Celtic's new Reo Hatate conundrum - James Dailey

Early last week I decided to make Reo Hatate’s performance levels in Celtic’s midfield since his arrival in January the focus of this column – and then he went and had a man-of-the-match performance playing right-back against St Johnstone.

Even prior to Saturday’s performance, the following question popped to mind: has Hatate been working out playing in Ange Postecoglou’s midfield?

There is some bias here – my own ‘eye test’ was the impetus for asking this question and diving into his performance data in a quest to find potential answers.

Let us begin with some attack-focused metrics in league play this calendar year and compare Hatate’s with Matt O’Riley’s:

Obviously, as is often the case with such comparisons, this is not pure apples-to-apples as the players have filled different roles over the sample period.

For example, O’Riley has tended to be more advanced defensively when he has played as one of the two attacking midfielders. He also served as Callum McGregor’s replacement for several games. With that context and caveats in mind, we can see the relative attacking output from both. O’Riley has had higher attacking production in a pervasive fashion across these metrics, including a custom metric which is open play xG assisted/open play key passes.

O’Riley’s 0.147 average xG assisted per key pass from open play is well above the team’s xG per shot over the sample period of approximately 0.12. Basically, the chances O’Riley has created for team-mates have been of high quality on average. In contrast, Hatate’s 0.116 for the same metric lags O’Riley significantly and was even below the 0.12 level.

This next radar offers metrics related to defending and pressing in league games:

The disparity in roles may be evident with this comparison, as O’Riley has often been the more advanced midfielder involved with the high press, so that is an important context to remember when considering the disparities in volumes. However, even when considering a metric like the tackle/dribbled past percentage, Hatate’s production benchmarks quite poorly over the period. The defensive action on-ball value (OBV) metric is similar but neither player benchmark well. 

As we bring these two components together, this OBV-focused radar provides some broader context for each player’s performance output over the period:

Attacking players at Celtic almost always benchmark well versus their broader peer group due to the extent of domestic dominance but, relative to one another, we can see how much more valuable O’Riley’s creative passing has been compared to Hatate’s, even if we account and adjust for his taking some set-pieces (roughly 16 per cent of his overall xG assisted, for example).

Conversely, Celtic players usually benchmark poorly in many defensive metrics that weigh the volume of actions and we again see that reflected in this radar. 

In fairness to Hatate, O’Riley has been producing at a high attacking level in a Celtic side under Postecoglou that has been creating chances at a blistering pace, so how has the Japanese player compared versus some players from the recent past?

Here is Hatate compared to McGregor’s 2018-19 season where the latter’s role, while in a different system, was not terribly different from Hatate’s current one:

We can see that despite playing in a side that generated about 20 per cent less xG from open play compared to Celtic over the period since Hatate’s arrival, McGregor’s attacking output was broadly similar to Hatate’s. 

The OBV radar again showed McGregor offering comparable or higher OBV across various metrics. 

Here was one of McGregor’s fellow midfielders that season, Olivier Ntcham, compared with Hatate:

Despite a reputation amongst many supporters of having not been the most active defensive player, we can see how even Ntcham’s output was significantly higher in that regard than Hatate’s has been. 

The comparisons are reflective of the many comparisons we have conducted with players of the recent past (which you can read here and here) and offer some inference that Hatate’s performance output in league games has been underwhelming to date.

While the sample size is only six games for the Champions League campaign, they offer additional context as to how his profile translated to the higher level of competition. This radar reverts back to comparing with O’Riley for the Champions League campaign:

The benchmarking for the percentiles in the radar is the same for the ones in the radars for league games and, despite the players fulfilling different roles for the team, neither one had performance output that translated all that well on average over the six games. 

As expressed in prior columns, I calculate the OBV metric excluding the shot component due to the relatively low volume of events and the issue is compounded further over such a small sample size of games.

Excluding, that component, the two players were comparable and ranked around the bottom third for the 98 midfielders across the Champions League group stage who played 300 minutes or more.

One of the qualitative concepts at play here is that a Celtic player should be a dominant force domestically if they are likely to be good enough to compete at the requisite level for the club to improve in Europe at a Champions League level.  

Returning back to the question asked at the beginning of this column, the answer seems to be ‘not so far’.

While Hatate still has room to develop and improve, Celtic’s ability to compete in midfield at Champions League level remains in question. Hatate played at full-back for part of his tenure in Japan and his initial foray at right-back was just the sort of dominant domestic performance level referred to.

As the Postecolgou-controlled build towards competing at a Champions League level continues there is a growing feeling that, if Hatate is to have a prominent role to play, it mightn’t be in midfield.