4 record--like St. Joseph's--could have the seventh best RPI.

Yeah - you're right about that. I was probably confused.

You weren't confused, the lax power page is stupid. It uses numbers talking about how they don't know the real numbers. Um, excuse me, then how did you calculate the rankings. I have no doubt the real numbers are well known (and if not officially, easy to reverse engineer over a bit of time). Rather the people who made the explanation page were too lazy to bother asking the script writers what they were. So they made a guess, which is a silly guess at that.

So then I WAS confused, it was just Laxpower's fault and not mine.