Last Year's and This Years' Teams in the Advanced Stats

calumnus

5:48p, 5/6/24

Last Year Player Min WIn Share per 40 min

Cone 1,138 min .066
Tyson 1,062 min .133
Aimaq 1,027 min .115
Kennedy 783 min .086
Celestine 725 min .098
Newell 681 min .040
Brown 453 min .051
Larson 195 min .088
Askew 146 min .033
Bowser 112 min -.014
Curtis 87 min .027
Okafor 70 min .090
Pavlovic 43 min .027

.100 is an average player. The sum of 5 average players playing 40 minutes is .500 which is the predicted winning percentage against an average schedule.

We had two above average players: Tyson and Aimaq.

A lineup of Cone, Tyson, Aimaq, Kennedy and Celestine is .498

We were essentially at best a .500 team with no depth and Kennedy not available early.

This year's team so far based on last year's WS/40

DJ Campbell .145
Ola-Joseph .138
Petraitis..132
Sissoko.123
Omot .112
Tucker .104
Stojakovic .031
Curtis .027
Pavlovic .027
Dort -.198
Nwankwo ?
Wilkinson ?
Last Player?

We already have 6 known, above average players. Just the top 5 would be .650 (21 wins). Nwankwo may be better than any of them. Stojakovic was a McDonalds All-American. Wilkinson is an excellent HS recruit. Plus year over year improvement And there is still one scholie to give.

A lot of caveats, we don't know how they will fit together as a team, the ACC is not an "average" conference schedule… but I do think Madsen has significantly improved the roster and Cal fans have reason to feel optimistic we will be significantly improved over last year.

stu

6:44p, 5/6/24

In reply to calumnus

I assume the win shares of some of our incoming players are based on playing in weaker conferences. If so, is there a way to correct for strength of schedule?

SFCityBear

6:58p, 5/6/24

In reply to calumnus

calumnus said:
Last Year Player Min WIn Share per 40 min

Cone 1,138 min .066
Tyson 1,062 min .133
Aimaq 1,027 min .115
Kennedy 783 min .086
Celestine 725 min .098
Newell 681 min .040
Brown 453 min .051
Larson 195 min .088
Askew 146 min .033
Bowser 112 min -.014
Curtis 87 min .027
Okafor 70 min .090
Pavlovic 43 min .027

.100 is an average player. The sum of 5 average players playing 40 minutes is .500 which is the predicted winning percentage against an average schedule.

We had two above average players: Tyson and Aimaq.

A lineup of Cone, Tyson, Aimaq, Kennedy and Celestine is .498

We were essentially at best a .500 team with no depth and Kennedy not available early.

This year's team so far based on last year's WS/40

DJ Campbell .145
Ola-Joseph .138
Petraitis..132
Sissoko.123
Omot .112
Tucker .104
Stojakovic .031
Curtis .027
Pavlovic .027
Dort -.198
Nwankwo ?
Wilkinson ?
Last Player?

We already have 6 known, above average players. Just the top 5 would be .650 (21 wins). Nwankwo may be better than any of them. Stojakovic was a McDonalds All-American. Wilkinson is an excellent HS recruit. Plus year over year improvement And there is still one scholie to give.

A lot of caveats, we don't know how they will fit together as a team, the ACC is not an "average" conference schedule… but I do think Madsen has significantly improved the roster and Cal fans have reason to feel optimistic we will be significantly improved over last year.

Can you explain Dort's win share? Is that a minus 0.198? How does a player get assigned a negative win share rating? It is hard to evaluate any meaningful statistic for him, as he only played 19 games over two seasons, averaging about a point and a rebound in 7 minutes per game.

SFCityBear

Growler91

7:00p, 5/6/24

In reply to calumnus

Great post.

calumnus

7:40p, 5/6/24

In reply to SFCityBear

SFCityBear said:
calumnus said:
Last Year Player Min WIn Share per 40 min

Cone 1,138 min .066
Tyson 1,062 min .133
Aimaq 1,027 min .115
Kennedy 783 min .086
Celestine 725 min .098
Newell 681 min .040
Brown 453 min .051
Larson 195 min .088
Askew 146 min .033
Bowser 112 min -.014
Curtis 87 min .027
Okafor 70 min .090
Pavlovic 43 min .027

.100 is an average player. The sum of 5 average players playing 40 minutes is .500 which is the predicted winning percentage against an average schedule.

We had two above average players: Tyson and Aimaq.

A lineup of Cone, Tyson, Aimaq, Kennedy and Celestine is .498

We were essentially at best a .500 team with no depth and Kennedy not available early.

This year's team so far based on last year's WS/40

DJ Campbell .145
Ola-Joseph .138
Petraitis..132
Sissoko.123
Omot .112
Tucker .104
Stojakovic .031
Curtis .027
Pavlovic .027
Dort -.198
Nwankwo ?
Wilkinson ?
Last Player?

We already have 6 known, above average players. Just the top 5 would be .650 (21 wins). Nwankwo may be better than any of them. Stojakovic was a McDonalds All-American. Wilkinson is an excellent HS recruit. Plus year over year improvement And there is still one scholie to give.

A lot of caveats, we don't know how they will fit together as a team, the ACC is not an "average" conference schedule… but I do think Madsen has significantly improved the roster and Cal fans have reason to feel optimistic we will be significantly improved over last year.
Can you explain Dort's win share? Is that a minus 0.198? How does a player get assigned a negative win share rating? It is hard to evaluate any meaningful statistic for him, as he only played 19 games over two seasons, averaging about a point and a rebound in 7 minutes per game.

A negative is worse than a zero. They cause losses. Imagine a player that commits a turnover every time they touch the ball and can't stop anyone on defense.
That said, as you say, the sample size is VERY small. I think it is safe to disregard it.

93Bear

9:42p, 5/6/24

Anyone able to post the win rate for our most recent players the year before last? I'd be curious to see what they did from their prior team/conference vs what they accomplished with us. Hoping this data may give some added insight to this analysis.

calumnus

10:13p, 5/6/24

In reply to 93Bear

93Bear said:
Anyone able to post the win rate for our most recent players the year before last? I'd be curious to see what they did from their prior team/conference vs what they accomplished with us. Hoping this data may give some added insight to this analysis.

Cone
Virginia Tech .103
Northern Arizona .089
Cal .066

Tyson
Texas .101
Texas Tech .144
Cal .133

Aimaq
Mercer .079
Utah Valley .177
Texas Tech .114
Cal .115

Kennedy
Xavier .014
UTEP .092
Memphis .123
Cal .086

Larson
Penn -.043
Cal .088

Askew
Kentucky .054
Texas .079
Cal .038

HoopDreams

10:17p, 5/6/24

I don't think win share is useful in comparing players on different teams or even the same team in different years

calumnus

10:32p, 5/6/24

In reply to HoopDreams

HoopDreams said:
I don't think win share is useful in comparing players on different teams or even the same team in different years

Why not? It isn't based on actual wins, it is the number of wins their individual production would be expected to produce. I agree it isn't baseball (basketball has more covariance with teammatea) and it isn't determinative, but it is an indicator. It is not zero.

Our results as a team were almost exactly what their prior year results on other teams predicted they would be in sum: a sub .500 team. Kennedy available sooner and Celestine healthy all year, plus Madsen finding his best lineup sooner, would have gotten us closer to .500, but not likely over.

That said, ignore it if you want.

75bear

12:41a, 5/7/24

In reply to calumnus

calumnus said:
93Bear said:
Anyone able to post the win rate for our most recent players the year before last? I'd be curious to see what they did from their prior team/conference vs what they accomplished with us. Hoping this data may give some added insight to this analysis.

Cone
Virginia Tech .103
Northern Arizona .089
Cal .066

Tyson
Texas .101
Texas Tech .144
Cal .133

Aimaq
Mercer .079
Utah Valley .177
Texas Tech .114
Cal .115

Kennedy
Xavier .014
UTEP .092
Memphis .123
Cal .086

Larson
Penn -.043
Cal .088

Askew
Kentucky .054
Texas .079
Cal .038

Well this bodes well for the upcoming season. We have a better overall team on paper (take this for what it's worth, obviously).

barsad

1:02a, 5/7/24

In reply to calumnus

calumnus said:

That said, ignore it if you want.

Thanks, I definitely will. From what I can tell from reading about the math on this stat, it is not useful for anything but comparing players WITHIN a team and their relative contributions to team success in one season. It cannot be used to compare teams or players across years, schedules, or pretty much anything else. The betting world wonks will really come up with anything if it thinks gamblers will pay for it.

calumnus

6:01a, 5/7/24

While it is arguably better when comparing players on the same team, it is definitely <intended> as a metric to compare the absolute value of players across teams based on the player's individual stats, not the team results. Obviously player performances in basketball are not truly independent as in baseball, but on the whole it was very predictive last year.

The bottom line is, while it may be difficult to predict our record going into the ACC, I am VERY confident and willing to bet, that our team strength will be improved as measured by our ranking in Ken Pom and Sagarin.

DaveT

6:37a, 5/7/24

Thanks for this post calumnus. It's great info.

You can find a very detailed explanation of basketball win shares here.

The biggest drawback in comparing players across teams or eras is the use of team defense in calculating win shares. While not a big component, it does impact the stat for individual players.

According to the article: "Getting back to my point in a previous answer about how defensive Win Shares are distributed, there's a tendency to reward playing lots of minutes on good defensive teams, even if there's scant evidence for a player's actual defensive abilities."

That's not a reason to ignore the stat when comparing players on different teams, but it adds context. Guys who get lots of minutes on good defensive teams may benefit slightly. We weren't a very good defensive team last year, so I doubt it had much impact on Win Shares for our 2023 roster. Also, several of our new portal transfers played sparingly last year, so again. probably didn't impact them much. As for the others, I guess you'd need to look at how good their team defense was and back that out to know how much the individual player's win shares benefited.

RedlessWardrobe

7:04a, 5/7/24

While all this stuff is good entertainment, its becoming obvious - the upcoming season can't start soon enough.