not sure how the NET formula works exactly, but part of it is margin of victory which is a terrible way to select the NCAA tournament qualifiers

So I guess if we were 22-5 and all our wins were by 1 point in ot we would be ***ed

HoopDreams said:

---

not sure how the NET formula works exactly, but part of it is margin of victory which is a terrible way to select the NCAA tournament qualifiers

---

I wonder how much the NCAA spent on something so ridiculous, complicated and arbitrary when there were plenty of good models already out and in use for decades, put to the test by gamblers.

I think margin of victory is only to distinguish between a one or two point win vs. a double digit win. I think there is no difference between winning by 20 or 30 or 40.

Go Bears?

NET explained
https://www.ncaa.com/news/basketball-men/article/2022-12-05/college-basketballs-net-rankings-explained

Thanks for posting. I didn't know the NET formula was changed 4 years ago to remove margin of victory (good!)

"With the changes announced in May 2020, the NET will no longer use …scoring margin."

barsad said:

---

NET explained
https://www.ncaa.com/news/basketball-men/article/2022-12-05/college-basketballs-net-rankings-explained

---

Scoring margin is no longer a specific component of NET rankings, but it would be a lie to say that scoring margin is no longer relevant.

While winning percentage, adjusted winning percentage, and scoring margin were removed in 2020 from the formula for NET rankings, those components still play into the other two components that remain, the Team Value Index and the Adjusted Net Efficiency. It wasn't the case that the NCAA decided that those components were not relevant. Rather, those components were already somewhat built into the other components, so they were effectively used multiple times in the original NET ranking formula, and when winning percentage, adjusted winning percentage, and scoring margin were taken out, it made for less redundancy, and based on the "data" they decided it was a better overall metric to remove those redundancies.

Adjusted Net Efficiency is impacted by scoring margins. Lets ignore the "adjusted" part for the moment. Net efficiency is offensive efficiency minus defensive efficiency, with efficiency being points per possession. So net efficiency is essentially just scoring margin adjusted by pace. If you have net efficiency and pace numbers, you can calculate scoring margin without knowing the points scored or given up.

For example, Cal's current net efficiency is negative 0.021 (offensive efficiency of 1.027 and defensive efficiency of 1.048). Cal's possessions per game is 72.9. Multiply 72.9 and negative 0.021 and you get negative 1.512. Rounded to a single decimal, Cal has scored 74.8 points per game and given up 76.3 points per game, a scoring margin of negative 1.5, i.e., the same as we get using net efficiency and pace rather than points scored versus points given up.

Bottom line, scoring margin is still in the formula, but adjusted for pace, plus adjusted for strength of opposition and location (home/neutral/away) of the game.

Similarly, Team Value Index is based on game results, factoring in opponent, location, and the winner. I.e., it is a blend of adjusted winning percentage and strength of schedule that is just more sophisticated than the RPI was.

So, in the end, NET Ranking is basically a more sophisticated version of RPI combined with a sophisticated way to factor in scoring margin, and the changes in 2020 eliminated redundancy to make a "better metric."

Cal8285 said:

---

Scoring margin is no longer a specific component of NET rankings, but it would be a lie to say that scoring margin is no longer relevant.

While winning percentage, adjusted winning percentage, and scoring margin were removed in 2020 from the formula for NET rankings, those components still play into the other two components that remain, the Team Value Index and the Adjusted Net Efficiency. It wasn't the case that the NCAA decided that those components were not relevant. Rather, those components were already somewhat built into the other components, so they were effectively used multiple times in the original NET ranking formula, and when winning percentage, adjusted winning percentage, and scoring margin were taken out, it made for less redundancy, and based on the "data" they decided it was a better overall metric to remove those redundancies.

Adjusted Net Efficiency is impacted by scoring margins. Lets ignore the "adjusted" part for the moment. Net efficiency is offensive efficiency minus defensive efficiency, with efficiency being points per possession. So net efficiency is essentially just scoring margin adjusted by pace. If you have net efficiency and pace numbers, you can calculate scoring margin without knowing the points scored or given up.

For example, Cal's current net efficiency is negative 0.021 (offensive efficiency of 1.027 and defensive efficiency of 1.048). Cal's possessions per game is 72.9. Multiply 72.9 and negative 0.021 and you get negative 1.512. Rounded to a single decimal, Cal has scored 74.8 points per game and given up 76.3 points per game, a scoring margin of negative 1.5, i.e., the same as we get using net efficiency and pace rather than points scored versus points given up.

Bottom line, scoring margin is still in the formula, but adjusted for pace, plus adjusted for strength of opposition and location (home/neutral/away) of the game.

Similarly, Team Value Index is based on game results, factoring in opponent, location, and the winner. I.e., it is a blend of adjusted winning percentage and strength of schedule that is just more sophisticated than the RPI was.

So, in the end, NET Ranking is basically a more sophisticated version of RPI combined with a sophisticated way to factor in scoring margin, and the changes in 2020 eliminated redundancy to make a "better metric."

---

Thanks for the explanation. I can't say I completely understand the math, but at least it sounds like scoring margin isn't as large a factor as in the past.

The efficiency factor seems to favor potential (or best team) rather than results, which seems reasonable as long as the other two factors favor results.

I didn't like the debate in FB where results didn't seem to factor as heavily as potential ... the long debate over most deserving team (earned on the field) vs best team (determined subjectively, and highly dependent on injuries).

I'm in the results on the field/court matter the most camp

HoopDreams said:

---

Cal8285 said:

---

Scoring margin is no longer a specific component of NET rankings, but it would be a lie to say that scoring margin is no longer relevant.

While winning percentage, adjusted winning percentage, and scoring margin were removed in 2020 from the formula for NET rankings, those components still play into the other two components that remain, the Team Value Index and the Adjusted Net Efficiency. It wasn't the case that the NCAA decided that those components were not relevant. Rather, those components were already somewhat built into the other components, so they were effectively used multiple times in the original NET ranking formula, and when winning percentage, adjusted winning percentage, and scoring margin were taken out, it made for less redundancy, and based on the "data" they decided it was a better overall metric to remove those redundancies.

Adjusted Net Efficiency is impacted by scoring margins. Lets ignore the "adjusted" part for the moment. Net efficiency is offensive efficiency minus defensive efficiency, with efficiency being points per possession. So net efficiency is essentially just scoring margin adjusted by pace. If you have net efficiency and pace numbers, you can calculate scoring margin without knowing the points scored or given up.

For example, Cal's current net efficiency is negative 0.021 (offensive efficiency of 1.027 and defensive efficiency of 1.048). Cal's possessions per game is 72.9. Multiply 72.9 and negative 0.021 and you get negative 1.512. Rounded to a single decimal, Cal has scored 74.8 points per game and given up 76.3 points per game, a scoring margin of negative 1.5, i.e., the same as we get using net efficiency and pace rather than points scored versus points given up.

Bottom line, scoring margin is still in the formula, but adjusted for pace, plus adjusted for strength of opposition and location (home/neutral/away) of the game.

Similarly, Team Value Index is based on game results, factoring in opponent, location, and the winner. I.e., it is a blend of adjusted winning percentage and strength of schedule that is just more sophisticated than the RPI was.

So, in the end, NET Ranking is basically a more sophisticated version of RPI combined with a sophisticated way to factor in scoring margin, and the changes in 2020 eliminated redundancy to make a "better metric."

---

Thanks for the explanation. I can't say I completely understand the math, but at least it sounds like scoring margin isn't as large a factor as in the past.

The efficiency factor seems to favor potential (or best team) rather than results, which seems reasonable as long as the other two factors favor results.

I didn't like the debate in FB where results didn't seem to factor as heavily as potential ... the long debate over most deserving team (earned on the field) vs best team (determined subjectively, and highly dependent on injuries).

I'm in the results on the field/court matter the most camp

---

Of course, the debate about most deserving versus best can still be had in the committee. The NET ranking is only a tool, only one factor among many the committee views. And aside from simple position in the NET ranking, NET ranking is also used to determine what games are Quads 1 through 4, and Quad 1 wins and Quad 3 or 4 losses are an important tool, too.

The basketball selection committee can sometimes seem to factor potential over results.

The good news is that it isn't nearly the big deal that the FB debate was this fall. The FB committee was picking 4 teams, with plenty of conference champs being left out. On the other hand, if we're taking about who gets to play in the tourney, the committee debate is over which teams that don't really deserve to play for the championship get the last spots in the tourney.

With 68 teams, there are 32 conference tourney winners and 36 at large teams. Most of the at large bids are pretty obvious based on results, but really, probably no more than 25-30 at large teams really deserve to be playing for a championship based on results. Sure, it is painful to teams that barely get left out, the "first four out" can always argue they deserved to get in more than the "last four in," and the "last four in" can always argue that the should have been one of the "last four byes," but in truth, not of those really deserved to be playing for an NCAA championship, so if the committee makes a bad decision, it isn't really a huge deal.

The committee is also using its "tools" to decide seeding, which may matter more for the championship than who gets in at the back end, but if seeds are off by a line or two compared to what was "deserved," well, a team needs to beat the best to win a championship, so it really isn't that big deal.

The good news for FB is that, with the expansion to 12, the committee decisions can't screw a deserving team as much as it did this year. The top four conference winners get the byes, and that will mostly be the current 4 power conferences, the committee mostly won't have big decisions to make, and the teams vying for 11th and 12th spots don't really deserve to be playing for a championship, so it won't be nearly such a big deal if the committee screws up (again, not that it won't hurt for the first couple of teams left out who might be more deserving).

I can't wait until football moves closer to basketball and does a 16, 20 or even 32-team format. I would be for taking all byes out of a playoff format, too, every team should have to earn it in the same number of games.
Football lacks any Cinderella stories like March Madness because of the small number of teams (more next year, though).
Controversies over bubble teams will always be the case no matter how many you invite to the Dance, that's part of the fun (and agony, if Cal one day gets left on the bubble).
The statistical tools and rankings are overkill at this point, I am OK with some X-factors letting some teams Dance. Like Loyola getting in because a 100-year-old nun roots for them. Or a certain blue-and-gold team getting in on Selection Sunday 2025 because their coach brought them from the dungeon (3 wins) to the light of day two years later (17 wins).