Better Box Score Metrics: PUL Early Season Review

Sifting through the first bits of PUL action to identify the top performers.

May 14, 2025 by Paul Würtztack in Analysis

Better Box Score Metrics is a statistical analysis column for Ultiworld subscribers, exploring data-driven analysis of ultimate. Enjoy this free preview! And get access to the entire series, as well as our Discord, where you can discuss the contents more thoroughly, as a subscriber.

The weekend ending May 4 marked something close to the chronological half-way point in the Premier Ultimate League’s 2025 season, but more like a third of the season in terms of actual games played, which is why we’re calling this an early season, rather than a midseason, review. (It does not include the games of this past weekend). The player off to the hottest start: Raleigh’s Alex Barnett.

In a league where the average turnover rate in the spring conditions has been 50 yards per throwaway, Barnett threw for 711 yards over two games with a single turn (along with 292 receiving yards). Add in 14 scores and Barnett is the leader in both EDGE and Player Efficiency Rating (PER).¹
(Note that the PUL doesn’t publish possession data, so we’re not offering E100 or E16 ratings).

Let’s run through a few early season leaderboards. Table 1 shows the Top 20 single-game performances, as rated by a standardized combination of EDGE (production) and PER (efficiency). In our most recent WUL update, we noted that we were temporarily calling it CV for combined value, but there’s the overlap with the statistical term Coefficient of Variation, and so we’ve landed on a new name (for now): as a nod to its EDGE and efficiency components, we’re calling it E+. It’s essentially the same method I’ve used over the past few years for identifying MVP candidates. In the PUL version, without E16 as a third component, I’ve used an EDGE:PER weighting of 3:1, weighing production more highly to account for the importance of playing in as many of the games as possible.

Best Single-Game Performances Through Wk 5

PUL WK5 - Table 1: Best Single-Game Performances Through Wk 5

Player	Team	Week	Match	OPP	DPP	G	A	B	T	Y	T	GmSE	EDGE-O	EDGE-B	EDGE	PER	EDGE+
Alex Barnett	RAL	Week 3	ATL @ RAL	17	3	3	4	0	0	583	0	0.45	5.78	0	5.8	160	167.8
Ella Juengst	NY	Week 4	RAL @ NY	16	5	6	0	2	0	412	0	0.45	4.32	0.9	5.2	145.9	158.3
Yina Cartagena	NY	Week 5	NY @ MINN	14	9	1	8	1	3	550	3	0.47	4.55	0.47	5.0	114.5	148.3
Katie Backus Meilstrup	RAL	Week 3	ATL @ RAL	13	2	2	1	2	0	356	0	0.45	3.26	0.91	4.2	143.4	146.4
Jenny Wei	RAL	Week 4	RAL @ NY	17	1	3	4	0	1	428	1	0.45	4.19	0	4.2	129.3	143.2
Audrey Lyman	RAL	Week 3	ATL @ RAL	17	2	1	3	0	1	502	1	0.45	4.1	0	4.1	132	142.9
Alex Barnett	RAL	Week 4	RAL @ NY	17	2	3	4	0	1	420	1	0.45	4.14	0	4.1	130.1	142.8
Genny De Jesus	NY	Week 5	NY @ MINN	14	8	9	2	0	1	302	1	0.47	4.1	0	4.1	123.8	140.8
Yina Cartagena	NY	Week 5	NY @ MKE	10	2	1	4	1	3	436	3	0.23	3.59	0.23	3.8	125.5	138.3
Eli Presberg	NY	Week 4	RAL @ NY	16	6	2	4	1	4	589	4	0.45	3.81	0.45	4.3	105.8	138.1
Madison Cannon	ATX	Week 3	INDY @ ATX	11	2	3	3	0	1	375	1	0.29	3.75	0	3.8	126.9	137.9
Claire Revere	RAL	Week 3	ATL @ RAL	14	1	1	6	0	2	456	2	0.45	3.95	0	4.0	117.8	137.7
Claire Bidigare-Curtis	RAL	Week 4	RAL @ NY	17	1	2	2	1	1	400	1	0.45	3.35	0.45	3.8	122.9	137.4
Laura Grencser	INDY	Week 3	INDY @ ATX	14	3	1	5	2	0	187	0	0.29	2.66	0.58	3.2	145.3	137.1
Quincy Booth	ATL	Week 1	ATX @ ATL	6	5	2	3	2	1	298	1	0.23	3.03	0.47	3.5	133.8	136.9
Ella Juengst	NY	Week 5	NY @ MKE	10	2	4	3	0	4	469	4	0.23	4.04	0	4.0	110.5	136.8
Kira Flores	DC	Week 3	ATL @ DC	3	6	1	4	0	0	270	0	0.35	3.05	0	3.1	146.5	135.4
Audrey Lyman	RAL	Week 4	RAL @ NY	17	1	2	3	0	2	460	2	0.45	3.55	0	3.6	116.1	133.1
Clare Frantz	DC	Week 5	DC @ LA	6	7	2	1	1	0	256	0	0.35	2.53	0.35	2.9	139.7	131.8
Amanda Murphy	DC	Week 5	DC @ LA	6	10	2	2	1	0	222	0	0.35	2.49	0.35	2.8	137.9	130.9

Raleigh and New York have each played two of the highest efficiency games of the season, one of which was against each other, and representatives of those two teams dominate the top of the table, including dual appearances by Barnett, Audrey Lyman, Yina Cartegena, and Ella Juengst. Undefeated DC puts three names on the board, leaving just three spots for the rest of the league.

To account for the wide range of games played by teams, Table 2 offers the Top 10 in EDGE per team game, as opposed to EDGE per game for the player, to remove the advantage for those whose teams played more games. For example, Barnett’s EDGE total is divided by three, even though she only played in two—and yet still leads the league. At season’s end, EDGE per team game will simply equal a player’s total for the year.

Best EDGE per Team Game Through Wk 5

PUL WK5 - Table 2: Best EDGE per Team Game Through Wk 5

Name	Tm	Gms	G	A	B	ThY	RecY	Y	T	EDGE-O	EDGE-B	E/TmGm	EDGE-O	EDGE-B	EDGE	PER	EDGE+
Alex Barnett	RAL	2	6	8	0	711	292	1003	1	9.92	0.00	3.31	5.78	0	5.8	160	167.8
Steph Wood	MINN	1	3	2	1	86	277	363	2	2.79	0.47	3.27	4.32	0.9	5.2	145.9	158.3
Emma Piorier	MINN	1	3	0	1	95	248	343	1	2.69	0.47	3.16	4.55	0.47	5.0	114.5	148.3
Yina Cartagena	NY	3	2	18	2	1247	157	1403	9	11.49	0.70	3.05	3.26	0.91	4.2	143.4	146.4
Laura Grencser	INDY	2	2	10	4	373	79	451.00	4	4.73	1.16	2.95	4.19	0	4.2	129.3	143.2
Ella Juengst	NY	4	11	4	3	132	1052	1184	6	10.33	1.24	2.89	4.1	0	4.1	132	142.9
Jolie Krebs	NY	4	6	14	4	1192	440	1631	16	10.41	0.91	2.83	4.14	0	4.1	130.1	142.8
Audrey Lyman	RAL	3	3	8	0	904	309	1213	10	8.39	0.00	2.80	4.1	0	4.1	123.8	140.8
Quincy Booth	ATL	3	5	9	3	725	388	1113	12	6.50	0.81	2.44	3.59	0.23	3.8	125.5	138.3
Amanda Murphy	DC	2	3	4	1	163	246	408	0	4.50	0.35	2.42	3.81	0.45	4.3	105.8	138.1

The Top 10 in EDGE-B block value (blocks weighted by the game’s scoring efficiency) are shown in Table 3. Raleigh again is represented at the top, but this time by Hannah Boeettcher and Mary Rippe.

Top 10 in EDGE-B Through Wk 5

PUL WK5 - Table 3: Top 10 in EDGE-B Through Wk 5

Name	Team	Gms	B	GmSE	EDGE-B
Hannah Boettcher	RAL	3	6	0.41	2.48
Mary Rippe	RAL	2	5	0.45	2.26
Karli Steiner	LA	3	6	0.29	1.76
Ashleigh Jentilet	DC	2	5	0.35	1.74
Marié Weverka	NY	4	3	0.46	1.39
Cassie Brown	NY	4	4	0.35	1.38
Dawn Culton	RAL	2	3	0.45	1.36
Morgan Lally	ATL	3	4	0.32	1.26
Sam Harris	NY	4	5	0.25	1.25
Ella Juengst	NY	4	3	0.41	1.24

The GmSE value shown is a weighted average of the game scoring-efficiency in which the player registered the blocks.

For the last three tables, we use rates, which also remove the games-played advantage. The first is Completion Percentage Plus (CP+), the BBSM take on standard completion percentage. At some point, we’ll get this into a separate reference paper, but to quickly summarize, CP+ is a measure of how well you maintain possession. It shares the basic framework of “completion percentage” tabulated by most leagues, but with five differences:

1. In CP+, the number of touches includes one’s own drops.
2. Completions are not penalized by having one of your throws dropped. In standard CP calculations, they typically are, simply because they are using “completions” in the team sense as the numerator.
3. The “touch” registered by scoring a goal is considered half a touch, since there’s no second half of the touch—the throw—which is the riskier half. A drop however is considered a full touch, as the drop itself foreclosed the option of making the throw.
4. We adjust the raw count of turns using GmSE adjustments, as in EDGE.²
5. To smooth out low sample sizes, and to avoid setting an arbitrary “minimum touches” threshold for inclusion, a set of touches is gifted to each player at league-average completion rate. For this week’s CP+ calculations, each player was gifted 36 touches and 31.8 completions. This gift will rise slightly through the season.

Table 4 shows the top 10 in CP+ after just a couple of games. Barnett, not surprisingly, is atop this leaderboard as well, but DC registers the most occupants, with Kira Flores currently the first among them.

Top 10 in CP+ Through Wk 5

PUL WK5 - Table 4: Top 10 in CP+ Through Wk 5

Name	Team	Gms	Tch	AdjT	CP+
Alex Barnett	RAL	2	102	1.5	0.959
Kira Flores	DC	1	41	0.0	0.945
Allie Milligan	PHL	2	29	0.0	0.935
Katharine Gilbert	NASH	1	54	1.7	0.934
Marge Walker	DC	1	28	0.0	0.934
Laura Grencser	INDY	2	83	3.8	0.933
Yina Cartagena	NY	3	180	10.2	0.933
Clare Frantz	DC	2	25	0.0	0.931
Audrey Lyman	RAL	3	160	9.5	0.930
Ella Juengst	NY	4	98	5.2	0.929

A good indicator of disc stinginess, CP+ by design considers every completion to be equal. But every completion is not equal, which is the reason we’ve developed PER. in PER, the numerator is not completions but goal-equivalents (GE), a value derived from a player’s yardage and scoring. The denominator is “personal possessions,” which is simply the sum of GE and Adjusted Turns. PER also compensates for the higher risk of throwing by weighting throw yards more heavily than receiving yards; the weighting is league-dependent and designed to ensure balance between cutters and handlers. The PUL publishes neither “effective yards” nor individual huck statistics, so we do not incorporate those as we do for the WUL and UFA respectively.

The PER value itself is a conversion of the standardized z-score set of all players, with 100 the league average. Although the relative values specifically refer to differences in standard deviations from the mean, informally we can consider a value of 120 to be 20 percent better than average. Table 5 gives the top 10 in PER. A player such as Jenny Wei shows how CP+ gives a little more context to the completions one has made. Wei’s CP+ of .910 is only 35th in the league, but she has generated much more production than average from her touches, with 664 yards and 12 scores in two games. Therefore, despite four turns, she climbs into the Top 10 for PER.

Top 10 in PER Through Wk 5

PUL WK5 - Table 5: Top 10 in PER Through Wk 5

Name	Team	Gms	G	A	ThY	RecY	ThT	RecT	PER
Alex Barnett	RAL	2	6	8	711	292	1	0	171
Amanda Murphy	DC	2	3	4	163	246	0	0	152
Clare Frantz	DC	2	4	2	139	236	0	0	148
Kira Flores	DC	1	1	4	180	90	0	0	146
Katie Backus Meilstrup	RAL	1	2	1	110	245	0	0	139
Yina Cartagena	NY	3	2	18	1247	157	8	1	135
Jenny Wei	RAL	2	5	7	341	323	2	2	132
Liz Hart	PHL	2	4	2	113	180	0	1	129
Chip Yen	LA	3	9	1	107	467	2	0	129
Laura Grencser	INDY	2	2	10	373	79	4	0	128

In our last table, we add it all up, again using E+ and a standardization with 100 as the average, and create a too-early pool of MVP candidates (Table 6). We include it now just for the sake of seeing how much it changes when we do our next update after a sizeable group of additional games.

Top 10 in E+ Through Wk 5

PUL WK5 - Table 6: Top Ten in E+

Name	Team	Gms	E/TmGm	PER	E+
Alex Barnett	RAL	2	3.31	171	150
Yina Cartagena	NY	3	3.05	135	139
Amanda Murphy	DC	2	2.42	152	137
Laura Grencser	INDY	2	2.95	128	136
Ella Juengst	NY	4	2.89	126	135
Emma Piorier	MINN	1	3.16	116	135
Audrey Lyman	RAL	3	2.80	124	134
Clare Frantz	DC	2	2.20	148	134
Steph Wood	MINN	1	3.27	106	134
Jolie Krebs	NY	4	2.83	110	131

Legend
G: Goals
A: Assists
B: Blocks
ThY: Throwing Yards
RecY: Receiving Yards
Y: Total Yards
EDGE-O: Efficiency-derived goal equivalents–offense, a composite measure of offensive production, in goal equivalents (GE)
EDGE-B: Efficiency-derived goal equivalents–blocks, a measure of block values, in goal equivalents
EDGE: EDGE-O + EDGE-B, a measure of total goal equivalents produced
CP+: Completion percentage plus, an attempt-weighted measure of how well a player maintains possession of the disc.
PER: Player Efficiency Rate, an index of a player’s production per turnover, adjusted for sample size. League average PER = 100.
E+: Composite value based on z-scores of EDGE and PER, to complement a player’s overall production with their efficiency.

---

1. The PER calculation has been modified slightly; a full update is forthcoming. ↩

2. Adjusted turns (AdjT) is just the EDGE process in reverse. In EDGE, the value of a turn in goal equivalents is based on the Game Scoring Efficiency (GmSE). In windy games with frequent turnovers, the value of any one turnover is less, and this is reflected in the GmSE. For AdjT, we take the GE value of the turn and divide it by the league-average GmSE. Therefore, a turn in a 0.25 GmSE in a league with an average GmSE of 0.33 will be equal to 0.76 AdjT (0.25/0.33). Over the course of a season, values tend to converge toward the raw count, but using AdjT ensures that aberrant wind-tunnel games do not exert excessive influence. ↩