top of page
  • Socks at Half-Mast

The Big Kahlina: Vol. 2

After 4 matches, I wrote this short piece about how well Kahlina had started the year. If you didn't read it, I would encourage you to take a quick look because I will be referencing the data that was discussed in it and I explained a few of the metrics that I will use in this piece, in that one.


A big point that I made in that first piece was that, after so few games, no definitive conclusions could be made. Kahlina had started out really well, but it wasn't possible to assume he'd continue that productivity. After 14 games, I think we can start to get a better reading of who he is as a keeper.


I want to preface this by saying I will be using the percentiles that FBref provides for Kahlina, however, I will continue to advise caution when interpreting them. Specifically, Kahlina is only 14 games into his MLS career, which means that his percentiles can only be shown in 2 ways:

  1. Over the past 365 days

  2. Over just the 2022 MLS season

In either case, the relative lack of time can distort who is he as a keeper. Fourteen games, in the grand scheme of things, is not a huge sample size. Therefore, while the 2022 season percentiles are good for us to see what kind of form he is currently in, they don't really give us much predictive data (I'm not trying to argue that percentiles are inherently predictive of performance, however, when a player has multiple years in a league, you can start to identify consistencies in their game, which enables you to make some predictions about their future performance). Comparing him to the MLS over the past year, even though he's only played 14 games, is also a bit unfair, because his performance may improve as he learns the league (it may also become worse as the league learns him).

 

One big aspect of goalkeeper play I wont be touching on in this post is the ability to play out from the back and general passing ability. While it is important for keepers to be able to play with their feet in the modern game, it's not a universal necessity. Some teams still do not really ask their keepers to play out from the back, while others insist on it.


Kahlina actually ranks in the 99th percentile for passes attempted over the past year (he's in the 95th percentile for 2022). Meanwhile, Andre Blake (who I will argue is the best keeper in the MLS currently) is in the 32nd percentile for this same category over the past year and the 24th percentile for 2022. Is this a reflection of a lack of ability on Blake's part or of a stylistic choice in Philadelphia? Or both? I don't know for sure, but I would guess it's more the former than the latter, as Blake is consistently in the top end of the percentiles for pass length and launch percentage. To me, that implies Philadelphia is making the choice to play long rather than out from the back.


In Kahlina's case, we know MAR wanted to play from the back. I assume Lattanzio will keep this, but we don't know for sure. Coupled with my inability to watch every single MLS game and goalkeeper, I didn't want to touch on this aspect of goalkeeper play. The main focus, therefore, will be on shot-stopping ability.

 

Let us start with a quick overview of some stats for Kahlina:

Time frame

Goals allowed (GA)

Goals against per 90 (GA90)

Shots on target against (SoTA)

Saves

Save percentage (Save%)

After 4 games

7

1.75

20

16

80%

After 14 games

18

1.29

64

49

76.6%

A few notes about what we see here. First, the GA90 has come down, which is fantastic to see. The SoTA has also come down on a per game basis, as after 4 games he was facing 5 SoTAs per game, while after 14 he's facing 4.6 (ok, so it's not a huge decrease, but it is one). The Save% is down, too, but I'm not super worried about that. An 80% Save% was always going to be difficult (if not impossible) to keep up unless you are truly elite (more on that later).


It's also important to note that in those first 4 games, Kahlina had faced 3 penalties (saving none). Since then, he's not faced one.


Now let's look at a few of the more advanced goalkeeping metrics:

Time frame

PSxG

PSxG/SoT

PSxG+/-

After 4 games

7.7

0.25

+0.7

After 14 games

19.8

0.27

+1.8

Kahlina continues to outperform his PSxG, which continues to be a good thing. I will reiterate that PSxG+/- can indicate luck--and there's always a certain amount of luck involved in sport--however, anyone who has watched Kahlina this year can tell you that this number should be viewed as a reflection of his ability and not due to luck alone. The PSxG/SoT has gone up a smidge, but this metric indicates he is facing slightly more difficult shots than he was through the first 4 games. As such, the fact that he continues to have a very positive PSxG+/- is encouraging. Also, please note that PSxG includes PKs, but PSxG/SoT does not.


Of course this whole discussion is meaningless if we don't compare Kahlina's performance to other keepers in the league. Below are the top 10 goalkeepers in either Save% or PSxG+/-. There is a lot of overlap in these categories except for Alex Bono, who isn't in the top 10 for Save% but is in the top 10 for PSxG+/-, and Joe Willis, who has terrible PSxG data. I also haven't included Stefan Cleveland of Seattle because, while he is in the top 10 for PSxG+/-, he has only played in 3 matches (one per month); although, he is actually tied with Kahlina in PSxG+/-.


The players are in alphabetical order by last name, except for Kahlina, who is first because he's the star of the show.

Player

Games played (minutes)

GA (GA90)

SoTA

Saves (Save%)

PSxG

PSxG/SoT

PSxG+/-

Kahlina, Kristijan (CLT)

14 (1,260)

18 (1.29)

64

49 (76.6%)

19.8

0.27

+1.8

Blake, Andre (PHI)

14 (1,260)

10 (0.71)

50

41 (84.0%)

12.6

0.22

+2.6

Bono, Alex (TOR)

13 (1,170)

25 (1.92)

79

55 (69.6%)

28.9

0.36

+3.9

Callender, Drake (MIA)

6 (540)

7 (1.17)

35

28 (80.0%)

10.8

0.31

+3.8

Clark, Steve (HOU)

13 (1,170)

14 (1.08)

55

42 (76.4%)

16.4

0.29

+2.4

Johnson, Sean (NYCFC)

13 (1,170)

10 (0.77)

38

28 (73.7%)

13.2

0.35

+4.2

MacMath, Zac (RSL)

15 (1,350)

19 (1.27)

76

60 (78.9%)

21.2

0.25

+2.2

Paes, Maarten (DAL)

12 (1,080)

10 (0.83)

34

26 (76.5%)

12.6

0.33

+2.6

St. Clair, Dayne (MIN)

12 (1,080)

13 (1.08)

59

47 (81.4%)

16.9

0.25

+4.9

Stuver, Brad (AUS)

12 (998)

15 (1.35)

51

37 (74.5%)

17.1

0.29

+2.1

Willis, Joe (NSH)

13 (1,170)

13 (1.00)

49

36 (77.6%)

12.9

0.22

-2.1

For my money, Andre Blake is the best goalkeeper in the league right now, especially if we talk about consistency over the past few seasons. Since 2020, Blake hasn't had a GA90 above 0.92 and his PSxG+/- was +5.8 in 2020 and +7.5 last year. Only Maxime Crépeau (now of LAFC, last year of Vancouver) had a higher PSxG+/- than Blake, but Crépau's has never had the consistency of Blake (-2.9 PSxG+/- in 2020 and -1.4 PSxG+/- this year).


Matt Turner is the notable absentee from this group and probably the primary challenger for Blake's title. He's just now coming back from injury. I haven't included him in this table because he's only played 5 games and he hasn't been great (1.40 GA90, -0.2 PSxG+/-). However, if he were staying in MLS I would expect those numbers to rebound substantially and I would have put him in. He's on his way to be the backup at Arsenal, though, so I didn't think it was necessary to include him. With that said, Turner's lowest PSxG+/- came in his debut season for New England in 2018 and was +1.0. Since then, he's had a PSxG+/- of +5.7, +8.2, and +5.9. While his GA90 has never been as good as Blake's--a fact that can also be partially attributable to team defense and philosophy--it's never been poor.


The only other goalkeeper who I think has a claim to best in the league is Steve Clark. He is a late bloomer who joined Portland in 2018 from DC United and took off. In 2018, his PSxG+/- was -1.6 with DC and jumped to +2.8 with Portland. In the 3 subsequent years with Portland, Clark had a PSxG+/- of +7.4, 0.0, and +4.9. The 0.0 PSxG+/- is odd and why I place Blake over him. At the same time, Blake's performance before 2020 wasn't great, so if you want to argue Turner or Clark (or someone else) is/was better, I'll hear it. I just think right now Blake is the current best.


St. Clair is the really interesting player of this group to me. He's only 25, but man do his numbers look good (and sustainable). He broke into Minnesota 2 years ago in 2020 as a 22 year old. In 13 games, he had a GA90 of 1.00 and a PSxG+/- of +2.2. He only got into 4 games last year, while Tyler Miller started 30. Last year, St. Clair wasn't great (2.50 GA90, -1.8 PSxG+/-). On the other hand, Miller was really good (+4.0 PSxG+/-), which explains why St. Clair didn't play. The situation has been reversed this year. St. Clair is getting the majority of game time and, as his numbers above show, is having a seriously good year. With only 34 career starts, it's hard to say what he is, but if I had to put money on it right now, I would bet he's going to be really good.


Sean Johnson gets a lot of flack from USMNT fans, however, I think he's a really solid MLS keeper (I'm not trying to advocate for him to be the national keeper, though). He's consistently in the positive for his PSxG+/-. In fact, his lowest is +1.7 in 2018 and 2019. He's at a career high this year so far (besting his 2020 +3.9). Johnson's Save% is usually in the low 70% range. 2020 was his career best at 77.8% and the only time he's been above 76% for an entire season. You can do a lot worse for a goalkeeper.


Joe Willis is the only member of this group who I'm willing say is a below-average keeper. His numbers on a GA90 basis look really good over the past 3 years at 0.96 (2020), 0.97 (2021), and 1.00 (this year). However, those numbers come with him playing for Nashville. Before then, he had played for DC United and the Houston Dynamo. With DC, he had a GA90 of 1.89 in 23 appearances (4 seasons). For the Dynamo, he had a GA90 of 1.60 in 90 appearances (5 seasons). He did spend one year (2014) in the USL Pro for Richmond, where he had a 0.78 GA90. With USL Pro being a 3rd division league, I don't think we need to include it in our analysis of him.


Willis' PSxG+/- seems to back up Nashville being the reason for his improvement in GA90. Even with Nashville, his PSxG+/- has never been great: +0.6 in 2020, -0.4 in 2021, and -2.1 this year. He was in the negative for Houston in 2018 (-0.7) and -3.8 (2019). Note that PSxG+/- only goes back to 2018 and is only available for some leagues (luckily the MLS is one). Basically, on a goal allowed per game basis, he has become elite with Nashville, but I don't buy it as a reflection of his skill. Before Nashville his GA90 numbers are bad; with Nashville, they're good. His PSxG+/- numbers have never been that good. To me, this indicates his GA90 is the product of the team defense being good in spite of their goalkeeper. He's not a terrible keeper, but I don't think he's that good.


The rest of the goalkeepers in the chart are players about whom definitive statements can't be made yet (at least in terms of their MLS performance) due to age, inexperience in MLS, or inconsistent playing time:

  • Callender is young at 24 and this is his 2nd professional season. He was in the USL League One (3rd division) last year. He started 17 games and had a 1.71 GA90. This year has been a good start for him, but with only 6 games at this level, it's impossible to say whether his good numbers will continue.

  • MacMath is 30, but hasn't played as much as he has this year since he was 24 (17 games). He played in 32, 34, and 29 games for Philadelphia between 2012-2014, then basically became a backup. From 2015 to 2021, he made only 52 total appearances (and, again, 17 of those were in one year). His Save% is at a career high this year and, due to his lack of playtime, I'm hesitant to read too much into his historical PSxG+/-. He's had some years where it's been positive: +2.9 in 2019 and +1.2 in 2020. However, he only had 8 and 6 appearances, respectively, in those seasons. Last year he had 7 appearances and a -0.7 PSxG+/-. The bottom line is that due to his lack of playing time--as well as the inconsistent nature of it--it's impossible to say how good he is. My gut tells me he's just average, at best, since he's spent so many years not being a starter.

  • Bono is similar to MacMath, albeit slightly younger (28). He played a lot for Toronto in 2017 (29 appearances) and 2018 (27 appearances), however, he went the next two seasons only making 10 appearances. He then played a lot last year (24 appearances) and has played a lot this year. His Save% has never been good, as it's usually in the high 60% range. That's where he is this year. His PSxG+/- is also inconsistent. In the two years where he wasn't playing much, he had a positive PSxG+/- of +0.9 (2019) and +2.3 (2020). He only had 7 and 3 appearances, respectively, in those years, though. In 2018, when he had 27 appearances, he had a PSxG+/- of -4.1. Last year he had a -2.2 PSxG+/-. His GA90 was really good in 2017 (1.12), but has never been below 1.67 since. Altogether, I think Bono is probably another average keeper, at best. None of his underlying numbers point to him being a good player who hasn't gotten a chance. In fact, when he has gotten a chance, he hasn't really made the most of it.

  • Stuver is just now getting a shot to be a starter. He made 33 starts for Austin last year at age 29. Before then, he had a total of 9 appearances in the MLS with Columbus and NYCFC. His Save% of 72.7% last year is okay, but his GA90 of 1.61 isn't that good. Both of those numbers have moved in the correct direction this year. His PSxG+/- of +2.7 last year might indicate that the goals allowed were due to team defense just as much as goalkeeper ability. Because he's really only in his 2nd year of meaningful playing time, it's too early to say what he is.

  • Finally, we have Paes who is 24 and in his first season in MLS. Before this year, he had been plying his trade in the Eredivisie. Last year for Utrecht, he made 17 appearances, had a 1.35 GA90 and a 65.6% Save%. There is no PSxG+/- data for the Eredivisie. Being that he is young and new to the league, I don't think that we can make to many conclusions about his ability. The numbers this year look really good, but it will take some time to know if they're a true reflection of his ability.

 

So what about Kahlina? Well, he's a true unknown.


FBref has no data on the Bulgarian First League, even though he was playing for Ludogorets (who have literally won the Bulgarian league every year since they joined in 2011; they also have an interesting history that includes the club going defunct, reforming, and merging, so check out the Wiki if you are so inclined). He also only made 19 total appearances for Ludogorets in 2 seasons.


Aside from this year, the only data on Kahlina is from 2019-20 when he was with Gorica in the Croatian top flight. He made 36 appearances and allowed 47 goals, which was good for a 1.31 GA90. There's no information on Save% or PSxG for any of his years except this one. Additionally, FBref only has data on him from his age 26 season. Before then he was playing for clubs and/or in leagues that it does not track. You can view his FBref page here.


As such, it's really hard to say what Kahlina is. His 14 appearances this year are the most he's made since 2020-21 when he made 14 for Gorica and 9 for Ludogorets. With another player, I'd be taking a very conservative approach in their evaluation; I have to do the same here. He's been good for Charlotte so far, but we have so little information on his historical performance that I really don't think we can make a statement yet on how good he truly is. The obvious exception to this is if you are someone who knows the Croatian or Bulgarian league. If you do and can speak to his performances in those leagues, let me know!

 

Below are some radar charts using FBref's scouting reports for 6 keepers: Kahlina, Blake, Turner, Clark, William Yarbrough, and Tim Meila. Each player is linked to their full scouting report.


These aren't the prettiest radars you'll ever see, but I think they'll get the information across all the same. If you're not familiar with these types of infographs, the basics of them is you want the lines to extend out towards the outer part of the circle. The closer they are to outer part of the circle, the better a player is in that category.


I have included two goalkeepers we haven't discussed yet: Yarbrough of Colorado and Melia of Sporting KC. Yarbrough is in the middle (ish) of the league for both Save% and PSxG+/-. Meanwhile, Melia has arguably been the worst keeper in the league this year. He's at the bottom of PSxG+/- and 2nd to last in Save%. Only Turner has a worst Save%, but, again, if Turner were staying in MLS, I wouldn't expect that to remain. They are both included as a reference point when comparing Kahlina to contemporaries. Blake, Turner, and Clark are arguably the 3 best keepers in the league or, at the very least, they're among the best. Yarbrough represents an "average" keeper, while Melia represents a bad one.


Finally, note that all of these percentiles are on a per 90 basis.

Kahlina, data via FBref
Kahlina, data via FBref

Kahlina's pretty good in most categories, with the exception of a few. The most notable areas of weakness are the PSxG and the PSxG/SoT. Very basically, with PSxG/SoT, the higher the number, the more difficult a shot the keeper has faced. In Kahlina's case, he's not facing difficult shots. This isn't necessarily a bad thing! It's not like we want our keeper to have to save difficult shots. If he is, it probably means that the defense is poor. While the PSxG percentile isn't great, I'm more concerned about his PSxG-GA, which indicates whether he is stopping more goals than expected. He is, so I remain happy.


What I think is important to note about his radar, though, is that he's not among the best in anything. His current profile is that of an above-average goalkeeper in MLS, but not an elite one.

Yarbrough, data via FBref
Yarbrough, data via FBref

Yarbrough is a prime example of a middling goalkeeper. His PSxG percentile and Save% are both really good (I mean the PSxG is in the 99th percentile range over the past 365 days, so 'good' is an understatement there). His goals against is also impressive. However, when looking at his PSxG-GA, PSxG/SoT, and shots on target against: they're all low! He's not facing a ton of shots and the ones he is facing aren't particularly difficult shots, yet he's not saving as many as you would expect. This profile suggests a keeper who is greatly benefiting from good defense.

Melia, data via FBref
Melia, data via FBref

Melia has been bad this year. So has Sporting, as they sit 14th in the West on 13 points. They have a -13 goal differential. In fact, they've allowed exactly double the amount of goals as they've scored (26 goals allowed to 13 goals scored). His profile shows someone who is decent at stopping crosses and that's about it. Yes, he faces a lot of tough shots (note the PSxG/SoT), but he's not saving anything. KC's poor defense probably has a lot to do with his poor radar, but it's not the only factor. His PSxG-GA still suggests someone who is allowing more goals than he should be.

Blake, data via FBref
Blake, data via FBref

This is the profile of an elite keeper. Yes, Blake benefits from a good defense in front of him, as he's not facing a ton of shots and the ones he does face aren't particular difficult. However, when he faces shots, he stops them. The primary trait I want in my keeper is an ability to prevent goals, whether the shots are easy or hard. Blake simply does that.

Turner, data via FBref
Turner, data via FBref

I wanted to include Turner's radar to show how one bad season can affect a profile. Remember, he's only played in 5 games this year after returning from injury and he hasn't been particularly good in them. It has absolutely affected his radar. With that said, I would note his PSxG-GA and PSxG/SoT. For the PSxG/SoT, he's facing tough shots, both this season and over the past year. Even facing tough shots--at least over the past 365 days--he's saving them.

Clark, data via FBref
Clark, data via FBref

Clark's radar shows a really good keeper, but it's not quite as impressive as Blake's. Many of the same traits we see in Blake, we see here, namely, decently low shots on target, not particularly difficult shots, and over performance in saving goals in relation to the PSxG.


The last point I want to make with all of these radars is something I've been touching on with most of them: it's not really possible to completely isolate goalkeeper play from defense. Metrics like PSxG help us do it to a certain degree, but not completely. What we see is that even the best goalkeepers benefit from good defenses that don't allow high quality shots. The best goalkeepers, though, are the ones who still make the saves even if they don't face a ton of shots

 

Kahlina has been a really solid addition to Charlotte. Many fans want to make the case for him being the best keeper in the league, and I just don't think we can. He's certainly been above average and I'm hopeful that as he learns the league, he will improve. He's only 29 (about to be 30), so he's still young for a goalkeeper. Better defenders in front of him, as well as a better attack that takes the pressure off of the defensive, could certainly benefit him.


It would fantastic if Charlotte had the best keeper in the league, but it's not necessary. Kahlina has already shown he can keep us in games, as well as win them for us, with his shot-stopping ability. Even if he only ends up being a top 10 goalkeeper in MLS, that's still incredibly valuable. Depending on which metrics you use, you could probably make a case that he is already a top 10--or even top 5--goalkeeper in the league. Regardless, Kahlina doesn't appear to be a goalkeeper that will hold a team back and that is a good thing for Charlotte and its fans.

73 views0 comments

Recent Posts

See All
bottom of page