Welcome Kim English

gmubrian

All-Conference
⭐️ Donor ⭐️
GOLD SPONSOR
I’m not a math guy by any stretch, but I think it only changes sharply like that when a big lead goes away in a short period of time, aka The Mason Special (trademarked)
If we were tied for the last 10 minutes (or whatever amount) and then they scored one point it would go from being in the middle (50/50) to plunging to the bottom. So it always shows a huge plunge if you lose a lead at the end of the game, whether it was 1 point or 30. That is why I suggest that it is an exaggeration, at least in that scenario.
 

gmujim92

Hall of Famer
If we were tied for the last 10 minutes (or whatever amount) and then they scored one point it would go from being in the middle (50/50) to plunging to the bottom. So it always shows a huge plunge if you lose a lead at the end of the game, whether it was 1 point or 30. That is why I suggest that it is an exaggeration, at least in that scenario.
Totally get that, but our win probability had to be in the 90s with an 8-point lead and just over 2 minutes left. Didn’t take long for it to go to zero lol.
 

gmubrian

All-Conference
⭐️ Donor ⭐️
GOLD SPONSOR
Totally get that, but our win probability had to be in the 90s with an 8-point lead and just over 2 minutes left. Didn’t take long for it to go to zero lol.
Agreed. The part above the 50/50 line is an accurate representation. It is the part below the line that is exaggerated in our SLU loss.
 

Petey Buckets

Starter
⭐️ Donor ⭐️
If we were tied for the last 10 minutes (or whatever amount) and then they scored one point it would go from being in the middle (50/50) to plunging to the bottom. So it always shows a huge plunge if you lose a lead at the end of the game, whether it was 1 point or 30. That is why I suggest that it is an exaggeration, at least in that scenario.
SmartSelect_20230112_151159_Chrome.jpg

For sake of comparison, here's Bonaventure. The graph is only dramatic when it really expects you to win.

Also, tied doesn't equal 50/50 - Kenpom bakes in pregame expectations. If we were tied on the road with Houston with ten minutes left, for instance, I wouldn't expect our wp to be higher than 33%. Kenpom expected us to lose to SLU by 6, so being really high means we had a really substantial lead that should have been safe.
 

GMUSSTN

Starter
I was at a Caps game once (year they won the cup) against the Rags and NY went up by 1 with under 10 minutes to go. I was like "ah crap this sucks" but was sitting behind a couple of Rangers fans who were like "no watch we're still going to lose." I didn't believe them, but the Caps scored a late one to tie and then won it almost immediately when OT started. Beginning to feel like those Rangers fans...
 

psyclone

All-American
View attachment 1619
For sake of comparison, here's Bonaventure. The graph is only dramatic when it really expects you to win.

Also, tied doesn't equal 50/50 - Kenpom bakes in pregame expectations. If we were tied on the road with Houston with ten minutes left, for instance, I wouldn't expect our wp to be higher than 33%. Kenpom expected us to lose to SLU by 6, so being really high means we had a really substantial lead that should have been safe.
Also, if one enlarges the graph, kenpom shows the score along the midline after each basket and indeed shows an "x" for each miss (I think) during a possession (or maybe each empty possession). So the graph includes both a team's probability of winning and the current score at each point in the timeline. Often they will also indicate significant "runs" , although none for the SLU game. I'm guessing the run must need to involve a double-digit for one of the teams as SLU had a 9-0 run at the end of the game and we had a 9-0 run at the end of the first half.
 

GMU79

Starter
⭐️ Donor ⭐️
Also, if one enlarges the graph, kenpom shows the score along the midline after each basket and indeed shows an "x" for each miss (I think) during a possession (or maybe each empty possession). So the graph includes both a team's probability of winning and the current score at each point in the timeline. Often they will also indicate significant "runs" , although none for the SLU game. I'm guessing the run must need to involve a double-digit for one of the teams as SLU had a 9-0 run at the end of the game and we had a 9-0 run at the end of the first half.
Seems like 9-0 runs at the end of the game always beats a 9-0 run before halftime. Alas.
 

hoops10

Specialist
I was at a Caps game once (year they won the cup) against the Rags and NY went up by 1 with under 10 minutes to go. I was like "ah crap this sucks" but was sitting behind a couple of Rangers fans who were like "no watch we're still going to lose." I didn't believe them, but the Caps scored a late one to tie and then won it almost immediately when OT started. Beginning to feel like those Rangers fans...

Except for tonight. Rangers scored with .8 of a second left in the game and tied Dallas 1-1. Rangers went on to win in overtime. Almost impossible to do.
 

GMUSSTN

Starter
Except for tonight. Rangers scored with .8 of a second left in the game and tied Dallas 1-1. Rangers went on to win in overtime. Almost impossible to do.
.8 is pretty crazy...I don't think I've ever been watching a game live where one went in with under 3 seconds, let alone 1. Still the 2019 Rangers were very different. As were the 2019 Capitals...
 

gmubrian

All-Conference
⭐️ Donor ⭐️
GOLD SPONSOR
@Petey Buckets and @psyclone You two should get some sort of commission as I just purchased a subscription to KenPom so that I could drill down into the charts to make sure I understand your points as well as to make my point. Don't know why haven't done it sooner. the $20 a year is peanuts compared to what I spend a year on Mason related stuff. Sometimes I can be penny wise but pound foolish...
 

gmubrian

All-Conference
⭐️ Donor ⭐️
GOLD SPONSOR
View attachment 1619
For sake of comparison, here's Bonaventure. The graph is only dramatic when it really expects you to win.

Also, tied doesn't equal 50/50 - Kenpom bakes in pregame expectations. If we were tied on the road with Houston with ten minutes left, for instance, I wouldn't expect our wp to be higher than 33%. Kenpom expected us to lose to SLU by 6, so being really high means we had a really substantial lead that should have been safe.

Also, if one enlarges the graph, kenpom shows the score along the midline after each basket and indeed shows an "x" for each miss (I think) during a possession (or maybe each empty possession). So the graph includes both a team's probability of winning and the current score at each point in the timeline. Often they will also indicate significant "runs" , although none for the SLU game. I'm guessing the run must need to involve a double-digit for one of the teams as SLU had a 9-0 run at the end of the game and we had a 9-0 run at the end of the first half.
I understand your points above. My point can be demonstrated by looking at the last two possessions of the SLU game graph. We were tied at 62 (what I was calling 50/50 earlier but to be 100% accurate to Petey's point the graph shows it as %51.4 chance of SLU winning). We miss a shot and then it plummets down to 69.5% for SLU (whithout a single point scored). Then they score the 1 free throw and it goes to 100%. See my screnshot of that portion of the graph below. So a missed shot caused the graph to plummet almost halfway down and a single free throw causes it to plummet the rest of the way. Whereas, it took 8 points to go from the top of the graph where we lead 62 -54 to where we are tied.

The rest of the graph up to the point where they tied us again is very representative of how we blew the lead. My point, and admittedly my nit with these graphs is very minor, is that when you have a very close game at the end, it exaggerates how bad your team blew a lead.
Screen Shot 2023-01-13 at 14.42.16.png
 
Last edited:

psyclone

All-American
For $20, there's really a wealth of data in Kenpom. Not that he needs me to shill for his site.

I don't pretend to know the formulas that he uses to calculate win probability, but it makes complete sense that if you take a 1 pt lead with your opponent having less than 2 seconds to go the length of the court and score, there's almost a zero chance that you will lose. Small lead with little time left yields big shift in win probability.
 

gmubrian

All-Conference
⭐️ Donor ⭐️
GOLD SPONSOR
For $20, there's really a wealth of data in Kenpom. Not that he needs me to shill for his site.

I don't pretend to know the formulas that he uses to calculate win probability, but it makes complete sense that if you take a 1 pt lead with your opponent having less than 2 seconds to go the length of the court and score, there's almost a zero chance that you will lose. Small lead with little time left yields big shift in win probability.
I wasn't saying it doesn't make sense. I was just saying if you are using the graph to make a point that we lost a big lead which was Petey's initial point that he was using the graphs to back up. He is right. But in this (SLU) case the graph exaggerates how badly we blew a lead. I wouldn't call losing a tie game by one point in the last few seconds blowing a lead but the graph makes it look that way. Thus, it exaggerates it.
 

sleeperpick

Hall of Famer
⭐️ Donor ⭐️
I wasn't saying it doesn't make sense. I was just saying if you are using the graph to make a point that we lost a big lead which was Petey's initial point that he was using the graphs to back up. He is right. But in this (SLU) case the graph exaggerates how badly we blew a lead. I wouldn't call losing a tie game by one point in the last few seconds blowing a lead but the graph makes it look that way. Thus, it exaggerates it.
that isn't how I read the graph
 

gmubrian

All-Conference
⭐️ Donor ⭐️
GOLD SPONSOR
Since we are talking KenPom data in this thread, I'll put this here. I just noticed an interesting tidbit in the KenPom data. Even though Ticket has only been out for 2 games, KenPom has already dropped him down to the "Limited roles (12-16% of possessions used)" category the same tier as Blake Jones. He is 3rd from the bottom overall.

On the flip side, Justyn is listed in the second from the top tier "Significant Contributors (20-24% of possessions used)" and 3rd from the top overall.

I am sure there is a logical reason for that once I read more into how it is calculated, but seems a little counterintuitive at face value.
 

Petey Buckets

Starter
⭐️ Donor ⭐️
Since we are talking KenPom data in this thread, I'll put this here. I just noticed an interesting tidbit in the KenPom data. Even though Ticket has only been out for 2 games, KenPom has already dropped him down to the "Limited roles (12-16% of possessions used)" category the same tier as Blake Jones. He is 3rd from the bottom overall.

On the flip side, Justyn is listed in the second from the top tier "Significant Contributors (20-24% of possessions used)" and 3rd from the top overall.

I am sure there is a logical reason for that once I read more into how it is calculated, but seems a little counterintuitive at face value.
Kenpom's role levels are more about usage rate than minutes. Usage rate is how many possessions a player "uses" while he's in the game, defined as ending the possession in either a field goal attempt, free throws, or a turnover. So what that number is actually telling you is that Ticket is fairly passive, and Justyn is more aggressive. The number is calculated using only possessions the player is on the floor, so time missed won't factor into it.

You opened up a big can of worms Brian. We're gonna turn you into a nerd!

download (17).jpeg
 

gmubrian

All-Conference
⭐️ Donor ⭐️
GOLD SPONSOR
Kenpom's role levels are more about usage rate than minutes. Usage rate is how many possessions a player "uses" while he's in the game, defined as ending the possession in either a field goal attempt, free throws, or a turnover. So what that number is actually telling you is that Ticket is fairly passive, and Justyn is more aggressive. The number is calculated using only possessions the player is on the floor, so time missed won't factor into it.

You opened up a big can of worms Brian. We're gonna turn you into a nerd!

View attachment 1622
Thanks!

Too late, I have been a nerd my whole life. But, I am definitely becoming more and more of a basketball analytics nerd everyday!
 

THANK YOU!

Our donors make all of this possible! Thank you for your continued support of MasonHoops, we couldn't do it without you.

Forum statistics

Threads
1,994
Messages
170,397
Members
857
Latest member
1874patriot
Top