Stuart Suss
PhD Student
Posts: 1439
Loc: Chester County, Pennsylva...
Reg: 11-21-04
|
03-11-19 12:57 PM - Post#280631
In response to 1LotteryPick1969
<<Any particular reason why Llewellyn's rating is so low?>>
The Win Score (WS, column G) represents a total of the Shooting Efficiency (column U) plus the efficiency number for everything else (rebounds, turnovers, etc . . .)
Gabe Llwellyn (-44.5) and Desmond Cambridge (-59) had very poor shooting efficiencies. Gabe's +17 on everything else still left him at -27.5. Desmond's +21 on everything else still left him at -38. A similar fate befell Gabe Stefanini (+35 on everything else) and Quinton Adlesh (+25.5 on everything else).
On the other hand, if you are Matt Morgan with a shooting efficiency of +70, you can get away with being only +5.5 on everything else.
Or you can be Jordan Bruner, +8.0 on shooting and +78.0 on everything else.
|
penn nation
Professor
Posts: 21298
Reg: 12-02-04
|
03-11-19 01:28 PM - Post#280638
In response to Stuart Suss
Thank you for this. I always look forward to this spreadsheet when you release it.
AJ looks like the leading candidate for POY. As mentioned in another post, if his FT shooting improves he will be completely unstoppable and a runaway leader next year.
|
1LotteryPick1969
Postdoc
Posts: 2280
Age: 73
Loc: Sandy, Utah
Reg: 11-21-04
|
03-11-19 02:17 PM - Post#280647
In response to Stuart Suss
Gabe Llwellyn (-44.5) and Desmond Cambridge (-59) had very poor shooting efficiencies. Gabe's +17 on everything else still left him at -27.5. Desmond's +21 on everything else still left him at -38.
Thanks. Helps me interpret.
You threw me for a while. Gabe Lewullis is now fixing bones in Allentown. I hope he is +infinity.
|
Stuart Suss
PhD Student
Posts: 1439
Loc: Chester County, Pennsylva...
Reg: 11-21-04
|
03-11-19 02:29 PM - Post#280650
In response to 1LotteryPick1969
I apologize to Jaelin for getting his name wrong.
I probably was also thinking of Dr. Gabe Lewullis who used to be a fixture of the summer basketball leagues, well into his middle age years.
|
Stuart Suss
PhD Student
Posts: 1439
Loc: Chester County, Pennsylva...
Reg: 11-21-04
|
03-09-20 03:53 PM - Post#303825
In response to Stuart Suss
The individual efficiency statistics have been updated and now include the 2019-20 season. These conference-only results are available back to the 2006-07 season (excluding post-season tournament and playoff games).
This season's results are limited to those who played at least 140 minutes in league games.
Some thoughts with respect to this year's numbers.
1. Box score metrics like this can only evaluate players based on numbers that are compiled. There is no number that reflects an individual's on the ball defense. There were five top defenders in the league this season. In alphabetical order they were, Justin Bassey, Obi Okolie, Trey Phills, Myles Stephens and Antonio Woods. These numbers underrate those defenders.
2. Box score metrics place a high value on rebounding. Therefore, they favor interior players over perimeter players. Players should be compared to others playing the same position.
3. For those reasons, this chart may be flawed for purposes of determining the player of the year or for first-team honors.
You can select the year by clicking on the appropriate box at the bottom of the spreadsheet. Within any particular year you can rank the players by category by utilizing the drop-down menu to the right of the category heading.
Please review the post at the top of this thread for explanations of some of the categories and the overall limitations of box score metrics. The specifics of the Hollinger Game Score are available.
|
penn nation
Professor
Posts: 21298
Reg: 12-02-04
|
03-09-20 03:59 PM - Post#303828
In response to Stuart Suss
So you're sayin' we should be debating AJ vs Knight instead of AJ vs Atkinson?
|
1LotteryPick1969
Postdoc
Posts: 2280
Age: 73
Loc: Sandy, Utah
Reg: 11-21-04
|
03-09-20 06:24 PM - Post#303851
In response to Stuart Suss
Thanks.
Looking at this through Princeton goggles, I'm surprised how high Friberg and Desrosier rank.
|
Stuart Suss
PhD Student
Posts: 1439
Loc: Chester County, Pennsylva...
Reg: 11-21-04
|
03-18-22 04:55 PM - Post#339539
In response to 1LotteryPick1969
In a day or two I will be posting the efficiency numbers for the 2022, 14-game, Ivy season. Because the rankings are so unusual, I want to state the following.
The formula for calculating the WS (win score) was created by David Berri, not by me. I do not have a personal stake in these numbers. I will defend them because I enjoy being argumentative.
Berri places a high value on rebounds and shooting percentage. Not surprisingly, the top two players in this year’s rankings are good rebounders with high shooting percentages.
Because rebounds are highly valued, interior players tend to score higher than perimeter players.
Each turnover is weighted at negative 1. Each turnover committed during a game is as if the player scored 1 fewer point. Players with fewer turnovers are rewarded by these numbers.
This does not purport to rank the “best” players or those players most deserving of first-team honors. One fan described the value of these numbers saying, “If I already had two stars on my roster, I would draft a player from the top of this list to get the rebounds, and not commit turnovers.”
Because WS numbers are cumulative, the more minutes you play, the higher your score will be. Therefore, I also offer the WS/40, the win score per 40 minutes. That category also identifies those players who were potentially underutilized.
Box score metrics like this can only evaluate players based on numbers that are compiled. There is no number that reflects an individual's on the ball defense. A player who boxes out so that his teammate gets a rebound or who sets a screen behind which his teammate scores is a player who receives no credit in these numbers.
Inefficient shooting is punished. Shooting efficiency is defined as points scored minus shot attempts. Two free throw attempts are considered one shot attempt. If a player attempts 16 field goals and four foul shots, that is considered 18 shot attempts. 18 points scored on those 18 shot attempts is considered a shooting efficiency of zero (breakeven).
There are no bonus points for degree of difficulty. If the shot clock is running out, a player who blows by his defender, then spins around the help at the basket at lays it in gets credit for only one basket. If the player rises up above his defender and buries a contested shot, he gets credit for only one basket. Creating one’s own shot receives no extra credit.
The raw win score represents a total of shooting efficiency plus everything else. Everything else consists of rebounds plus assists plus blocked shots plus steals minus turnovers minus personal fouls (with the designated weights for each category).
When the chart is published, you will find the player who had a shooting efficiency of 21.5, but who was a negative 8 on everything else for a total WS of 13.5. You will find the player who had a shooting efficiency of negative 7, but who was a plus 55.5 on everything else for a total WS of 48.5.
It goes without saying that this is a team game. In 2019, Harvard tied Yale for the regular season title. No Harvard player was ranked in the top 15. But, Harvard had 5 of the top 26 ranked players and 8 of the top 38. Teams that “spread the wealth” win games, even if they do not produce highly ranked players.
Later this weekend, I will post a link to the numbers. I will be intrigued by your comments.
|
Stuart Suss
PhD Student
Posts: 1439
Loc: Chester County, Pennsylva...
Reg: 11-21-04
|
03-20-22 09:58 PM - Post#339677
In response to Stuart Suss
I offer the individual efficiency statistics for the 2021-22 Ivy season.
Please review the preceding post which explains some of these numbers and their limitations.
I previously noted that you will find the player who had a shooting efficiency of 21.5, but who was a negative 8 on everything else for a total WS of 13.5. You will find the player who had a shooting efficiency of negative 7, but who was a plus 55.5 on everything else for a total WS of 48.5.
Additionally, you will find the player who had a shooting efficiency of negative 12, but who was a plus 67 on everything else for a total WS of 55.
And for those of you who were comparing players based on their assist to turnover ratio, I offer the better alternative, the pure point rating found at the right end of the chart.
You should be able to sort the rankings based on any column heading and to look at the rankings back to 2007.
|
LocalTiger
Masters Student
Posts: 440
Age: 58
Reg: 11-15-17
|
Individual efficiency stats 2007-2014 03-21-22 11:49 AM - Post#339700
In response to Stuart Suss
Thanks for all the work and the helpful admission that this data,
as all data, has limitations in drawing real world conclusions.
A few observations:
1) All 5 members of the second team All Ivy are in the top 15
here, including Gainey at 1, Wrighrt at 4 and Choh at 5.
Only two of the first teamer scored as high with Tosan at 3
and Kirkwood at 11. As you said, front court players will score better
in general on these metrics, and three guard made first team-
Llewellyn at 22, Dingle at 29 and Swain at 44.
2)Following your suggestion that pure point rating is a better basis for comparing playmaking efficiency, both Tosan and Llewellyn fare
much better than Dingle (which was the argument).
3). As a Princeton fan, I am happy to see our top 6 players in the top
25 here, but Yale fans could certainly question the predictive value of these metrics. Their best guys seem undervalued here.
4). In general, these metrics seem to favor a role player who does his
job and knows his limits over a player who tries to be a difference maker for his team. I am not sure what to do about that, but it is an observation.
Edited by LocalTiger on 03-21-22 11:50 AM. Reason for edit: No reason given.
|
palestra38
Professor
Posts: 32891
Reg: 11-21-04
|
Re: Individual efficiency stats 2007-2014 03-21-22 12:20 PM - Post#339703
In response to LocalTiger
Efficiency is only an ideal stat as long as you have teammates who can do what you do well enough to draw away double coverage. No way Penn is 9-5 in the League (and the Dartmouth loss was without him) without Jordan Dingle. Frankly, I don't think there was a single player in the League with as much impact on the results as Dingle. When you take 1/3 of the possession time, there usually is a reason for it.
|
LocalTiger
Masters Student
Posts: 440
Age: 58
Reg: 11-15-17
|
Re: Individual efficiency stats 2007-2014 03-21-22 08:26 PM - Post#339720
In response to palestra38
Your opinion is strongly held, but the logic is a little hard to follow.
You think Slajchert is the best rookie, and that Penn's returning talent
should win the League next year. But those same players are so
offensively challenged that Dingle has no choice but to play inefficiently.
Which is it?
Edited by LocalTiger on 03-21-22 08:26 PM. Reason for edit: No reason given.
|
Stuart Suss
PhD Student
Posts: 1439
Loc: Chester County, Pennsylva...
Reg: 11-21-04
|
Individual efficiency stats 2007-2014 03-21-22 09:01 PM - Post#339727
In response to LocalTiger
May I respectfully request that Local Tiger and Palestra 38 move the debate about the quality of Jordan Dingle's teammates to another thread.
By posting these efficiency statistics, I was hoping to start another discussion. Local Tiger, Palestra 38 and others are welcome to join that discussion.
Jordan Dingle ranks 28th (Row 29 is the 28th ranked person as Row 2 is the first ranked person). Azar Swain ranks 43rd. Yet, the coaches unanimously voted each of them to the 1st team, All Ivy. Most fans would have voted identically. Jaelin Llewellyn, who was also voted to the first team (not unanimously) ranked 21st.
As I have acknowledged, box score metrics like this are unable to count non box-score aspects of the game, playing defense, boxing out, setting screens and the like.
My question: What are these efficiency numbers failing to count that produced such shocking rankings for Dingle and for Swain?
|
palestra38
Professor
Posts: 32891
Reg: 11-21-04
|
Re: Individual efficiency stats 2007-2014 03-21-22 09:36 PM - Post#339734
In response to LocalTiger
They were very young, and return everyone consequential. The teams that won won with veteran talent, almost all of which is gone. Plus, Penn had no big men of note until the playoff game. I think you'll see the efficiency ratings increase dramatically next year for Penn. But I think these ratings are almost meaningless when it comes to individual excellence since they penalize those who dominate possession time, an utterly meaningless distinction. Look at shooting percentage for true excellence. If a guard shoots over 50% from 2, he is very hard to defend. 3 point shooting really is as much of a team factor as individual--you have to have teammates who can get you the ball while you are open--it's very very hard to create your own 3. The evidence that Dingle was the best player in the League was his shooting percentage from 2, when he almost always was subject to a double team. Jordan Dingle shot 56.5% from 2 in the League this year, and 85% from the line. Those are numbers of an unstoppable player. His 3 percentage was a bit lower than one would like, but as I said, many were taken out of necessity. To compare, Brendan Barry shot 36% from 2, Choh, 47%, Kirkwood 47%, Llewellen 49%, Swain, 47%. Only Gabbidon and Tosan were comparable from 2, and both had far more help and shot few 3s.
|
penn nation
Professor
Posts: 21298
Reg: 12-02-04
|
Re: Individual efficiency stats 2007-2014 03-21-22 10:10 PM - Post#339743
In response to Stuart Suss
Dingle and Swain have impressive scoring stats, but it takes an awful lot for them to produce them.
Swain had 479 FG attempts this year and Dingle was next at 445. #3 was Llewellyn all the way down at 376. Dingle was only 14th in overall FG% while Swain was 18th. They were 13th and 14th, respectively, in 3 pt shooting as well.
May I respectfully request that Local Tiger and Palestra 38 move the debate about the quality of Jordan Dingle's teammates to another thread.
By posting these efficiency statistics, I was hoping to start another discussion. Local Tiger, Palestra 38 and others are welcome to join that discussion.
Jordan Dingle ranks 28th (Row 29 is the 28th ranked person as Row 2 is the first ranked person). Azar Swain ranks 43rd. Yet, the coaches unanimously voted each of them to the 1st team, All Ivy. Most fans would have voted identically. Jaelin Llewellyn, who was also voted to the first team (not unanimously) ranked 21st.
As I have acknowledged, box score metrics like this are unable to count non box-score aspects of the game, playing defense, boxing out, setting screens and the like.
My question: What are these efficiency numbers failing to count that produced such shocking rankings for Dingle and for Swain?
|
SRP
Postdoc
Posts: 4919
Reg: 02-04-06
|
03-21-22 10:15 PM - Post#339745
In response to palestra38
Stuart, how is this different from the usual problem of not correcting for usage (which some used to call the "Kobe Bryant problem)?
Suppose you think about an offensive possession as exploiting opportunities to score or assist. These opportunities are of different quality, and a better player can score or assist more often at a given level of opportunity quality.
Opportunities arrive with a degree of randomness for any given player. A player who only shoots or makes an assist-type pass with great opportunities, and hence has very high efficiency, could well be less productive than one who is intrinsically better at any opportunity quality level but who also uses more-plentiful, lower-quality opportunities.
For example, being able to dribble by a primary defender and challenge a help defender with a 40% success rate could be more valuable than only being able to hit an uncontested layup at an 80% rate if the former opportunity occurred more than twice as often. (Coaches and teams that can manufacture more great opportunities are rightly lauded, but that just changes the absolute numbers, not the principle that a high-usage player who is relatively good at available opportunities may be worth more than one who has to, or chooses to, wait for very good ones to come by.)
|
LocalTiger
Masters Student
Posts: 440
Age: 58
Reg: 11-15-17
|
Re: Individual efficiency stats 2007-2014 03-21-22 10:20 PM - Post#339749
In response to Stuart Suss
I was trying to have that discussion. As I said,
the All-Ivy second team was generally more highly rated
on these numbers than the first. Some of that, as you
pointed out, reflects a bias for inside players, or at least strong rebounders (Wright). It also struck me, however, that role
players like Friberg and Martz score better than their teams' stars.
you had cautioned that the numbers are better in comparing players
at the same position. Perhaps, I would add, if they play similar roles
in terms of offensive burden. On that basis, comparing Llewellyn,
Dingle and Swain ( and perhaps Lilly) by these metrics does tell us something about their overall effectiveness.
I am surprised Swain does not score higher, and the low ranking of Yale players overall is the biggest puzzle to me.
|
iogyhufi
Masters Student
Posts: 681
Age: 27
Reg: 10-10-17
|
Re: Individual efficiency stats 2007-2014 03-21-22 10:40 PM - Post#339754
In response to Stuart Suss
The usage issue is very real — Bart Torvik tries to control for that with his PORPAGATU! stat (what a mouthful). Link to an explanation: https://www.bigtengeeks.com/new-stat-porpagatu/
In that metric, the coaches' choices are largely vindicated: Tosan is rated 4.4, Dingle 3.7, Wright 3.3, Kirkwood 3.3, Swain 3.0, Llewellyn 3.0.
The one thing I think that's hard to capture with any metric is the particular abilities of the other players on the floor. For example, the reason that Swain's production in the ILT was so impressive to me is that Yale almost always had two complete non-threats to shoot on the court at all times. That shrinks the paint for drivers and makes it easier to give help defense or send double teams. Matt Painter showed exactly what I mean after Azar burned Purdue for those early 11 points: he started icing every ball screen Swain tried to use, knowing that the players who were being left open weren't likely to make him pay for so doing. In my mind, this should be a credit to both Swain and Kirkwood, since they both played on teams that had a lot of minutes taken up by players who weren't major concerns for opposing defenses.
|
SomeGuy
Professor
Posts: 6415
Reg: 11-22-04
|
03-23-22 06:39 PM - Post#339809
In response to SRP
Maybe because I am a little older, I view this as the Dominique Wilkins problem. The year he tore his Achilles, I believe the Hawks offensive and defensive efficiency improved statistically (at least in the unsophisticated way we measured these things back then) once he got hurt. Yet the Hawks lost more. There are situations (often outcome determinative ones) where having a high usage player who can get his own shot is more important than the efficiency stats can pick up.
Of course, I say this from memory without looking up whether it is actually true. Hopefully nobody goes back and proves me wrong.
|
Go Green
PhD Student
Posts: 1153
Age: 52
Reg: 04-22-10
|
03-24-22 09:29 AM - Post#339821
In response to SomeGuy
Maybe because I am a little older, I view this as the Dominique Wilkins problem. The year he tore his Achilles, I believe the Hawks offensive and defensive efficiency improved statistically (at least in the unsophisticated way we measured these things back then) once he got hurt. Yet the Hawks lost more. There are situations (often outcome determinative ones) where having a high usage player who can get his own shot is more important than the efficiency stats can pick up.
Plenty of people on this Board insisted that Dartmouth was a better team without Evan Boudreaux for the same reason.
Yeah. You betcha...
|
|