What's wrong at Carlisle United. Again.

With the hangover from 23/24 turning into a multi-year headache, it's time to dig deep into the stats and work out what's going wrong...

This is a companion piece to this one from last season, which gives an indication of how long things have been going awry for…

Confounding pre-season expectations of not being absolutely terrible, Carlisle United sit in the relegation zone of League Two and, already, are four points away from safety. The side themselves came into the season an unknown quantity full of known ones. My own pre-season prediction had them 9th and, in the round of pre-season predictions, that was probably one of the more pessimistic ones going around. Carlisle have, instead, served to massively undershoot even that prediction. Perhaps more concerningly from that prediction, the recent run of what looked like easier games pre-season has seen Carlisle slip from “this might be a problem” to “this might be a crisis”. What on earth has gone so wrong this season?

The surface level stats

Carlisle’s xG of just under 15 is 11th in the xG table, well behind Walsall at the top (albeit, so is everyone) and in a cluster of about 15 sides topped by Chesterfield and propper up by Crewe whom we could throw a blanket over and describe them as creatively mediocre.

Of those sides, however, only five actually underperform their xG - Gillingham and Colchester barely, and Carlisle, Morecambe and Bromley - 23rd, 24th and 18th respectively. Evidently from that, it’s fair to say that we’ll need to look at the attack a bit more carefully.

Defensively, Carlisle are in a league of their own.

Well, not quite. Carlisle have conceded the most in the division but their xGA isn’t the highest by a long shot - it’s 11th highest. Similarly to their xG for, Carlisle sit roughly in the middle of a mass of mediocrity with Wimbledon the only real outliers in the division.

What Carlisle do top is the xGA difference vs goals, having conceded 11.41 more than stats suggest they should. Accrington are the only side close to this at 11.29; the next worst are Swindon on 5.57.

When it comes to the volume of shots conceded, Carlisle are decidedly mid-table - 11th fewest shots conceded, 15th fewest shots on target conceded but no side concedes  higher percentage of their shots than Carlisle.

Attacking

Let’s start with the good news first - Carlisle’s set piece xG is 5.07. Carlisle have scored 5 set piece goals. Carlisle’s attacking set-pieces work.

(As an aside, do you know who the worst attacking set piece side in League Two are? MK Dons - given Carlisle were always a side looking to exploit attacking set-pieces, this stat could well fall as they go deeper into the Williamson era. It’s also worth noting that Jacob Blain, the first team analyst running the rule over set pieces in 22/23 when Carlisle were the best in the league at it is now in Carlisle’s recruitment team - what hand he still has in set pieces, I don’t know, but it’s a fair thing to point out.)

That’s the good news done with. Take the set pieces out and you’re left with 9.12 open play xG and 6 open play goals (taking Charlie Wyke’s penalty out of the reckoning). Morecambe at least can say their underperformance is about half-and-half set piece and open play, Carlisle cannot. Their “not scoring” issue is that they simply don’t score from open play.

(As another aside, Carlisle are one of only two sides to not have had a single goal come from their bench this season)

This is borne out by the types of attacks Carlisle have - no side breaks quickly less than Carlisle (their 4 shots from that game state is lowest in the league tied with Swindon - Walsall top this with 16). No Carlisle player is in the top 50 in the league for xG per shot (Luke Armstrong is 51st) and, in general, Carlisle’s xG per shot is joint 5th lowest in the division. Is Carlisle’s issue one of strikers not doing the job or construction?

Chance Creation

Carlisle have the fourth most possession in the league, averaging over 54%.

Carlisle have completed the fourth most passes in the league at 4456 and their 77.7% success rate is also 4th in the league.

Carlisle have attempted the 8th fewest passes in the final third in the league. 

As a proportion of their passes, Carlisle’s proportion of their passes noted as forward passes is the 5th lowest in the league. By contrast, their proportion of lateral passes (aka - going sideways) is 6th highest.

To answer that earlier question - it’s construction. It’s definitely construction. Of the sides with more possession than Carlisle, all three are in the top 8 of touches in the box - Carlisle are 15th. All three are in the top 6 of xG created, Carlisle aren’t. Seven of their players are in the top 30 xA creators in the league - no Carlisle players are in that list.

I could go on. Carlisle have a lot of possession - it’s not in the right areas. Carlisle don’t go through sides - the best attacking sides in the division do.

Defence

As with attacking, let’s look first at set pieces as, in contrast, this is some of the worst news.

Only three sides have conceded more than six set piece goals in League Two this season - Carlisle, Grimsby and Accrington, all of whom are on 8. Now, while it’s obvious that conceding goals is considered a “bad thing”, we need to establish something first. And to do that, we’ll borrow the following chart from Analytics FC:

Now, the reason I’ve borrowed this is because I want it to be clear that, over a full season, there aren’t really that many outliers across the general trendline of positive correlation here. The relegated sides all tend to be poor at both open play and set piece defence and no relegated side are above the average line when it comes to having good open play defence.

The reason I wanted to establish this is because set piece xGA (and xG for, for that matter) tends to be around 25-30% of a team’s total xG. Anything quite a bit higher or lower than that is likely to regress towards the norm and means that, from a coaching perspective, a side could be considered unlucky or needs to work a bit more on them.

With that established, it’s time to make clear that Carlisle’s set piece xGA is 26% of their total xGA. Accrington’s is 40%. Accrington’s set piece issues are very likely to sort themselves out - Carlisle’s are reflective of issues stopping goals from set pieces. As for Grimsby, they sit at 30% but they also have 13 points more than Carlisle - they will almost certainly regress as the season goes on.

It’s also important to note that Carlisle’s 8 concessions came from 41 set piece shots - Accrington’s from 52, Grimsby’s from 58. To put that another way, Carlisle concede a goal every 5.2 set piece shots, Accrington 6.5, Grimsby 7.25.

To put that another, far scarier way - Carlisle concede one goal for every 0.51 xGA at set pieces, Grimsby 0.83, Accrington 0.67. Once you dig down into Carlisle’s set piece defending, the stats become much, much more concerning that the bare numbers suggest.

Here, as an example, are stills from two of AFC Wimbledon’s goals vs Carlisle. As you can see, in both Harry Lewis is surrounded by players. Neither corner actually went into that area but the marking of Lewis meant that he was static in goal while attackers were able to be mobile and go into space. There’s no command of an area, because Lewis can’t actually move - in the second, it was noted on the Brunton Bugle podcast, that a Wimbledon player was stood behind Lewis and had a hand on the shirt of the Carlisle defender in front of Lewis therefore meaning the keeper can’t move. If your keeper can’t claim the ball, then that’s one problem sorted for an attacking side.

This set up is common for Carlisle when defending corners (the above vs Cheltenham. It’s a mix of man and zonal marking, it seems, with a consistent trend of there being three players (two defenders and one  attacker) planted right in front of Lewis. In this instance, the ball went short and there was space to cross in, which was eventually cleared from the near. However, in this instance (and it’s not the only one I’ve seen), the Carlisle “man” markers track the man closest to the near post in this picture, the man on Lewis runs from in behind and only Robson tracking that run allows Carlisle to clear otherwise a dangerous situation would have occurred. It’s a system that challenges a cross given anything at the 6 yard line will have players coming out to clear well, but one that blocks the keeper and leaves a vulnerability to the short corner - quite simply, Carlisle are asking for balls either onto a keeper that’s being blocked off or short and then into a corridor of uncertainty between defence and GK. It also leaves space outside the box to give time to shoot - as Omari Patrick did to devastating effect in the Tranmere game.

From open play, while this is already getting too long to go into the intricacies of PPDA, etc, suffice it to say the following - Carlisle have the most errors leading to a shot in the league, only Grimsby have more errors that led to a goal. Carlisle’s PPDA against is also higher than their PPDA which does also play into the issue of where Carlisle are holding possession in that Carlisle are allowing opposition more time on the ball than they are allowed and therefore, when they turnover the ball, are more likely to turn it over in a deeper area where it’s going to a) take longer to get to a dangerous area and b) is seeing Carlisle almost always be pressed higher up the park, hence why any errors are so dangerous.

Harry Lewis

Obviously, this section was coming because Lewis has quite comfortably become one of the worst signings in the club’s history. However, we can use the stats to detail just how badly things have gone because however bad you think it’s been, I promise you that it’s worse.

Lewis is obviously in the spotlight, particularly after his three stooges moment vs Cheltenham where a near post shot was parried down and then the spin simply took it around and beyond Lewis even nearer to the near post. It is the sort of goalkeeping cock-up that you have to watch a few times just to gauge firstly if it actually happened and then a few more to gauge if it actually complied with the rules of physics. You could defend him and say he was just unlucky. The stats, however, just don’t bear that out.

Opta’s Goalkeeping stats place Lewis comfortably as the least effective goalkeeper in the division with their stats as below. (Image is from Opta - go look at them yourselves, they’re great)

Tommy Simkin of Walsall places second bottom of the League Two regulars with a goals prevented stat of -4.4. Alex Cairns was bottom of the table last season with -14.4 (having played every minute for Salford!). It should also be noted that Accrington, the only side close to Carlisle on the xGA/GA stat, have alternated goalkeepers - Michael Kelly (ex Carlisle!) started the season and Billy Crellin took over after being loaned in late in the window after Kelly’s form was unimpressive.

While goalkeeping statistical analysis is beyond doubt more complex when done properly than this, the indication is that, were Lewis replaced simply by a league average goalkeeper who saves no more than they should and lets in no more also, Carlisle would save themselves 0.79 goals every single game.

If we go back to 22/23, Harry Lewis’ save percentage at Bradford was 74%.and, across the season, prevented over 8 goals going in the back of the Bradford net. Even last season at Carlisle, with Lewis’ minutes amounting to around 20 games, his goals prevented stat was only around -0.31 a game. That’s still not great and would, over a full season, put you around the bottom of your colleagues at other clubs but it’s still around 0.5 goals better per game than Lewis has managed this season and would already mean over 6 fewer goals would be in the back of the Carlisle net.

While there’s no reason to directly point a finger at any single player for Carlisle’s rancid form, the reality is that that regression is absolutely wild. It’s the sort of statistic that makes you wonder what Lewis himself would say is the answer. Undoubtedly, luck must be some of it - take the second vs Wimbledon where he parries a long shot to see it fall straight to the feet of an attacker for an open goal. It’s also fair to say that Lewis does pull off the occasional good save. 

But also look at the other three Wimbledon goals - all from set pieces. Two where Lewis doesn’t really leave his line and is bossed about by an attacker and one where he almost falls in slow motion to try to save it. You could talk about Grimsby’s opener where his positioning leaves a finish across him wide open. Or the very first concession of the season where Lewis drops a cross.

There’s undoubtedly a good goalkeeper there in Harry Lewis. His full season in 22/23 was beyond any doubt good. While there may be many other areas of mismanagement going on, the fact that Lewis hasn’t simply been taken out of the firing line and allowed the chance to get his confidence back (given it must surely be shattered at this point), is damning because Lewis’ Carlisle spell has not just been bad - it’s been catastrophically bad.

So, how would you fix it?

As I hope this shows, Carlisle have three key tactical issues to sort out - passing further up the park, creating chances and defending set pieces.

While “Drop Harry Lewis” isn’t one of those, honestly, he just needs to be taken out of the firing line. If it doesn’t work, it doesn’t work but a fortnight of him just training and talking to get his mojo back is what he needs right now and would give everyone a chance to reset.

Carlisle’s passing and chance creation are indelibly linked and the solution is, ultimately, in how they deal with pressing. Carlisle don’t have either the tactical freedom given to them or the confidence to just play through a press – while Carlisle can have the ball where they don’t hurt sides, a press (especially early - 17 of Carlisle’s 28 conceded have come in the first half, 10 of which are in the first 20 minutes) rattles them if they’re not in a flow. 

Carlisle are reticent to go direct under Williamson (not under Simpson, but Carlisle’s pressing was more intense there) but it’s a valid tactic and Luke Armstrong is nothing if not a willing runner, if Carlisle can get a player alongside him and a shift into something like a 3-4-3 with Amrstong in the middle of Burey and Sadi/Charters/Adu-Adjei (when the latter is fit), may well give them the option of going a bit more direct and turning a situation where they’re being pressed into an attack. If nothing else, having the option to play through/over a press will give opposition pause to think before they press too intensely and get caught out.

As for defending set pieces - allowing the keeper to move would be a start. By comparison to the best set piece defences in the league (Barrow and Bromley), those sides don’t sit on their own keeper. While the box is congested, their keepers all have the space to be able to come out and claim the ball every time. Both are also more aggressive in defending the near post. Those are two quick fixes that will help.

So, who’s the problem?

And on this point I’ll clam up!

While many fans are quick to blame players for commitment, etc, the reality is that there’s a LOT of issues at play here (and I don’t have distance covered stats!). Carlisle are underperforming but that doesn’t mean such underperformance will regress back to the mean. Like last season, the only parallel to what Carlisle are doing is in Forest Green - Forest Green ended last season with a xG difference that was 20 better than their actual goal difference. Carlisle are already at 13!

(For reference, Carlisle in 23/24 L1 were 19.6 worse off than their xGD, 22/23 Forest Green were a ridiculous 32)

I don’t need to remind people what happened to Forest Green Rovers last season but they serve as a warning that underperformance, while statistically weird, is realistically quite sticky. 

These issues did exist under Paul Simpson - his tactics were nothing if not predictable and that carried its own issues - but they’re getting worse. Although a small sample size, on Paul Simpson’s departure, Carlisle were in the top 8 in the division for the amount of touches in the box, they’re now 15th. Carlisle’s xG per shot remains, as then, near the bottom of the league. Carlisle’s defensive duels rate remains, as then, near the bottom of the league.

Mike Williamson may have been Mr Right at some point but he, regrettably, is hard to consider as Mr Right Now. As those metrics indicate, Carlisle have become, at best, slightly better in some areas but it doesn’t compensate for the other areas in which things have gone backwards. Williamson would probably have been a fine appointment in May with a full pre-season to put ideas in and recruit. He probably isn’t who you want in a relegation dogfight.

And, ultimately, that must turn focus onto the ownership. I think every Carlisle fan understands Plan A would have been to give Simpson this season, use that time to get a sporting director in and to work on a plan for the next three seasons from there so that the Piatak family could learn football on the job.

They have not been afforded that time and, while few would doubt their commitment to improving the club off the pitch, the reality is that they have made an appointment that hasn’t worked and, as a result, will have to make a decision to sack Williamson (unless things drastically improve) that reflects directly upon them and reflects directly upon their competence. If they end up looking clueless, it’s a consequence of their own actions and, ultimately, a consequence of the extremely fine margins and steep learning curve that makes running football clubs an extremely difficult thing to do.

It really doesn’t matter who the problem is, we can see what the problems are. The challenges for the players, for Mike Williamson and his coaching team, for Greg Abbott and his recruitment team and for the Piatak family is how they go about fixing them.

Because it’s all well and good sorting the loos at Brunton Park out, but if they’re playing host to National League football next season, then the Piataks will have own the club and own the failure.