Although the execution did vary depending on the source, the RPI formula itself was pretty straight forward. If had the skillz to write an Excel macro, the data, and a computer that wasn't busy mining bitcoin, anyone could have calculated RPI within a few ten-thousandths of the NCAA's "official" calculation.
Has anyone seen the formula for the NET? There's no shortage of articles on the interwebs from August (when the NET was announced) about what data goes into it, but there's been nada since. What concerns me most is that Gavitt went on record back then saying that there was not yet a decision to release the formula (and even intimated that the formula itself had not been defined). This would be unprecedented, no? Even the BCS computer algorithms were public (even if only the upper sextile of the population - and zero SEC fans - could understand them), were they not?
As we all know, the NCAA is nearly as transparent as a sheet of wax paper that's been regaled with a toddler's crayon, but are the schools, coaches, fans, pundits, etc. actually going to stand for some esoteric formula to determine the participants in the country's most prestigious (and profitable) tournament?
Formula is not available.
I was under the impression there was a "secret sauce" that may not be algorithmic (i.e. subjective)...meaning no excel-spreadsheet-on-steroids could do it. BUT I can't seem to find where I read that. In any case, something about it is meant to be kept secret.
My guess is that they don't want teams to schedule to the formula. But someone's going to reverse engineer it anyway.
Sorry I can't link to the tweet, but John Gasaway tweeted out yesterday that the NCAA sent out an "explanation" of the NET to D-I teams yesterday.
The NET will weight in the following order:
1) Team Value Index -- some nebulous algorithm of game results based on opponent, location, and who won
2) Net Efficiency -- Offensive E - Defensive E
3) Winning Percentage -- straight up
4) Adj Winning Percentage -- weighted based on location and result
5) Scoring Margin -- max differential of 10pts , with OT games always 1
So, not much new info and maybe more questions than answers.
What's under the hood of the TVI?
How much will each component be weighted?
Why have winning percentage used twice?
Was the acronym solely intended for basketball-related puns?
Quote from: barfolomew on November 05, 2018, 01:05:09 PM
Sorry I can't link to the tweet, but John Gasaway tweeted out yesterday that the NCAA sent out an "explanation" of the NET to D-I teams yesterday.
The NET will weight in the following order:
1) Team Value Index -- some nebulous algorithm of game results based on opponent, location, and who won
2) Net Efficiency -- Offensive E - Defensive E
3) Winning Percentage -- straight up
4) Adj Winning Percentage -- weighted based on location and result
5) Scoring Margin -- max differential of 10pts , with OT games always 1
So, not much new info and maybe more questions than answers.
What's under the hood of the TVI?
How much will each component be weighted?
Why have winning percentage used twice?
Was the acronym solely intended for basketball-related puns?
You forgot the "line of code" in the equation where Syracuse gets an "auto-bid"......
Quote from: burger on November 05, 2018, 01:47:22 PM
You forgot the "line of code" in the equation where Syracuse gets an "auto-bid"......
They had to cut that line out to make room for the line that makes UW a pre-season top 5 in at least three different categories.
Quote from: barfolomew on November 05, 2018, 01:05:09 PMWhy have winning percentage used twice?
Looks to me like it's factored in several times:Quote from: barfolomew on November 05, 2018, 01:05:09 PM
1) Team Value Index -- some nebulous algorithm of game results based on opponent, location, and who won
2) Net Efficiency -- Offensive E - Defensive E
3) Winning Percentage -- straight up
4) Adj Winning Percentage -- weighted based on location and result
5) Scoring Margin -- max differential of 10pts , with OT games always 1