How the Selection Committee SHOULD Look at Resumes

[RoyalShock]

Yes. The committee could AND should adjust for home and road. A simple category called "adjusted top 50" would work well IMO.

Adjusted top 50 = top 30 home games, top 50 neutral games, top 70 road games.
Adjusted 51-100 = 31-81 home games, 51-100 neutral games, 71-121 road games.
and so on...

That would effectively mean a home win over 31 = neutral win over 51 = road win over 71. It would be super easy to calculate for every team and then the analysis I used in my original post could be followed using the adjusted records vs each category.

Call it "Tier 1 record", "Tier 2 record", and so on.

[1979Shocker]

Combine the NCAA and NIT tournaments into one large tournament. Get rid of the post-season conference tournaments to allow an additional 3 rounds. That would allow most teams into the main tourney, meaning the other minor tournaments would no longer be necessary. Let a computer figure out the seeds and locations.


I'm only joking about all of this. Anyone that has "The Onion" link, feel free to post it.

[Jamar Howard 4 President]

Do you know what the 30%, 54% and 59% would be for teams in the 20-40 RPI range and also the 1-20 range?

	RPI 1-20	RPI 21-40	RPI 41-60
vs top 25	55%	31%	30%
vs 26-50	69%	54%	54%
vs 51-100	80%	73%	59%

I developed these numbers purely from 2016. I bet if you combined the last 3 years you would get a more accurate distribution, but this should be close enough to get the idea.

[im4wsu]

	RPI 1-20	RPI 21-40	RPI 41-60
vs top 25	55%	31%	30%
vs 26-50	69%	54%	54%
vs 51-100	80%	73%	59%

I developed these numbers purely from 2016. I bet if you combined the last 3 years you would get a more accurate distribution, but this should be close enough to get the idea.

Thanks, JH4P. Not too much to differentiate 21-40 and 41-60 groups until the 51-100 level.

[Jamar Howard 4 President]

Thanks, JH4P. Not too much to differentiate 21-40 and 41-60 groups until the 51-100 level.

Yeah, I was surprised there wasn't a little more separation between those two groups. I would expect a little more separation if I used a larger sample size, such as 3 years rather than just 1, but who knows. 20 teams, averaging a handful of games in each category would seem to be a large enough sample size, but common sense tells me that there is no way the 41-60 group beats the top 25 equally to the 21-40 group. I just have to think there are some oddities in this year's numbers that would even out over several years.

[proshox]

Something about JH4P’s first post in reminded of the book “Paradox of Choice” that reviewed similar situations over custody battles. A fun, but impossible experiment with blind resumes would be interesting to me.

• The experiment would have the following elements/tricks

• Without telling the committee, each blind resume comparison will be reviewed two times throughout the process. There would be a randomized order in comparisons, but the goal would be for the committee to not realize they have already looked at this particular comparison.

• The key difference between the first and second review would be how the question is framed by the chair for each review:

• First time through - Which team should be included based on the blind resume comparison?

• Second time through – Which team should not be included based on the blind resume comparison?

I wouldn’t be surprised if the same Blind Resume is chosen for both questions. With either question the “average” team that is missing great wins or terrible losses will be ignored. The bi-polar team with a couple of signature wins and a couple of terrible losses will be chosen because you have something to point at to justify your decision.

I suspect this year’s committee spent a lot more time on “why a team should be included.”

[60Shock]

It should not happen, but just how much weight does the name brand of Syracuse versus little ol' St. Mary's carry into the equation?

[Cdizzle]

It should not happen, but just how much weight does the name brand of Syracuse versus little ol' St. Mary's carry into the equation?

I don't know the exact unit conversions, but roughly the difference between an NCAA tourney 10 seed and an NIT 2 seed. Whatever that comes out to in grams.

[ShockdaWorld]

I wouldn’t be surprised if this year’s committee spent a lot more time on “why a team should be included.”

Of course. Pick a name brand and find excuses to put them in the tourney. Syracuse is the perfect example.

[proshox]

Of course. Pick a name brand and find excuses to put them in the tourney. Syracuse is the perfect example.

Probably a fair statement, but in my controlled experiment team names would not be available.

[ShockdaWorld]

Probably a fair statement, but in my controlled experiment team names would not be available.

I agree with your theory/experiment. I was just responding to what happened this year that would be taken out of the equation if what you are saying were ever to come to fruition.