I hope the Gophers are highly efficient tonight against Illinois!

[Some guy]

Let me dig up that box score...

In that game, calculated offensive possessions for our offense was 87.4. 68 points. That's .778 per possession or an offensive eff of 77.84
Defensively 92.6 possessions. 95 points allowed. 1.02 pts per possession. Def Eff of 102.59.

That's a spread of -24.753. (that's Chicago St level performance).
If we lose that 78 to 88 our spread would have been -5.74 instead.

17 points essentially has us at #54 right now instead of #50.

In a game where the team quit.
Not forgiving the team for quiting. But the last 10 minutes of that game the team stopped playing hard and it alone is dragging the team down 5 spots.

As it should, because those are clearly some of the most important possessions of the season.

 

[builtbadgers]

I just shake my head.

The implications are sillier than we have even discussed. There is a real possibility that sending IW to the hoop with 25 left in the shot clock up 8 points is better for your team than working the clock to win the game . It isn't just a different way of measuring things, it might actually be counter intuitive to the point of competitive sports since its inception.

Pace of play has nothing to do with it. Sending IW to the hoop is a good play if it results in a good shot and a conversion more often then a turnover. Of course it is all situational. If there were 40 seconds left in the game and 25 left on the shot clock then taking time of the clock is more important and can still lead to points on the possession.

 

[WoodburyTim]

How do they compare to the last Illinois game where we puked all over our self. they will add up to the point differential in pig slaughter loss and comfortable victory. All the conference champions and NCAA champions are super in OE-DE total every single year without exception.

This is not solid statistical analysis. Looking at something historically and then determining a statistic to judge past success is not the same as then saying that statistic is now how success should be judged in the future.

For example, if I go back and study the last 50 years of successful students and determine that they scored well on a "builtbadger statistic" I calculated after the fact, it would be informative, but it wouldn't necessarily be helpful once I told students that their "builtbadger statistic" would be how they would be judged moving forward. It changes behavior. It is a pretty accepted statistical premise.

 

[Some guy]

Pace of play has nothing to do with it. Sending IW to the hoop is a good play if it results in a good shot and a conversion more often then a turnover. Of course it is all situational. If there were 40 seconds left in the game and 25 left on the shot clock then taking time of the clock is more important and can still lead to points on the possession.

In an effiency rating you shouldn’t actually try to win the game on the possession.

For the game it is better to take 30 seconds with a 35% chance of scoring.
For the effiency rating it is better to run the court to take 10 seconds for a 40% chance of scoring.

So pace of play does matter.
ESPECIALLY if you are better at one style of play compared to another.

I suppose pace of play doesn’t matter at all if you are equal in all situations.

 

[Dano564]

One added point, this is just on the KenPom (Efficiency) side of the equation. The NET evidently also gives bonuses for margin of victory on top of this, so when you win by 10 or more, you get a kickback that adds more weight.

That's why teams like Wofford and Lipscomb are higher in NET rankings than they are in KenPom rankings. Both ranked in both higher than us.

 

[Dano564]

Pace of play has nothing to do with it. Sending IW to the hoop is a good play if it results in a good shot and a conversion more often then a turnover. Of course it is all situational. If there were 40 seconds left in the game and 25 left on the shot clock then taking time of the clock is more important and can still lead to points on the possession.

The formulas adjust for pace of play. Bottom line is you want to score the most points on average per possession. 1.28 per possession it tops right now. Same on defense, keep the other team as low as possible. So if you are getting your ass kicked, you are better off slowing the pace of the game down so you don't dilute your sample more. (Better to go 0.7 points per possession over 50 possessions than 120 possessions against a team you don't play well against). By doing that, you might eliminate your chances to win, but the formulas will like you better.

You don't want 120 of your 3000 possessions to come from your worst game.
You want your worst game to be 40-50 possessions of your 3000 possessions.
Also, if you are dominating a team, run the floor.
Wofford has a few 40 / 50 / 60 point win games. We have none, and that's hurting us.

 

[Some guy]

One added point, this is just on the KenPom (Efficiency) side of the equation. The NET evidently also gives bonuses for margin of victory on top of this, so when you win by 10 or more, you get a kickback that adds more weight.

That's why teams like Wofford and Lipscomb are higher in NET rankings than they are in KenPom rankings. Both ranked in both higher than us.

Which actually makes sense to cap it at 10.

Efficiency of every possession matters. But due to margin of victory measure, for the most part closer games are given higher weights.

For instance, if you win by 8 you’re given a higher rating for being more efficient than your opponent leading to an 8 point victory plus an 8 point margin of victory.

If you win by 58 you are rewarded for being more efficient than your opponent leading to a 58 point victory but they don’t double count the victory by also getting you a margin of victory bonus of 58.


Basically when the game is within 10 points effiency is being counted multiple times but when it is more than 10 it’s only being counted once.

 

[Some guy]

The formulas adjust for pace of play. Bottom line is you want to score the most points on average per possession. 1.28 per possession it tops right now. Same on defense, keep the other team as low as possible. So if you are getting your ass kicked, you are better off slowing the pace of the game down so you don't dilute your sample more. (Better to go 0.7 points per possession over 50 possessions than 120 possessions against a team you don't play well against). By doing that, you might eliminate your chances to win, but the formulas will like you better.

You don't want 120 of your 3000 possessions to come from your worst game.
You want your worst game to be 40-50 possessions of your 3000 possessions.
Also, if you are dominating a team, run the floor.
Wofford has a few 40 / 50 / 60 point win games. We have none, and that's hurting us.

Yup.

 

[Dano564]

Which actually makes sense to cap it at 10.

Efficiency of every possession matters. But due to margin of victory measure, for the most part closer games are given higher weights.

For instance, if you win by 8 you’re given a higher rating for being more efficient than your opponent leading to an 8 point victory plus an 8 point margin of victory.

If you win by 58 you are rewarded for being more efficient than your opponent leading to a 58 point victory but they don’t double count the victory by also getting you a margin of victory bonus of 58.


Basically when the game is within 10 points effiency is being counted multiple times but when it is more than 10 it’s only being counted once.

My point is, if you win by 112 to 58 over 90 possessions, your Efficiency Spread for that game is a whopping 65.
If you call off the dogs and win 70 to 45 over 70 possession, your Efficiency Spread for that game is only 35.


So my point is, heavily weighting efficiency this much is fine, but then giving a bonus for margin of victory from 1 to 10 points (with added weight for each point) is essentially double-counting.

That's why some of these teams are outperforming their KenPom rankings with weaker schedules.

 

[Bordergopher]

True, next year we won’t even have a tourney.

They will just crown the champ based on who gives up the fewest points per possession and who scores the most.

They will throw in a magic multiplier to adjust for strength of schedule. And voila, there is your theoretical national champion

Ha! Love this! Overemphasis on the stats is idiotic.


Sent from my iPhone using Tapatalk

 

[GoGophersUMN]

The formulas adjust for pace of play. Bottom line is you want to score the most points on average per possession. 1.28 per possession it tops right now. Same on defense, keep the other team as low as possible. So if you are getting your ass kicked, you are better off slowing the pace of the game down so you don't dilute your sample more. (Better to go 0.7 points per possession over 50 possessions than 120 possessions against a team you don't play well against). By doing that, you might eliminate your chances to win, but the formulas will like you better.

You don't want 120 of your 3000 possessions to come from your worst game.
You want your worst game to be 40-50 possessions of your 3000 possessions.
Also, if you are dominating a team, run the floor.
Wofford has a few 40 / 50 / 60 point win games. We have none, and that's hurting us.

This pretty much sums up why efficiency stats shouldn't be used for important decisions, other than to break a tie when everything else is too similar.

The goal in the tournament is to win games. You get the same credit for winning by 1 as winning by 100. Yet in the regular season, you're penalized for behaviors that could help your team such as playing quickly while down to bring the game closer, intentionally fouling at the end of the game, and resting starters when you know you've won. If you're looking at efficiency scores, there could be situations where it's better to lose in a particular way than win naturally.

I love math and love that there's a way to make things more objective but if more efficient teams aren't always better, it shouldn't be weighed as heavily as thing like record, strength of schedule, or even the eye test.

Hypothetically you could have 2 teams from the same conference play the same exact schedule and have team A have the same wins, plus win the head to head game(s), plus win a few more games that team B lost, but still have team B look better because they're more efficient. How does that make any sense?

 

[GopherJake]

I’m sorry, but I’m with Pitino on this. This stuff is all idiotic. This whole analytic approach- if it is indeed weighted heavily enough to worry about- is going to change the coaching, get players hurt, when they should be on the bench, not allow lesser players to ever see the floor

I am (at least was) a stats guy. I am also a math guy. I don’t give a rats ass which team is more efficient, how many 25 point wins a team has, instead of 15 point wins, how many 7 point losses versus 22 point losses. All that really matters is who won. Not how they won.

When star players get hurt up by 9 with 12 seconds left going strong to the basket, this will not be worth it.

It’s stupid.

/rant

 

[Dano564]

I’m sorry, but I’m with Pitino on this. This stuff is all idiotic. This whole analytic approach- if it is indeed weighted heavily enough to worry about- is going to change the coaching, get players hurt, when they should be on the bench, not allow lesser players to ever see the floor

I am (at least was) a stats guy. I am also a math guy. I don’t give a rats ass which team is more efficient, how many 25 point wins a team has, instead of 15 point wins, how many 7 point losses versus 22 point losses. All that really matters is who won. Not how they won.

When star players get hurt up by 9 with 12 seconds left going strong to the basket, this will not be worth it.

It’s stupid.

/rant

Efficiency is a stat derived from trying to predict wins and success, which is fine looking back on history. But when you put the cart in front of the horse and use it forward looking, the goals of a basketball game change which influences the way a game is typically played.

 

[Dano564]

KENPOM updated.

We passed VCU. Our efficiency spread is now 12.50. Up from 12.32. (A small move). We are currently 53rd.