2019 Bracketology Thread

[jhu72]

The formula is very simple. The 25% vs 50% vs 25% is well understood. It's the inputs for each team that Admin was referencing I think. For some reason, various outlets tend to have different results. I have no earthly idea why. I have been trying to figure it out since I started putting up my RPI stuff.

To be claer, the ordering of the teams in the final RPI tends to be the same, so it's not likely we are all putting out different lists, but you'd think the specific underlying inputs would be the same as well, which doesn't seem to be the case.

And no, Hampton does not count.

I know you're way deeper into the weeds on this stuff than I'll ever be, so maybe what I'm going to say is already obvious. But the one thing that immediately comes to mind that'd produce small differences like that is that IIRC you're supposed to drop an opponent's game(s) against the team you're rating from the calculation. In other words, if you're computing Hopkins' RPI you enter Towson as 10-4 not 11-4, UNC as 8-6 not 8-7, Maryland as 11-2 not 11-4, and so on. (And the same for opponents' opponents, so e.g. Georgetown as a Towson opponent goes in as 13-3 not 13-4.)

If some sites are doing that correctly* and others aren't, that'd account for minor discrepancies in input of the kind you're describing.


* It's also possible I have it backward and you AREN'T supposed to do that but some people do. But I'm pretty sure the way I have it here is the right one.

Years ago that was a very common error.

[Homer]

Laxref, i question who is using the 25-50-25 and nothing else.

According to the article posted on the front page of your website, the NCAA selection committee.

[laxreference]

You are right Homer, as far as I know anyway. I do remove those games from my calc. The biggest thing that I haven't gotten answered is whether the Hampton D1 games count or not. I have currently just excluded any games against Hampton from my calculations.

But at the end of the day, all my knowledge of their calculation is second-hand, as opposed to being directly from the NCAA. So it's possible we all have it wrong.

[ABV 8.3%]

The formula is very simple. The 25% vs 50% vs 25% is well understood. It's the inputs for each team that Admin was referencing I think. For some reason, various outlets tend to have different results. I have no earthly idea why. I have been trying to figure it out since I started putting up my RPI stuff.

To be claer, the ordering of the teams in the final RPI tends to be the same, so it's not likely we are all putting out different lists, but you'd think the specific underlying inputs would be the same as well, which doesn't seem to be the case.

And no, Hampton does not count.

Are you sure about Hampton not counting? According to the NCAA RPI ranking, it shows VMI and Furmans W-L record reflecting the Hampton win. (4 wins for Furman as an example ) Are you saying that Furman and VMI (NJIT too) RPI formula is skewed to only using the first .25% W-L, not counting anything else from playing Hampton? Is this the case?

[Big Dog]

Get credit for simply showing up at the game.

As the saying goes, 80% of life is just showing up.

[laxreference]

The formula is very simple. The 25% vs 50% vs 25% is well understood. It's the inputs for each team that Admin was referencing I think. For some reason, various outlets tend to have different results. I have no earthly idea why. I have been trying to figure it out since I started putting up my RPI stuff.

To be claer, the ordering of the teams in the final RPI tends to be the same, so it's not likely we are all putting out different lists, but you'd think the specific underlying inputs would be the same as well, which doesn't seem to be the case.

And no, Hampton does not count.

Are you sure about Hampton not counting? According to the NCAA RPI ranking, it shows VMI and Furmans W-L record reflecting the Hampton win. (4 wins for Furman as an example ) Are you saying that Furman and VMI (NJIT too) RPI formula is skewed to only using the first .25% W-L, not counting anything else from playing Hampton? Is this the case?

That's the issue. I'm not sure. I had heard that they were totally excluded, which is what I implemented in my script. But I have not heard definitely from the NCAA at all whether that is true.

Subjectively, it would make sense to make to me to remove them completely because I see no reason to penalize teams that play Hampton. If we could see the underlying RPI scores and inputs, we'd know for sure, but their list only showed the rank, not the actual RPI value.

[runrussellrun]

WIlling to take MAXIMUM RISK that the n$aa is open and transparent on this issue

[10stone5]

Get credit for simply showing up at the game.

As the saying goes, 80% of life is just showing up.

—— Woody Allen

[admin]

ABV, concurring with you, whenever a ranking speaks in generalities, I flinch. For sure with conferences, as you're saying, but also total goals scored vs. goals against, etc. and this includes the total wins of your opponent and the total losses. i.e. RPI. Heck, the fact that RPI uses the team's winning percentage bothers me. I want to be able to understand why A, B, or C is happening and when the formula speaks in such generalities, not only is this method (I believe) prone to errors, when there are errors, it's difficult to point directly at what needs to be fixed.

[admin]

Pretty good piece. https://www.insidelacrosse.com/article/ ... eria/54516

[Big Dog]

Pretty good piece. https://www.insidelacrosse.com/article/ ... eria/54516

Disagree.

The author points out a his perceived problem but doesn't have the cajones to recommend something else. Or, at least recognize that 'something else' would have its own problems (as many have discussed on this blog). Or to recognize that similar to men's football and basketball where there are 5 so-called Power conferences, and a bunch of mid-majors, in LAX there are 2 power conferences, and a bunch of mid-majors. The mid-majors get fewer bids, by definition, and oftentimes, just the AQ.

If the author thinks the RPI should go, he should own up to it.

[admin]

His examples of why RPI and SOS are unfair were pretty good. No?

[Big Dog]

His examples of why RPI and SOS are unfair were pretty good. No?

Perhaps, but that's not exactly new news. The issues with the RPI have been known in b'ball for 30+ years. (B'ball coaches learned long ago to schedule hard out of conference games and a few cupcakes for home.) As a result, others have entered the field (Sagarin, Pomeroy, Massey, et al)

One issue that the author ignores is the small number of games in a lax season; it means that any statistic used will have a higher variance, relative to b'ball and hockey. For example, NCAA hockey can readily use hard statistics (pairwise) as they play ~40 games in the regular season.

[laxreference]

Pretty good piece. https://www.insidelacrosse.com/article/ ... eria/54516

Disagree.

The author points out a his perceived problem but doesn't have the cajones to recommend something else. Or, at least recognize that 'something else' would have its own problems (as many have discussed on this blog). Or to recognize that similar to men's football and basketball where there are 5 so-called Power conferences, and a bunch of mid-majors, in LAX there are 2 power conferences, and a bunch of mid-majors. The mid-majors get fewer bids, by definition, and oftentimes, just the AQ.

If the author thinks the RPI should go, he should own up to it.

I don't know. It wasn't highlighted especially well, but it seemed like the main issue he pointed out was that High Point took a huge penalty for having the Bonnies and Mercer on their schedule. I thought it was helpful to know that scheduling those teams cost them as much as their two regular season losses. I do think that this is the crux of the issue that people have with RPI.

But rather than focus on picking (or recommending) a new ranking system, I think the conversation should be about what behavior we want to incentivize. So yes, I agree with in you in that I wish he would have taken the second half of his post in this direction. If we want to prioritize scheduling difficult teams, we can design an algorithm to reflect that. If we want to prioritize wins on the field, we can design an algorithm to reflect that. If you want to balance those two factors evenly, we can design an algorithm to reflect that.

If the NCAA/teams agree that the current model incentivizes the wrong things AND they can agree on what they do want to incentivize, that is 80% of the work. Then it just becomes a question of whether there should be any human subjectivity at all.

At the end of the day, I think there is value in pointing out what you don't like, assuming that you follow it up with maybe not a proposed ranking system, persay, but at least some idea of what behaviors you do want to incentivize.

[ICGrad]

Pretty good piece. https://www.insidelacrosse.com/article/ ... eria/54516

Disagree.

The author points out a his perceived problem but doesn't have the cajones to recommend something else.

If the author thinks the RPI should go, he should own up to it.

I think he clearly owns up to his assertion that it should go.

Look, gathering up a list of the top 12-15 teams in consideration for an at-large, putting the names on a wall, and having a blindfolded monkey throw darts at the names would be a fairer system than what we have now, a system which keeps rewarding the same teams over and over because they lose a bunch of games to good teams.

[Homer]

But rather than focus on picking (or recommending) a new ranking system, I think the conversation should be about what behavior we want to incentivize.

This is a really good point. Whatever the committee chooses to prioritize doesn't just affect selection itself, it also in turn sets the incentives for programs and conferences in how they approach scheduling. You get more of whatever you subsidize and less of whatever you tax.

The problem I think some people don't really want to confront is that the incentives that make for the most intuitively fair selection process don't necessarily align with the incentives that create the most entertaining, viewer-friendly regular season.

It's no accident that so much of this debate surrounds Hopkins. What the current system allows Petro to do is keep as many of the traditional marquee opponents on the schedule as he wants, while also playing PSU, Rutgers, etc., with no real downside in terms of tournament selection. Stay north of .500 and you'll probably be okay. Under a different system you might have to think twice about the wisdom of playing a schedule like that, but as it stands now you can keep alums and TV happy and be putting your team in a good position for the postseason. No trade-off!

Start tinkering with that and you'll start seeing some big-name programs changing the way they schedule in a way a lot of people won't like. Maybe that's not a bad thing; I know runrussellrun would be all for it, and I'm not saying he's wrong. But anybody with a take on this should definitely be grasping the nettle of how these things are all interconnected, and presenting their position accordingly.

[admin]

I can't remember who shared the rankings that included "expected wins/losses" but, whoever that was, thought they'd find FIFA's rankings interesting.

PS This methodology is better than RPI.

[cuseman4133]

Pretty good piece. https://www.insidelacrosse.com/article/ ... eria/54516

Disagree.

The author points out a his perceived problem but doesn't have the cajones to recommend something else. Or, at least recognize that 'something else' would have its own problems (as many have discussed on this blog). Or to recognize that similar to men's football and basketball where there are 5 so-called Power conferences, and a bunch of mid-majors, in LAX there are 2 power conferences, and a bunch of mid-majors. The mid-majors get fewer bids, by definition, and oftentimes, just the AQ.

If the author thinks the RPI should go, he should own up to it.

Let's not forget Foy and company still thought High Point was an at-large team in the NCAA Tournament even after their loss to Jacksonville when they weren't. It was their first Bracketology. Nobody at US Lax Mag or College Crosse had High Point as an at-large in any of their Bracketology posts.

And he notes the obvious: "If you don't lose to two bad teams, your resume is better." No duh. And "if you didn't schedule a winless first-year program and someone in your actual conference, your RPI would be better." (Same with VMI). No. crap.

Terry doesn't have an impactful voice in terms of changing the sport, especially at the NCAA level. I don't know why people think it's such a great post when he states the obvious that everyone knows and doesn't bring up an alternative. Lacrosse isn't that big to get its own ratings formula. Face it. A ton of lax people are so insular that they don't look outside their own bubble.

[Drcthru]

+1

[CU77]

Lacrosse isn't that big to get its own ratings formula.

Hockey is smaller than lacrosse and gets its own ratings formula. And the committee doesn't second-guess it. And it has been that way for decades. And it has not noticeably impacted scheduling.