sth

I’ve had a fascination with Southampton since they came back to the Premier League. Maybe it’s due to having replaced their manager year upon year without any tremendous signs of decline in team performance. Maybe it’s due to the hilarity of Liverpool constantly poaching their players over the years for huge sums of money. Maybe it’s due to having a good track record of selling high on players when the opportunity arises. Whatever it is, Southampton continue to be an interesting team and this year is no exception.

At the beginning of the season they were projected by most to be around 8-10th place with some bookies having their point tally nestle around 50 points. They’re currently on pace for around 47 points with models predicting similar results, so they’re headed for a point total a bit below their over/under but not disastrously so. For what it’s worth, when we were doing season previews on the site, I projected that their best-case scenario would be finishing around 7th and being the best of the rest. So, on the face of it, one could argue “Well they’re more or less projecting where they finished so what’s all the hubbub about them this season?”

Well here’s the thing…

I think Southampton have been considerably better than what their standings project. Dare I say, they might even be good.

I know it’s weird to say that a team 3 points behind Tony Pulis’ Theater of Dreams could be considered so, but I don’t think it’s such a far-fetched thing to suggest. How they’ve done it is through shot volume. Whereas Arsenal last season looked to be gaming xG by destroying opponents in shot location, Southampton’s plan is to focus on volume in attack and shot locations defensively. Some of what’s happening in attack is probably due to having players like Sofiane Boufal and Nathan Redmond, who are more concerned with getting a shot off and figuring out the rest later (especially in the case of Boufal). In terms of the percentage of shots in attack that are classified as Big Chances, Southampton rank in the bottom 4, and the three teams below them are either in lower mid-table (Burnley), fighting relegation (Hull), or doomed (Sunderland). Even then, all the rudimentary shot data screams of this team being better than it has shown; xG models project them to be around the Europa League spots. They have an impressive statistical resume for a team that’s stuck in midtable purgatory.

I’ve gone through video to come up with possible reasons as to why Southampton are settling for lower quality shots among a healthy volume at 14.7 per game. One thing that stood out was that Southampton can lack the ability at times to occupy the central areas with any conviction. Both Oriel Romeu and Steven Davis would be classified as players who don’t take risky passes to find a better shot opportunity, primarily wanting to circulate the ball to the outside. You could survive and probably do just fine if you have one of them, having two of them particularly in a double pivot seems riskier. This is compounded by a lack of centerbacks that are game changing passers. CBs that can be comfortable on the ball are such a commodity in today’s game, and having one that isn’t good on the ball represents a problem. The advantages of having one includes being able to bypass self inflicted problems in the midfield. Against Tottenham, they had problems whenever the ball was given to Maya Yoshida. He’s fine when it comes to simply recycling the ball to his fullbacks, but things get hairy when he’s asked to do more.

I don’t really know how much these things truly contribute to their inability to get a higher % of good quality shots. Putting a number on such things isn’t easy. It’s fair to ask if playing someone who’s more adventurous alongside Oriel Romeu would brighten things up. Steven Davis seems like a good soldier who’s got some mobility, and can recycle the ball to teammates. I’m sure that’s valuable in some way but it just seems to put more strain on Romeu, who isn’t exactly Cesc Fabregas as a passer.

It’s not all bad however. The more I watch, the more I really enjoy Nathan Redmond as a player. He still has problems with shot locations, but he is a clever player who can operate in tight spaces. His constant movements into the more advantageous areas of the pitch is one way for Southampton to compensate for having Tadic stationed out in a wider position whenever they play in a 4-3-3. I have a soft spot for wingers who can operate in half spaces both by passing and dribbling, and I really think Redmond is underrated in it. Imagining a full season of sequences like this would get me excited if I was a Southampton fan.

While at Norwich, Redmond did well relative to his teammates in terms of being able to create shots for himself or for others. He was the type of purchase that other mid-range PL sides should’ve been looking at (Hello Everton). Again, it feels like the point is being beaten to death, but probably the major thing that’s keeping him from being a star PL player is his lack of shot discipline. Now you can ask multiple questions from this, primarily; is it the systems he’s played in that have hindered him from getting better shots, or is an Andros Townsend/Ross Barkley situation where the basic concept of what a good shot constitutes isn’t getting through to them?  I hope it’s just been him being unlucky to have played on two straight teams that didn’t prioritize shot quality.

Defensively, Southampton are quite passive. They are more concerned with defending in compact situations in what resembles a 4-4-2 shape. They’re not a team that will hurry opponents like Tottenham/Man City/Liverpool. In fact, they’re perfectly fine at times with giving the opposition backline loads of space to just pass it to each other, so long as there’s no threatening outlets through the middle. For the most part this has worked, though when the setting gets more chaotic, that structure can collapse like Jenga blocks.

Numbers

Despite the things I brought up about the deficiencies, which are legitimate on some level, every general shot metric we use to identify team quality paints them in a positive light.

  • Greater share of total Shots? Check. They’re between 56-59% whether it’s all shots, shots on target, or unblocked shots
  • Greater share of Big Chances? Check. Once again hovering around the mid 50’s.
  • Expected Goals? Yes, they’re doing fine in that department as well, to varying degrees depending on the model at hand.

We care about these things because we care about things that can be controlled. Variance is something that is at the mercy of every team, and it could be purely conversion rates or even stuff like injuries. An example of this is people in the analytics community pegging Liverpool to bounce back this season. One of the reasons why was that they produced impressive shot numbers without much in the way of conversion luck. In conjunction with a reasonable enough offseason of transfer purchasing, those are usually the teams to be looked at for bouncing back and have a better season.

To some extent Southampton are doing what Liverpool did last year, but without even the league average conversion rate that Liverpool had. As it currently stands, Southampton are in the bottom 20 in conversion rate and bottom 10 in save rate of teams since 2009. The only other team to have accomplished this feat was the fabled 2015-16 Aston Villa side, and they were the living embodiment of The Benny Hill Show. The level of variance that has gone against Southampton has been the equivalent of going from their current standing of mid-table obscurity to fringe Europa League contender, which has been the space they’ve occupied over the past few years. Tactics obviously matter, and people way smarter than me would hypothesize logical reasons for what Southampton aren’t doing to maximize their talents. But when you’re carrying the anvil of historically bad variance on your back, I mean what the fuck can you really do? The chance that they revert towards a mean next season is high, and any movement there will give them a tangible boost.

A comparison that could be made with Southampton this year is Tottenham, especially in attack. The journeys the two teams take are vastly different, but the end results are quite similar in attack. They both believe in volume first and location second. Southampton are basically a diet coke version of Tottenham. And yet one of these clubs have gotten much friendlier variance than the other. It should be noted that even without variance Tottenham’s attack has been better in both volume and shot location because they have considerably more talent to work with. Also, while finishing skill is still something of a murky topic, there are hints that suggest the likes of Harry Kane, Christian Eriksen and Son Heung-Min could be above average finishers.

I wish I had an amazing answer for why Southampton are being screwed over, but I really don’t. There are things they could improve on when it comes to where their players are positioned on both ends, and that could tilt the balance of play more in the positive direction, but let’s call a spade a spade; they’ve just been exceedingly unlucky, especially in defence. Until this season, all the public goalkeeping data available suggested that Fraser Forster was a fine enough shot stopper, and this season he’s just been putrid. Something resembling a normal season from him and there wouldn’t be a need for articles like this to surface. The more seasons we witness, the more likely we will see weird outliers at both ends, a list that Southampton exist on. Until very recently, they were flirting with bottom 10 levels of conversion on both sides of the pitch, which no one in the Opta era has ever accomplished. Sometimes shit just happens, and while 38 games on the face of it sounds like a massive number, it’s still a small enough sample size to have variance run amok.

Future

Southampton this season have ebbed and flowed a bit in quality and currently reside as a solid team, but without any variance help whatsoever, which is a recipe for a mediocre finish. I’m certainly not saying that Southampton have been good enough to contend for a top 4-6 finish, they’re frankly just not talented enough. But they’ve been much better than either their goal difference or place in the table indicate. They’ve clearly decided that with the mishmash of talent in attack, beating opponents by volume was the recipe for success. Defensively, they’ve done fine in both volume and location. All of this should’ve worked if they could stop a beach ball in goal or weren’t carrying a ~25% conversion rate. Everton are 17 points ahead of them in the table, and I would very much argue there’s very little difference in quality both anecdotally and quantitatively between the two clubs. It’s clear that somehow, someway, Southampton have the ability to plug and play with managers and maintain being a ~55% shot or xG team. That is not a small feat, and if the time comes that the club retreats to being a lower rung club in the PL, it should be among the very first things mentioned on their tombstone.

In terms of going forward this summer, the club should do more of what they always do: bet on 18-23 yr old players that they can sell for two-three times what they paid for. A version of Nathan Redmond who takes better shots is worth a lot more than what they paid for him last summer, particularly because he has shown this season to be a positive contributor with his playmaking. Boufal as a player divides opinion; I’m still a fan of his and think he’s got the ability to be a good player, but it can’t be denied that he’s been a disappointment this season (though having a knee injury early in the season certainly didn’t help things). The process again was the right one, it just hasn’t worked out so far. This team could go with some more creativity within the squad, particularly in the deeper midfield positions, which could open their team to take better shots. It’s also fair to suggest that it could be time to start looking for a younger replacement for Dusan Tadic. He’s been very good for major parts of his PL tenure, but he’s turning 29 in November and though one could guess his game would age better than most because it isn’t heavily reliant on athleticism, he’s still at the end of the age curve where things can get hairy fast.

By points alone, this is the worst season Southampton have had since their first season back in the PL in 2012-13. There have been things to like this season even with their disappointing position; particularly the signing of Manolo Gabbiadini, and the growing evidence that Nathan Redmond could be a very good PL player. Fraser Forster can’t possibly be this bad next season considering his past success, and Virgil van Dijk having a contact reportedly until 2022 means the club can sit back and let the market bid crazy money. What is done over the summer will be very intriguing, and if they keep thinking rationally in the market instead of overreacting to their points tally because of variance being totally against them, there’s a good chance Southampton could once again be a top 6-7 side by the end of next season. Claude Puel’s tenure has had a tricky start, but if he stays on, there’s plenty to work with.

  • allanderek

    Thanks for a very interesting read.

  • Jamie Kitching

    A great read. I’d love to see a comparison of shot locations over the last couple of seasons to see the difference

  • NunyaBusiness

    Puel happened.

  • Paul Tiensuu

    Thanks for a good article, very sensible look at this. I’d also add one thing: extremely tough luck with injuries. For the most of the first half of the season, when Southampton still played EL too, they had to play their 3rd choice player on either RB (Martina, the first 2 being Pied and Cedric) or LB (McQueen, the first 2 being Bertrand and Targett), sometimes both. On second half, they lost their first CB pair. This was more a transfer failure though, maybe they should have signed a ready-to-play replacement to Fonte and not trust that VVD can rest fit for all season. But I understood that they chose Puel largely because they wanted somebody who can bring in young players and work with what he gets, like he did at Nice, instead of buying ready made players.

    Anyway, it is always tricky to cope with a lot of defence injuries. A weak, inexperienced defence not only makes you vulnerable at back, but the that vulnerability also makes it harder to play attacking football. It’s risky to trust a defence line with two of McQueen, Stephens, Martina and Yoshida to support a very attacking formation, and may ask from the midfielders to drop deeper to support them.

    And yet, Puel did formidably well with that. In the 2-1 win against Inter, they had a backline of Martina-Dijk-Yoshida-McQueen against Inter’s first team and won shots 21-5, including a missed penalty. The same defence line only conceded 6 shots at Prague for a 12-6 shot difference in an unlucky loss that cost them the place in the EL knockouts, and dominated Inter at Milan, too, for a 13-8 shot difference, missing a bunch of big chances in the end, reducing Inter to 0 big chances. They lost a lot of matches like this, with that defence line, dominating the game, losing to a team that hardly got shots (against Palace and WBA, for example), that Forster just let slip in. Looking at it that way, they actually did fine even with the 3rd choice laterals, until VVD got injured. With Van Dijk out, that Yoshida-Stephens CB pair made them way too vulnerable at back.