Tossup

Remember back in 2023 when we started keeping track of throw-ins?

Okay, I get it. Not on your to-do list today.

But, hey, we still have a week before Opening Day in Kansas City, and I’ve been banging around on this question for a while.

So we’re gonna go there.

My impression is and has been for some time that the Thorns, going back to Parsons days, aren’t great with throw-ins. Not “Jesus wept do you know how to play this game” bad! Just not as good as many of their opponents.

Part of that is the lack of a big touchline cannon, like Jess McDonald used to provide The Damned Courage and Sam Staab provided the Washington Spirit and now presumably provides the Chicago Red Stars. I’ve mentioned those two specifically for their usefulness around the penalty area; their long throws are damn near as good as a corner kick for attacking the Thorns goal.

The other is the perception that the Thorns are “doing it wrong”, and largely because;

1. The Thorns throwers – and I’m looking especially at you, Kling – seem to take forever dawdling along the touchline, pointing and bluffing. “Slow and static” is a fan complaint I hear from myself, too. And
2. The ball always seems to go harmlessly out or back; even though the rules say you can’t be offside on a throw-in the Thorns almost never try to have a play on that frees up an attacker as the thrower hucks a long throw over the head and away.

Anyway, I thought it would be entertaining to take a hard look at several matches from 2023 to see if they could tell us anything about all this ball throwing, much like our look at corner kicks helped get a sense for how the Thorns did with that kind of setpiece.

The 2023 Throw-in Database:

Here’s the final throw-in table from 2023:

Opponent	Advantage gained	Advantage lost	Opponent gain	Opponent loss
Houston	38%	38%
Louisville	24%	28%
ACFC	30%	25%
NCC	35%	27.5%
Houston	26%	26%	18%	32%
Chicago	23%	6%	55%
San Diego	39%	22%	63%	26%
Seattle	52%	26%	66%	11%
Orlando	16%	19%	55.5%	22%
Chicago	20%	30%	58.3%	16%
Washington	36%	20%	40%	8%
Kansas City	23%	14%	32%	45%
Gotham	17%	23%	31%	5%
NCC	66%	6%	16%	4%
Washington	26%	45%	64%	8%
Louisville	28%	14%	41%	18%
Seattle	31%	23%	56%	12%
San Diego	20%	40%	41%	22%
Gotham	29%	41%	69%	11%
Angel City	29%	19%	40%	45%
Gotham	26.6%	29%	44.1%	14.7%
Average	30%	25%	46%	19%

My initial thought was to look at two endmembers; a match the Thorns threw-in well and another they did not.

Those were; 1) Carolina here August 20th (the Hubly Sending-Off had the best Portland won-lost ratio of the season at 66% to 6%) and then 2) Washington away the following week, the worst ratio of the season – 26% “for” – possession from the throw retained – to 45% “against”.

I first pulled up the Paramount+ replay of the Courage match (and let me tell you; watching a soccer match for throw-ins is like watching a dirty movie for the dialogue. I do not recommend it for entertainment) and ran into an immediate problem.

The Thorns throw-in pattern was utterly banjaxed by the red card.

The Thorns took 14 throws. Only four were taken in the first half. Of the remaining 10, nine were taken after the 73rd minute as the Thorns were just killing off the game (and nicely, too, but still…).

That’s why the ratio was so good; all the throws were super cautious, intended only to keep possession and succeeding.

Ugh.

That still left Washington away as the “bad” match and I screened that and annotated the throw-ins.

I still needed another match, a “good” one. The next-best Thorns throw-in ratio was the Seattle away win in June (52% of the throws retained, 26% lost). So I screened that match.

I still wanted an “apples-to-apples” comparison between opposing clubs for throw-ins, meaning the reverse fixtures with either Seattle or Washington.

I went purely by entertainment value for the Smith hattie, but in so doing I also wondered about Washington; those wenches can throw! Both their matches recorded the highest “plus-minus” ratios of any Thorns opponent. Why? What were they doing differently? Were they doing something differently? So for the June match I tracked both squads.

So I kept track of Washington’ s throw-ins in the June match.

We’ll get there. But first, let’s review how the whole “throw-in” business works as a technical part of the game of soccer.

Throw-ins: How Do They Work?

Luckily we have a nice little study, The undervalued set piece: Analysis of soccer throw-ins during the English Premier League 2018–2019 season (Stone, Smith, and Barry) published in the International Journal of Sports Science and Coaching in 2021.

This work analyzed throw-ins from 380 matches in the English Premier League 2018-19 season. The entire study is worth a look, but I’ll give you the tl:dr from the abstract:

“Higher final league position was correlated to increased throw-in first contact success and possession retention. 83% of throw-in’s resulted in a successful first contact, 54% resulted in possession being retained and 8.8% of throw-ins led to a shot at goal from the possession achieved after a successful first contact. Throw-in’s which went backwards or laterally in direction resulted in increased first contact success, retaining of possession, and shot creation. The least efficient throw-in was forwards and long, which resulted in both reduced first contact success and possession retention. Findings highlight, that throwing the ball laterally or backwards should be a focus for coaches and players during attacking training. In contrast, a team’s defensive strategy should reduce the opportunities to throw backwards or laterally with a higher press and look to force a long forward throw-in, therefore, increasing the likelihood of winning possession and counter attacking.”

Stone, Smith, and Barry (2021)

So, unsurprisingly:
1) Better teams are better at throw-ins,
2) Throwing back downfield or perpendicular to the touchline is most effective, and
3) Forward throws, especially long forward thrrows, are least likely to retain possession.

For the purpose of my analysis I broke the study’s throw-in types down – from forward, lateral, and backward, long, medium, and short – into a nine-region grid:

Short and directly backwards along the touchline was identified as “Region 1”. Short and lateral as “Region 4”, Short and forward, “Region 7”.

Longer lateral throws went into Regions 5 and 6, longer diagonals into Regions 2 and 3 (backwards), Regions 8 and 9 (forwards).

I also kept track of where inside the regions throws went:

So a tiny two- or three-yard lateral dink was a “4-short” or “4S”. a longer throw into the middle of Region 4 a “4-medium/4M”, and one near the far edge, eight or nine yards or so, is a “4-long/4L”.

Right at the far edge the throw became a “Region 4/5”.

Other information I tracked included:
1) Throw-in time (from the time the thrower raised the ball overhead to release),
2) Area of the field where the throw originated. This was generally divided into thirds; attacking, midfield, and defending, and from there into forward, middle, and back areas (so “attacking midfield”, or “defensive midfield”). Throws within 10 yards of both attacking and defensive corners were also noted (“A10/D10”),
3) Thrower,
4) Intended receiver,
5) Region of throw and length inside the region,
6) Initial outcome – either retained possession or lost,
7) Subsequent outcome – was the ball advanced from a successful throw-in? Was it a good throw that was then lost (for example, a successful throw-in to a tightly-marked player who was then tackled for loss) This could also include a “neutral” outcome, such as being knocked back into touch, and
8) Any other notes or comments.

The raw data ended up looking like this:

With that in mind, let’s look at our three matches, beginning with the Washington Spirit match here.

Portland vs Washington, June 23, 2023:

In the fourth week of June the two clubs looked like this:

At the time Washington were top of table on 22 points in 12 matches, with the Thorns second on 20 points with the same played. So no real mismatch in quality at the time, although as we now know the Thorns ended up with a large advantage in the “final table position” of Stone, Smith, and Barry (2021).

This was the match where I tracked both sides. Here’s the Thorns raw data:

Here’s Portland sorted by thrower:

Here’s Washington’s:

How did this break down? Let’s start with Portland in general.

Portland’s Throw-ins by the numbers:

Total throws: 22
Forward throws (Regions 7,8,9): 7 of 22 (31.8%)
Lateral throws (Regions 4,5,6): 12 of 22 (54.5%)
Backward throws (Regions 1,2,3): 3 of 22 (13.6%)
Total lateral and backward throws: 15 of 22 (68%)

So, good! The Thorns were listening to our researchers; more back and lateral, less forward!

Results
Forward throws: 4 of 7 successful (about 57%)
Lateral throws: 10 of 12 successful (83.3%)
Backward throws: 3 of 3 successful (100%)

Right again, Thorns! The average success rate of the lateral and backwards throws ran near 90% compared to only close to 60% for the forward throws. Worth noting that one of the two lateral throws that was lost was a “long” throw into Region 5.

Player Statistics:

Natalia Kuikka (RB) took only 5 throw-ins, or only about 23% of Portland’s total. She did not throw any back to her defenders, hucking four lateral throws (three short to Region 4, one longer to Region 5) and a forward throw.

She was also the most successful thrower, connecting all four that landed in regulation time – the fifth didn’t make it into play before the final whistle.

Reyna Reyes (LB) was the Thorns throwing quantity champion here in June; 15 of 22. Of these:
Six (40%) went short and forward (Region 7),
Six went short and perpendicular to the touchline (Region 4),
Two (13%) went short and backward (Region 1), and
One (6%) went moderately-long and perpendicular (Region 5).

Of her throws a total of 11 were successful (73%); all the backward throws, 5 of 7 lateral throws (about 71%), and 4 of 6 forward throws (66%).

So far that tracks with out studies; more backward and more shorter, more better.

Hina Sugita (RW) took the team’s only other two throws, one back, one lateral, both short, both successful. There’s an interesting thing about Hina-san’s throws, though, which we’ll discuss.

Here’s how the Thorns’ throw-ins looked from overhead. The arrow indicates the direction of attack, red dots are successful throw-ins, purple unsuccessful. A black outline indicates a throw-in that produced a productive attack:

The left-right imbalance is interesting, suggesting a lot of Portland attack went down the left flank, and I don’t recall that, either from matchday or watching the replay, but the data from the pitch doesn’t lie.

How about Washington?

Washington took about 30 throw-ins.

(I say “about” because the Paramount+ director had an annoying habit of cutting to close-ups of players, coaches, random fans, and pieces of paper blowing across the field as players – and especially Washington players – picked the ball up. I know for certain I missed several Spirit throws that way, so the Washington data set is not pristine.)

Most of Washington’s throw-ins by players other than Samantha Staab were short, lateral-or-backward throws. Of the 20 non-Staab total taken I was able to observe 15 throws. Of those 11 (73%) went to Regions 1 or 4.

Staab is a different matter, as we’ll see.

Of the non-Staab throws 12 (80%) were successful which, again, fits our researchers’ model.

Washington’s throw-ins were largely split up between three players:

Garielle Carle (LB) took 6 of the total (20%), nearly all lateral or backwards. I’m pretty sure she went forward-short once. Connected on 5 of 6 (83%).

Dorian Bailey (RB) took 9 of 30 (30%). Something like 7 of 9 were short lateral or back throws. She was successful on at least 7 of these (77%), including both forward throws. I missed one, too, so possibly more.

Sam Staab (CB) was the third of the main trio, taking another 9 throws. As we’ve discussed, she doesn’t dink around – the woman can seriously huck – and Washington doesn’t waste her time tossing short or backwards. Of her throws:
Lateral: 4 throws were lateral but either very long (into Region 6) or at the near edge of Regions 5 and 6. Of those 3 of the 4 (75%) connected – only the single Region 6 throw failed.
Forward: 5 were forward and long, either into Region 8 or the border of Regions 8 and 9. These were less successful, connecting only 3 times (60%).

So far, not so good; less than 3/4 to 2/3 of Staab’s long throws were connecting…but…one long forward throw (in the 46th minute) landed in the Thorns penalty area, wasn’t effectively cleared, and found the foot of Ashley Sanchez who scored to level the match at 2-2.

So there’s that.

I didn’t plot out the Washington throw-ins, but the heavy presence of Bailey and Staab suggests a similar heavily-right-flank weighted attack from the Spirit, as well…

Here’s an intriguing statistic, though.

Timing is Everything:

As I mentioned, one thing that tends to draw a lot of Thorns-fan-critical attention is that it seems to take Thorns throwers foreeeeeever to release the ball. “Slow and static” is something you hear a lot about the Thorns throw-ins.

So I kept the clock time for the two teams’ throwers. Here’s the throwers’ shortest, longest, and average throw-in times, listed from slowest to fastest thrower’s average time:
Carle (WAS): 0:01 to 0:17, average 0:08
Kuikka (POR): 0:02 to 0:10, average 0:07
Three tied at 0:05 –
Reyes (POR) 0.01 to 0:15
Staab (WAS) 0.01 to 0:12
Bailey (WAS) 0:02 to 0:10
and
Sugita (POR) 0:01 to 0:03, average 0:02

I’m not sure why Hina-san has such a quick release, but her N is also very small, so this could be an artifact. Still, her throws in all three games were taken very quickly.

Conclusions from The June Match vs. Washington:

1. Both clubs had throw-in tactics that conformed to the Short, Smith and Barry (2021) model; mostly short and backwards or laterally.
2. Both executed the tactics effectively in general.
3. The individual average time each teams’s throwers took to release were generally within the margin of error of the overall average, and time did not seem to be a critical factor in success; Carle and Kuikka, the two slowest throwers, were also the most effective.
4. Parsons had a distinct use for Staab; it was not generally as successful at connecting as his team’s short throws, but it was effective as a one-time strike- and goal-producing tactic once during the match.

Okay, so. How about the reverse fixture?

Washington vs. Portland, August 27, 2023:

Here’s the squads:

By Matchday 17 the two clubs had begun to diverge; Portland (which had gone top after the June win) was still in second. Washington was on a four-match winless streak (2D 2L) and had dropped to 5th en route to an 8th place finish.

Klingenberg instead of Reyes is the most impactful change from June for our study purposes, although Parsons going from the 4-3-1-2 to a 4-2-3-1 seems to have made a defensive upgrade to the Spirit.

Portland’s Throw-ins by the numbers:

Total throws: 21
Forward throws (Regions 7,8,9): 10 of 21 (47.6%)
Lateral throws (Regions 4,5,6): 7 of 21 (33.3%)
Backward throws (Regions 1,2,3): 4 of 21 (19%)
Total lateral and backward throws: 11 of 22 (52.3%)

This one was not in line with our researchers’ recommendations; barely half of the Thorns throws in D.C. went out laterally or backwards compared to about 68% in the Washington match here.

Results
Forward throws: 5 of 10 successful (50%)
Lateral throws: 6 of 7 successful (85.7%)
Backward throws: 3 of 4 successful (75%)

The average success rate of the lateral and backwards throws was only 80% compared to the nearly 90% in the first match and the 50% for forward throws compared even more unfavorably to the 60% in the earlier fixture.

Unsurprisingly the scoreline reflected that; 1-all instead of 4-2 Thorns.

Player Statistics:

Natalia Kuikka (RB) took the bulk of the throw-ins in D.C.; 10, or almost half of Portland’s total. Continuing her trend from the match here she threw backwards only once. Kuikka also took only two short lateral throws (Region 4) and tried seven forward throws; five short to Region 7, two longer into Region 8.

Kuikka did poorly overall with her throw-ins. She missed on her only backwards throw and one of her two lateral throws. Her forward throws were, as our researchers could have predicted, largely ineffective, only three of the seven (about 43%) succeeding.

Kuikka’s throw-in times ranged from 0:02 to 0:14, averaging 0:07, the same average she posted in the Spirit match here.

Meghan Klingberg (LB) took 8 of 21 throw-ins (her substitute Reyna Reyes took two more). Of these:
Two (25%) went forward; one short (Region 7), one very long (Region 9),
Four (50%) went short and perpendicular to the touchline (Region 4),
Two (50%) went short and backward (Region 1).

Klingenberg threw well in D.C., completing 7 of 8. Her only miss was the speculative very long forward throw.

Kling’s release time was nearly identical to Kuikka’s; 0:02 to 0:14, average 0:07.

Hina Sugita took a single throw to Region 1. She took no measureable clock time to complete the throw, which started a dangerous attack.

Here’s how all that looked:

Much more balanced between Klingenberg on the left and Kuikka on the right, but also much less successful on Kuikka’s side.

Was There Anything Different on the Pitch?

From the video review the Washington press didn’t look tighter on the Thorns throwers or receivers in D.C. than it did here in June.

Here’s a series of screenshots from the June match, first Reyes at 22′:

Weaver is tightly marked and though she received the throw she was immediately dispossessed.

Now here’s Kuikka fourteen minutes later:

Again, Washington is pretty tightly marked up, but Dunn is still open for the throw (and boots the ball away with her first touch, but oh, well…)

Last, here’s Sugita at 50′:

Smith is, again, well covered and though the throw connects Smith is quickly tackled.

Now here’s the reverse fixture in Washington, Klingenberg in the first minute:

Washington is well-positioned but Rodriguez still finds space to receive. She doesn’t have much space, though; she’s forced to go back to Kling who has to boot it long downfield.

Now here’s Kuikka at 36′:

Tons of space for Coffey to take and turn.

I don’t see a lot of difference between the two matches on the screen. The similar throw-in time averages follow the eye-test; suggesting that Kling and Kuikka in D.C. were not having to fight through a tighter press than Kuikka and Reyes did in the earlier match.

Conclusions from the August match vs. Washington:

1. Portland’s throw-in tactics largely violated the recommendations of the Short, Smith and Barry (2021) model.
2. The results, especially Kuikka’s long forward throws, thus proofed the rule by showing the predicted lack of success.
3. The average throw-in times between Kuikka and Klingenberg, as well as between long forward (Kuikka) and short and lateral or backwards (Kling) throws were not significantly different.
4. The fullbacks’ average throw-in times in D.C. were statistically indistinguishable from the fullbacks’ times here in June. (As were Sugita’s consistently shorter times…)
5. Despite having seen them in June Sam Staab’s long throws were still effective as a secondary or additional weapon in August: she hit Ashley Sanchez in the 13th minute only this time Sanchez shot tamely at Bella Bixby.

One last match to look at,

OL Reign vs. Portland, June 3, 2023:

The squads:

Same Portland formation and much the same XI, particularly the same fullbacks as Washington v. Portland in August.

In June the Reign were in second (5W 3L 1D, 16 points) while Portland was in first (4W 4D 1L, 16 points) on goal difference, so no real daylight there.

Here’s the raw data:

Portland’s Throw-ins by the numbers:

Total throws: 32
Forward throws (Regions 7,8,9): 17 of 32 (about 53%)
Lateral throws (Regions 4,5,6): 10 of 32 (31.2%)
Backward throws (Regions 1,2,3): 5 of 32 (15.6%)
Total lateral and backward throws: 15 of 32 (46.8%)

This one was similar but even worse to August with regards to Portland listening to our researchers’ recommendations; now less than half of the Thorns throws at Lumen went out laterally or backwards compared roughly half in D.C. and about 68% in the Washington match here.

Results
Forward throws: 9 of 17 successful (about 53%)
Lateral throws: 6 of 10 successful (60%)
Backward throws: 5 of 5 successful (100%)

The average success rate of the lateral and backwards throws was 73%, less than the 80% in D.C. and well below the nearly 90% in the first Washington match.

The 53% success rate for forward throws is statistically similar to the 50% success in D.C. but much less than the 60% in the June match.

At this point I think we can feel confident that our little analysis confirms that the Short, Smith, and Barry (2021) conclusions are sound.

Player Statistics:

Natalia Kuikka (RB) was again the top throw-in taker; 15 or 32, almost half of Portland’s total. She still didn’t throw backwards much, either; only three throws to either Region 1 or 2. This match saw Kuikka take four short lateral throws (Region 4); not much, but at 26.6% of her throws a tic up from her20% in D.C. As usual, Kuikka threw forward a lot; eight of 15, 53%.

Kuikka threw fairly well overall, connecting on 11 (73%). Her throw completion ranged from 100% (3 of 3 for backward throws) through 75% (3 of 4 for lateral throws, and her only fail was an attempted long-ish throw to Region 5) to 5 of 8 forward throws (62.5%). Again, Kuikka’s throw success shows the difficulty of completing forward throws.

Kuikka’s throw-in times ranged from 0:02 to 0:14, averaging 0:05, significantly faster than in D.C. and here in June. That suggests that Seattle might have been slower to close her down, but the screenshots (see below) aren’t particularly conclusive.

Kuikka also had an odd sort of hockey assist on her one long lateral throw – the Region 5 toss mentioned above – in the 17th minute. She was trying to drop it on Crystal Dunn but hucked it too far over Dunn’s head for the fail she got tagged with.

But. Sam Coffey was lurking, hawked the loose ball, passed to Sugita, who passed to Smith, who scored the opening goal.

Go figure. Sometimes you can fall on your face and still come up looking like a star…

Reyna Reyes (RB) came on for Kuikka at 70′. She took four throws; two in quick succession around the 80th minute; first trying to go short forward to Sugita. This was knocked out for another throw, which Reyes tossed back to Hubly (who started it around the back).

She tried again in the 83rd minute. This was knocked into touch again, and Hina-san grabbed it and took the quick throw in to Smith (who was tackled for loss).

Reyes’ final throw went short and forward up to Michelle Vasconcelos, whose shot went wide.

Her average throw-in time was slow (0:07) but at that point the Thorns were killing off the match.

Meghan Klingberg (LB) took 12 of 32 throw-ins. Of these:
Five (41.6%) went forward; three short (Region 7), two long (Region 8),
Six (50%) went short and perpendicular to the touchline (Region 4),
One (about 8%) went short and backward (Region 1).

Klingenberg was the poorer of the two starting fullbacks, completing only 8 of 12 (66%). That total included completing her only backwards throw but only 4 of 8 (50%) of her lateral throws. Oddly, Kling was a champ throwing forward, hitting four of five (80%) including two long throws into Region 8.

Kling’s release time was nearly identical to Kuikka’s as well as her similar to her own D.C. time average; 0:02 to 0:11, average 0:06.

How’d this all look?

Lots of purple there, so not so great.

What’s peculiar about this is the lack of left-flank attacking-third throws. That’s Weaver Territory, and she usually creates enough Chaos Muppet havoc up there to force clearances into touch. That didn’t happen at Lumen, and I’m not sure why.

Okay, then! Screenshots? Yes; first Kuikka in the 9th minute.

Tons of space to dime Coffey. But Sam has no one unmarked upfield, so she goes back to Kuikka:

Kuikka tries to pass inside to Smith, but her pass is too slow and a Seattle midfielder (Quinn, I think) steps in to pick it off.

Here’s Kuikka again five minutes later. Portland’s defensive third this time:

She doesn’t have much; Seattle is doing a good job marking up:

Gotta go to someone, howsabout Kelli Hubly?

Nah. That’s too far. Let’s go short to Dunn.

Who’s covered like a blanket and is forced into a bad pass for a turnover.

This is where I keep coming back to Staab and McDonald. That long throw can be a weapon on defense, too. Kling can’t huck, so she’s forced into a predictable (and, though initially successful, ultimately futile) short well-covered throw.

Conclusions from OL Reign vs. Portland:

1. Portland’s throw-in tactics, again, failed to follow the recommendations of the Short, Smith and Barry (2021) model.
2. The poor results from the forward throws, and the relatively large number of poor forward throws, showed why those recommendations are sound
3. Kuikka, in particular, seemed really averse to throwing behind her. I’m not sure if it’s her or the centerback (Hubly, usually) who’s there to receive,
4. The average throw-in times were shorter in Seattle than the two Washington games but not grossly so, and the Thorns’ averages in all three games were generally similar to Washington’s in June.
5. Based on this I think our perception of the Thorns throwers as “slow” is a preconception bias. The fullbacks from 2023 appear to have been within the league averages.

General Conclusions:

1. Based on the matches I screened from 2023 the conclusions of the Short, Smith, and Barry (2021) study are well-supported by our small study; throw-ins are most effective going perpendicular to the touchline or backwards.
2. Forward throws – especially long throws – are typically ineffective. so…
3. My initial hypothesis (“There’s no offside on a throw! Why not huck it long downfield..?”) was disproved, at least for Portland in 2023.
4. The Thorns, while not being significantly awful at throw-ins last season, lost more than they should have…
5. Not because their throw-in timing is too slow, but…
6. Because the fullbacks tried to force the ball upfield too often.

Hopefully this coming season the coaching staff will think about this and work on it. We’ve got a new fullback (Muller) and, possibly, a new regular starter (Reyes) who can be coached into a more productive throw-in pattern.

Throw-ins appear to be a very minor tactical element of soccer. But even minor elements can add up. There’s no reason that the Thorns can’t be better at this technique, and if they can they should be.

Next: NWSL is Back!