Brackets and Rankings: 2024-2025

[Xville]

It's an interesting concept. It essentially creates a mythical team that is squarely on the bubble, then measures how that mythical team would do against each team on your schedule, and then compares your actual performance to the mythical team's mythical performance. So Oklahoma has done better against their actual schedule then this mythical 45ish place team would have done, and Xavier has not done quite as well.

Any idea how exactly Torvik does his team rankings? Im curious because in that, Xavier is 45 and Oklahoma is 60. Pretty big discrepancy, and unlike a lot of the other computer rankings previously mentioned.

[xubrew]

Any idea how exactly Torvik does his team rankings? Im curious because in that, Xavier is 45 and Oklahoma is 60. Pretty big discrepancy, and unlike a lot of the other computer rankings previously mentioned.

I do not. I believe that it's a similar concept, but as far as how he goes about calculating it I really don't know. I'm not downplaying the importance of his metric. I've just never looked in to how he goes about calculating it.

[Xville]

I do not. I believe that it's a similar concept, but as far as how he goes about calculating it I really don't know. I'm not downplaying the importance of his metric. I've just never looked in to how he goes about calculating it.

Cool thank you. Yeah I was just curious ...it looks like quite a mix of WAB, Kenpom, and Net "stuff" in there but didn't understand how he sorted it.

[MHettel]

WAB = Wins Above Bubble for those of you who were planning to look it up (like I just had to).

And Oklahoma's WAB is 36 to Xavier's 54.

What the fuck does it mean? What are the inputs to it? Is it like WAR for baseball where everyone uses it to evaluate players and yet nobody can explain what it means?

I hope to god the committee understands these metrics. Otherwise, why have a committee at all. Just let the computer decide.

[MHettel]

It's an interesting concept. It essentially creates a mythical team that is squarely on the bubble, then measures how that mythical team would do against each team on your schedule, and then compares your actual performance to the mythical team's mythical performance. So Oklahoma has done better against their actual schedule then this mythical 45ish place team would have done, and Xavier has not done quite as well.

So instead of taking actual results of actual games, this approach compares simulated results of a fake team to your teams simulated performance against the fake team. Is that right?

And we use that?

[xubrew]

So instead of taking actual results of actual games, this approach compares simulated results of a fake team to your teams simulated performance against the fake team. Is that right?

And we use that?

Not really. My initial description was a little confusing.

It's the win probability of each opponent you've played against a bubble team. Each D1 team has a value (that's always changing). If you're playing a team like...say...Cleveland State at home, and it's determined that the win probability of a bubble team beating Cleveland State at home is 80%, you essentially get a +0.2 if you win the game, and a -0.8 if you do not. So, the actual results matter. Then they're all tallied and averaged for each team, and each team is then ranked based on that average. Some opponents will move you way up the list if you win. Others will barely move you at all.