Sorry Boomer [Massey]

[Policarp]

For you Boomer....peeked at the LTURd bleedtechblue posts.......they are painfully aware who is ranked at #52 and how very very low CUSA overall is being ranked.

Their beloved Sagarin has LTURd at #99.....I could almost hear the tears falling.

[VObserver]

You're right, but in all fairness 4 games is not enough of a sample size for an algorithm to produce accurate results. So there are bound to be some wacky numbers this early in the season. A good model is useless without a good data set.

I dont disagree with this; but a construct like the Massey Composite will still likely be more accurate en globo than any one idividual or organization's opinion.

[CharlieK]

I dont disagree with this; but a construct like the Massey Composite will still likely be more accurate en globo than any one idividual or organization's opinion.

Since it all originates with subjective rankings and equations, that could never be verified by direct competition, it's either 100% or 0% accurate, depending on how you look at it and what you define accurate as. It's almost certainly not exactly how a tournament of all teams would turn out, so maybe it's 0% accurate. But then again it doesn't claim to be, it's only essentially one dude and his computer's opinion...it 100% accurately reflects that opinion.

Numbers I care about:

1. Record
2. Sunbelt Conference Placing
3. Bowl final score

[STPacronym]

Since it all originates with subjective rankings and equations, that could never be verified by direct competition, it's either 100% or 0% accurate, depending on how you look at it and what you define accurate as. It's almost certainly not exactly how a tournament of all teams would turn out, so maybe it's 0% accurate. But then again it doesn't claim to be, it's only essentially one dude and his computer's opinion...it 100% accurately reflects that opinion.

Numbers I care about:

1. Record
2. Sunbelt Conference Placing
3. Bowl final score

But, think of the number of recruits we can wow with the Massey rankings!

[CharlieK]

But, think of the number of recruits we can wow with the Massey rankings!

Only if they have the correct secondary sexual characteristics.

[CajunVic]

Only if they have the correct secondary sexual characteristics.

They all have these, some don’t master them as quickly as others, some don’t master them at all.

[VObserver]

Since it all originates with subjective rankings and equations, that could never be verified by direct competition, it's either 100% or 0% accurate, depending on how you look at it and what you define accurate as. It's almost certainly not exactly how a tournament of all teams would turn out, so maybe it's 0% accurate. But then again it doesn't claim to be, it's only essentially one dude and his computer's opinion...it 100% accurately reflects that opinion.

Numbers I care about:

1. Record
2. Sunbelt Conference Placing
3. Bowl final score

The Massey composite isnt one dude and his computer. It is 42 dudes or organizations of multiple dudes making independent evaluations. The composite obtains the mean and standard deviation of the ordinal rankings of each team according to all 42 data services, then ranks them from 1-130.

[CajunVic]

The Massey composite isnt one dude and his computer. It is 42 dudes or organizations of multiple dudes making independent evaluations. The composite obtains the mean and standard deviation of the ordinal rankings of each team according to all 42 data services, then ranks them from 1-130.

That’s some fancy words, but makes absolute sense.

[CharlieK]

The Massey composite isnt one dude and his computer. It is 42 dudes or organizations of multiple dudes making independent evaluations. The composite obtains the mean and standard deviation of the ordinal rankings of each team according to all 42 data services, then ranks them from 1-130.

Lighten up, Francis, it was a joke. I know it's more complicated than that.

But since we are talking about it, my point stands. It could be 1 million dudes and their computer and it would still be an amalgamation of OPINIONS. Not a massaging of data based on measuring a measurable thing. If I asked 1 million people their opinion on the best color, would the ranking somehow become a factual indication of "bestness"? Or would it be a ranking of favorite? There is no way to pin accuracy to a subjective ranking like this one, no matter how many subjective opinions you get and add up, divide, or square root.

[CajunVic]

Lighten up, Francis, it was a joke. I know it's more complicated than that.

But since we are talking about it, my point stands. It could be 1 million dudes and their computer and it would still be an amalgamation of OPINIONS. Not a massaging of data based on measuring a measurable thing. If I asked 1 million people their opinion on the best color, would the ranking somehow become a factual indication of "bestness"? Or would it be a ranking of favorite? There is no way to pin accuracy to a subjective ranking, no matter how many subjective opinions you get and add up, divide, or square root.

I don't think they are doing square root.