Tuesday, September 27

2016 USAU Women's Data

(This entry is best read in PDF form b/c I care less and less about formatting things multiple times for multiple formats and if you view it there, you can see all columns of all tables.)

So!  There I was preparing to spectate at USAUltimate’s 2016 Club National Championships in Rockford, IL.  I went to look up stats for the division I’d be watching (Women’s) and I was wholly unable to find anything useful.  Not on Ultiworld, not on USAU’s Triple Crown Tour page, not on Skyd, not on reddit/r/ultimate, not on team pages… nowhere.  

This is a problem as I like to have some concept of what it is that I’m watching.  I buried my head in baseball stats as a kid, dive into basketball stats year round, and play fantasy football largely without watching football games because I like numbers.  

I wondered what I was going to be watching at #NationalsIL and I couldn’t find much of anything. How many throws per game?  What’s the average completion percentage? What about Possessions per goal?  Which defenses give up the highest completion percentage?  Which defenses are most likely to convert their break opportunities?  Which teams convert the highest percentage of their “Shots” (aka throws into the endzone)?

I wasn’t concerned so much with individual player stats because without the background knowledge of overall team stats, individual stats are mostly meaningless.  I’m also certain that my “eye test” will tell me which players are outstanding and which fall in with “the great unwashed masses”.

So I sought out women’s games from 2016 against club competition (Not against All-Star Teams) and hoped they would be reasonably evenly matched.  I went to my team’s Ultiworld subscription and ended up with 16 games. With a few weeks between the end of my coaching season and the start of #NationalsIL, I had a bunch of charting work to do.  I charted the 16 games available from Ultiworld and then added 1 game filmed by Georgetown Ultimate at MA Women’s Regionals.  Since then, I’ve found more games (from folks like Hallie’s Dad) but simply lacked the time to chart them.  The games I charted are as follows (Team which received to start is listed first):

6ixers 13 v 8 Schwa ESC
Nemesis 13 v 5 Heist ESC
Underground 6 v 13 Phoenix ESC

Brute Squad 13 v 12 Molly Brown PFF
Schwa 8 v 13 Brute PFF
Nightlock 10 v 13 Scandal PFF
Scandal 11 v 13 Fury PFF
Traffic 13 v 10 Fury PFF
Brute Squad 15 v 14 Molly Brown PFF
Riot 13 v 11 Brute PFF

Brute 10 v 13 Riot PEC
Riot 13 v 8 Nightlock PEC
Molly Brown 13 v 6 Scandal PEC

Scandal 13 v 15 Molly Brown USO
Fury 11 v 15 Molly Brown USO
Brute Squad 15 v 7 Scandal USO

Scandal 14 v 11 Green Means Go MAR

Teams which received to start the game went 10-7 and outscored their opponents 11.71 to 11.00 per game.

There were 5049 Throws (3633 Off, 1416 Def), 542 of which were Incompletions (230 Off, 212 Def) for an average completion percentage of 89.27% (90.92% Off, 85.03% Def).  Of those 5049 Throws, 547 (374 Off, 173 Def) were “shots” (aka “a throw into the endzone”) and of those 547 shots, there were 386 goals (267 Off, 119 Def)  yielding a shooting percentage of 70.57% (71.39% Off, 68.79% Def).  There were 386 Points Played, 267 resulted in a Hold (69.17%) while 119 resulted in a Break (30.83%).  There were 595 Offensive Possessions and 330 Offensive Possessions yielding Off. Poss/G of 2.228 and Def. Poss/G of 2.773.  Overall Conversion % of Offensive Possessions were 44.87% while Defensive Possessions were 36.06%.

Offensive teams averaged 6.11 Throws per Possession while Defensive teams averaged 4.29 Throws per Possession (Overall: 5.49 Throws/Poss).  Offensive teams average 13.61 Throws per Goal while Defensive teams averaged 11.90 Throws per Goal (Overall 13.08 Thr/G).  Offensive teams took a shot at the endzone on 62.86% of their possessions while Defensive teams took a shot at the endzone on 52.42% of their possessions (Overall 59.13% of Possessions ended in a scoring opportunity.)

In the first half of games, there were 204 Goals (52.84% of Total Goals.  152 were Holds [74.51%], 52 were breaks [25.49%]) while in the second half there were 182 Goals (47.16% of Total Goals.  115 were Holds [63.19%], 67 were breaks [36.81%]).  

That’s a buncha numbers in paragraph form.  Let’s take a look at what the “average game” looks like broken out by halves and by Winning Team (left side) and Losing Team (right side):

H1

2016W





Winner
Loser






H1


Pts
G
EZ
Cmp%
Inc
Thr
Poss
Poss
Thr
Inc
Cmp%
EZ
G
Pts


O
5.24
4.65
5.82
93.28
3.29
49.00
7.88
10.35
60.65
6.12
89.91
6.18
4.29
6.76


D
6.76
2.47
3.47
87.16
3.65
28.41
6.12
3.29
13.24
2.71
79.56
1.24
0.59
5.24


T
12.00
7.12
9.29
91.03
6.94
77.41
14.00
13.65
73.88
8.82
88.06
7.41
4.88
12.00


















H2

2016W





Winner
Loser






H2


Pts
G
EZ
Cmp%
Inc
Thr
Poss
Poss
Thr
Inc
Cmp%
EZ
G
Pts


O
4.71
3.59
4.65
92.36
3.65
47.71
7.24
9.53
56.35
6.35
88.73
5.35
3.18
6.00


D
6.00
2.82
3.53
86.26
3.59
26.12
6.35
3.65
15.53
2.53
83.71
1.94
1.12
4.71


T
10.71
6.41
8.18
90.20
7.24
73.82
13.59
13.18
71.88
8.88
87.64
7.29
4.29
10.71


















GM







Winner
Loser






GM


Pts
G
EZ
Cmp%
Inc
Thr
Poss
Poss
Thr
Inc
Cmp%
EZ
G
Pts


O
9.94
8.24
10.47
92.82
6.94
96.71
15.12
19.88
117.00
12.47
89.34
11.53
7.47
12.76


D
12.76
5.29
7.00
86.73
7.24
54.53
12.47
6.94
28.76
5.24
81.80
3.18
1.71
9.94


T
22.71
13.53
17.47
90.63
14.18
151.24
27.59
26.82
145.76
17.71
87.85
14.71
9.18
22.71


The first thing to remember when looking at this is just how thin the margin between winning and losing is.  The losing team throws 17.71 Incompletions while the winning team throws 14.18 Incompletions.  That’s a difference of 3.53 Incompletions per game resulting in a final score differential of 4.35 (13.53 to 9.18 is the average final score.) Of course, that’s a little simplistic, which is why my second thing to look at is a table of simple derived statistics (Top five rows are the winning teams, middle rows are the losing team, bottom rows are the difference from Winning minus Losing):

Win
Possessions
Throws
Conversions
Pos/G
Pos/Pt
Pos/EZ
Thr/G
Th/Pt
Thr/Pos
Thr/EZ
EZ%Th
EZ%Pos
Shot%
Conv%Pos
Conv%Pt
Off.
1.836
1.521
1.444
11.743
9.728
6.397
9.236
10.827
69.261
78.652
54.475
82.840
Def.
2.356
0.977
1.782
10.30
4.272
4.373
7.790
12.837
56.132
75.630
42.453
41.475
Total
2.039
1.215
1.579
11.18
6.661
5.482
8.657
11.552
63.326
77.441
49.041
59.585













Lose
Possessions
Throws
Conversions
Pos/G
Pos/Pt
Pos/EZ
Thr/G
Th/Pt
Thr/Pos
Thr/EZ
EZ%Th
EZ%Pos
Shot%
Conv%Pos
Conv%Pt
Off.
2.661
1.558
1.724
15.66
9.166
5.885
10.15
9.854
57.988
64.796
37.574
58.525
Def.
4.069
0.698
2.185
16.86
2.893
4.144
9.056
11.043
45.763
53.704
24.576
17.160
Total
2.923
1.181
1.824
15.89
6.420
5.434
9.912
10.089
54.825
62.400
34.211
40.415













Diff
Possessions
Throws
Conversions
Pos/G
Pos/Pt
Pos/EZ
Thr/G
Th/Pt
Thr/Pos
Thr/EZ
EZ%Th
EZ%Pos
Shot%
Conv%Pos
Conv%Pt
Off.
-0.826
-0.037
-0.281
-3.919
0.562
0.512
-0.912
0.973
11.273
13.856
16.901
24.315
Def.
-1.713
0.279
-0.404
-6.562
1.378
0.229
-1.266
1.794
10.369
21.927
17.877
24.315
Total
-0.884
0.034
-0.245
-4.706
0.241
0.048
-1.255
1.463
8.502
15.041
14.830
19.171

The things which stand out the most to me are the differences in:
Defensive Possessions per Goal (Winning team at 2.356, losing team at 4.069),
Throws per Goal (Winning teams take 4.706 fewer throws per goal),
Shooting Percentage (Winning teams shoot 75% or higher while losing teams shoot under 65%),
Possession Conversion Percentage (Winning teams convert 54% of O-Line Possessions and 42% of D-Line Possessions while Losing teams convert 37% of O-Line Possessions and 24% of D-Line Possessions
Point Conversion Percentage (Winning teams Hold on 82.84% and Break on 41.475% while losing teams Hold on 58.525% and Break on 17.16%)


With that as a primer, what can we learn about the individual teams involved?

If we consider each team in aggregate (not splitting for O/D or by halves), and each team’s opponents in aggregate, there are some basic notions which emerge from the games considered:

  1. Throws per possession range from 4.034 (Phoenix in 1 game v Underground) to 6.986 (“Riot’s Opponents” in 3 games vs Riot).
    1. For teams with more than 2 games, Throws per possession range from 5.006 (Molly Brown) to 6.848 (Brute Squad)
  2. Completion Percentage ranges from 82.00% (Underground in 1 game v Phoenix) to 93.4% (“Nightlock’s Opponents” in 2 games vs Nightlock).
    1. For teams with more than 2 games, Completion Percentage ranges from 87.8% (“Molly Brown’s Opponents” in 5 games vs Molly Brown) to 92.1% (Brute Squad in 6 games)
  3. Conversion of Possessions ranges from 17.24% (Heist in 1 game v Nemesis) to 65.00% (“Nightlock’s Opponents” in 2 games v Nightlock).
    1. For teams with more than 2 games, Conversion of Possessions ranges from 33.53% (“Molly Brown’s Opponents” in 5 games v Molly Brown) to 52.70% (Riot in 3 games)
  4. Shooting Percentage ranges from 42.11% (“6ixers Opponent” in 1 game v Schwa) to 92.86% (Traffic in 1 game v Fury)
    1. For teams with more than 2 games, Shooting Percentage ranges from 65.17% (“Molly Brown’s Opponents” in 5 games v Molly) to 82.98% (“Fury’s Opponents” in 3 games v Fury)
  5. Throws per Goal range from 30.80 (Heist in 1 game v Nemesis) to 7.462 (Traffic in 1 game v Fury)
    1. For teams with more than 2 games, Thr/G ranged from 17.103 (“Riot’s Opponents” in 3 games v Riot) to 9.795 (Riot in 3 games)

Now for some theories:

Brute Squad is better at taking the easy passes than other teams I’ve watched this season.  Not just in Women’s, but in Men’s as well (I have watched on mixed game in 2016 and I didn’t watch all of it.  We had to travel to the stadium at the US Open to watch the Women’s and Men’s finals).  They have a tendency to take the pop-out (“dishy”) on shallow in-cuts which serves to change the angle of attack and get the disc to a thrower who will have seen the field before becoming the thrower.  I think this will serve them well at Nationals.  I’m not certain that ultimate requires “a crunch-time go-to option” the way basketball does.  It might, but that said… it might more require a group of players willing to just be simply and do what the situation requires before any individual excellence comes into play.  Then, when it is time for said excellence to shine, said excellence shall make all the difference.  More like soccer (a weak link game) than basketball (a strong link game).  A team needs game-breakers on offense and defense in ultimate, but would be wise to refrain from over-reliance on the top of the roster.

One way to compare a team’s aggression (taking shots) to their ability to possess (throws per possession) is to subtract Throws/Goal from Throws/Possession.  For every team which fails to complete 100% of their passes, this number will be negative.  In the Women’s division (in the games I saw on tape) these numbers ranged from -2.612 to -25.490.  Of teams with 2 or more games, Riot was the leader at -4.633.  The next closest? Fury at -7.267 (Followed by Molly at -7.472, Brute at -7.827, Scandal at -8.032, Schwa at -10.270, and Nightlock at -11.690).  The distance between Riot and the rest of the group is notable.  The interesting bit is should teams just shoot earlier in their possessions or must they develop more comfortable ways of maintaining possession until ideal shots present themselves?

A way to compare these is to use Throws/EZ to (and Shooting %) to determine the efficacy of Shots and the frequency with which teams take/generate shots at the endzone.  But… I’ll leave that alone for now.  I shall leave you, dear reader, with a decent amount of this data to sift through yourself.  I can type theories on this all night and day and on, but I need to pack for my trip to watch #NationalsIL (So Ill!), so here are some groupings of data.  I double-checked the underlying data (and speadsheet math) repeatedly, but I could well have messed up some things.  I did my best.  I hope they help illuminate your understanding of 2016 USAU Women’s Ultimate!

6 comments:

parinella said...

Like with all of your posts, I have trouble reading every single word because there are so many, but the first thing I noticed was the break percentage in the second half (36%) was so much higher than in the first half (25%). I wonder if a team can get around this by playing deeper into the rotation and the strategic options in the first half instead of riding the same players and tactics until they start failing.

dusty.rhodes said...

Sure sure... can't please 'em all... when I don't use words and replace them with numbers, I get complaints. When I type words to paint the numbers I get complaints... and "all of [my] posts"? Like all... two of them this year?

But to the point: I suspect (based on what I can recall from the videos) that the better teams in these games start to better leverage their advantages in the second half. I suspect this may have to do with the depth of talent on each team. That is the team most likely to win has a useful depth of 20 or more players whereas the team less likely to win has a useful depth of ~16 players (give or take).

I've been struggling with this as a coach this season in terms of how best to actively use the considerable depth of the team I coached rather than being forced to use depth or being constrained to using the same old strength over and over again. Similarly, the belief that what is working will continue to work is always dangerous. But changing too much or too soon is also dangerous.

Then again... I'm but a neophyte as a frisbee coach. I have tons to learn.

Most of these games felt rather close in the first half, but in the second, the tune changed and the team which would eventually win pulled away rather than letting the team who was behind change things up and get into a new rhythm/pattern. Haven't quite found a way to quantify that... but so it goes.

parinella said...

Posts, comments on those other Frisbee sites, it's all the same.

Do the winning teams also get broken more in the second half? In close games, does the break percentage go up? If it's that the deeper team pulls away in the second half by racking up several breaks, that's one possible explanation. But it feels like I've been in a bunch of games where neither team gets broken much in the first half but a lot in the second half, whereas games that start sloppy but get cleaned up are rare. O players get tired and make lazy cuts or don't cut at all, or run the same play one too many time. Half of defense is just being close enough to catch any swill.

Unknown said...

مكافحة حشرات بالرياض
شركة مكافحة حشرات بالرياض
شركة تسليك مجارى بالرياض

Panharith said...

Like with all of your posts!




Goldenslot สล็อต
goldenslot mobile

Unknown said...

Hello sir, This is the great research you did on the USA women's team data. I always read all your post about the sports and i always enjoyed it. Thank you so much.here you can make your DESMUME speed up .update your old desmume and enhance the speed of your emulator and it design for Windows and Macintosh Nintendo.