Kyogo vs Giakoumakis revisited: Who should Celtic start?

While this may have excited the traditionalists for whom anything other than ‘two up top’ is hipster nonsense, the reality was that Celtic maintained their recent shape. Furuhashi sat behind the Greek striker in what could be termed the ‘10’ role. Recently this has been filled by Sead Haksabanovic and Reo Hatate.

It really is a variant on a 4-2-3-1. However, while formations are of some interest, the reality is that in and out of possession, different shapes manifest around this framework depending on where the spaces are (with the ball) or the danger is (without the ball). I digress.

What some supporters are debating – given it is highly unlikely Ange Postecoglou will revert to a military 4-4-2 – is which striker should be the primary point of this formation?

Giakoumakis has four goals in his last six outings, while Furuhashi has two in his last nine appearances.

Case closed? Of course not. All this being preamble for a proper performance data comparison. 

For the purposes of this exercise, I will be using a combination of StatsBomb and my own data.

Build-up

By build-up, we mean involvement in linking the play. There are three inputs to this:

Receive Packing Score is a fancy way of saying they successfully received and controlled a forward pass that took opponents out the game. It is a proxy for movement, positioning, finding space and control.

Giakoumakis is significantly more available to teammates in terms of receiving the ball. He is also more involved in the build-up play himself, completing over 10 passes per 90 minutes compared to Furuhashi’s nine.

The Japan international completes slightly more progressive runs – moving the ball at least 10 yards inside the opposition half with the ball at feet.

This is maybe a surprise as perceptions say Giakoumakis is the penalty-box striker while Furuhashi is more adept at creative movement and passing… but that is not what has been happening.

Creativity

Let’s build up from build-up and look at chance creation.

Whilst Furuhashi is creating 1.04 chances per 90 to Giakoumakis’s 0.79, their overall expected assist numbers are almost identical with 0.13 and 0.12 respectively.

Where the Greek has the advantage is in passes into the danger zone – the space in front of goal within the opposition box. He provides 0.53 of these per game as opposed to 0.21 from Furuhashi.

Goal threat

All of this has been a teaser to the main dish: goal threat.

Let’s deal with this in two ways. Firstly, goal threat preparation. That is, getting into the positions to score.

The first couple of metrics are essentially about giving up the chance to score – so the negative aspect. Here, Giakoumakis is more likely to lose the ball in the opponent's defensive third (3.28 to 2.9) and is also more likely to turn over the ball such that his teammates are now out of the game (a pack turnover).

Meanwhile, Furuhashi is positioning himself to get more big chances – simply a situation where you think ‘he should score’. 

But Giakoumakis is also getting into a higher volume of promising positions. He has more touches of the ball inside the opposition box and he takes slightly more shots.

Now let’s look at execution:

Simply, Giakoumakis is scoring at a rate of 0.92 goals per 90, whilst Furuhashi’s rate is 0.83. However, Furuhashi has the better expected goals rate of 0.9 to Giakoumakis’s 0.89. 

The suspicion is that because Giakoumakis’s actual goal output is currently higher than the Japanese’s, this leads to the perception that he must be first choice. The reality is their respective expected goals numbers are virtually identical.

StatsBomb data

Finally, we’ll consider StatsBomb radars. The first covers SPFL matches only for both players this season.

The most striking aspect of this is how similar the profiles are. It is almost as if there is actually a coherent system underpinning these individual performances.

Giakoumakis can claim the edge in aerial duel wins (unsurprisingly) as well as touches in the box as seen above. 

Meanwhile, Furuhashi forces more turnovers and has a better-expected goals value per shot taken while also assisting more.

Now we’ll look at StatsBomb’s On-ball Value (OBV) concept to compare the pair:

Kyogo scores negatively for dribbles and carries while scraping over zero on all the other striker metrics – pass, shot, defensive and then overall OBVs.

Giakoumakis scores really well on shot OBV – being in the 99th percentile – and while his pass OBV is zero, this high shot OBV rating elevates him to the 75th percentile on overall OBV.

Summary

With Furuhashi on eight goals for the season and Giakoumakis seven in a squad necessarily rotated due to the relentless schedule, both are delivering.

The underlying metrics suggest the differences between the two players are tiny in performance terms. What may muddy the water further is the consideration that both are playing in a way that the manager demands to fit the team system. That is, the manager has recruited two players with different personal characteristics who nevertheless are capable of following the same team orders with high effectiveness.

Last season, we’d have said that Furuhashi was the primary striker, of international standard and one of Celtic’s most valuable assets. 

Meanwhile, Giakoumakis was a journeyman striker who had been wisely recruited since his strong points are valuable, and his more technical downsides can be accommodated.

This season it looks like the Greek is improving as he becomes more familiar with the team and style, while Furuhashi is perhaps not as sharp. 

He was out for a long time with a grade three hamstring tear; that is a serious injury for an explosive player, and he may not have recovered his previous raw speed and reaction capabilities.

Those, though, are simply some subjective thoughts on explaining the underlying data and the fact that their respective performance metrics are coalescing. 

What is also true is that Celtic have two highly-capable strikers who are both terrific fits to the team style of play. 

So which one starts?

It is a satisfying problem to have, but the sheer volume of matches means both will likely continue to rack up significant minutes regardless.