Goodnight, Sweet Prince


In the sixth article in the 2016/17 season review series, I consider the work of the wingers. Arguably the strongest section of the team, last season was the story of three excellent players. Celtic play with either three attacking midfielders behind a striker, or three forwards of which two are wide. Sinclair has made the left forward position his own, usually playing as an inside left, ceding the wing to the advanced Tierney. On the right, Forrest and Roberts competed for the advanced wide position. Forrest would provide width and play as a traditional and direct winger. Roberts would tend to drift inside more, taking up positions as an alternate “10” or mimicking Sinclair’s tendency to play in one.


Sinclair has played for some of the biggest clubs in England, but seems to have finally found a home. Not all moves work out, but having been plucked from Bristol Rovers at age 16, he joined perhaps the largest playing pool in professional football, at Chelsea. Limited first team opportunities and numerous loan spells followed. A familiar pattern at a club which had 38 loan moves outwards last season. Finally, at 21, a successful spell at Swansea City, partly under Rodgers. This was followed by another big club move to the Manchester City galacticos. Again, few opportunities presented themselves and following a dispiriting loan at West Bromich Albion, it was off to free-falling Aston Villa, and relegation. Sinclair, wary perhaps of another wrong decision, and having no doubt benefitted financially from such big money moves, seems to appreciate what Celtic gives him – adoration from the fans, a manager who trusts him and on field success. He was involved in every squad when fit, and played more minutes bar Gordon, Brown and Lustig. This has been only the third time in his career he has passed 40 matches in a season and forth passing 30. At 28, he is at his peak but does not have too many miles in his legs.

Forrest completed his second highest number of minutes. Only once has he passed 3,000 minutes, in 2011/12. He has never started as many matches (34 was the previous high) but only completed 12. In 161 starts for Celtic, he has only completed 90 minutes on 82 occasions, and only 25 times in the last three seasons. Missing an average of over 8 matches a season to injury, Forrest is a player both Deila and Rodgers manage carefully in terms of game time. Forrest was involved in 46 out of 59 matches, only failing to make the match day squad once. He started the season as first choice on the right. Starting most games, he was injured in December for four matches, but on return was again mainly a starter. It was only the last five matches of the season that Forrest had a continuous run as a sub.

The on-loan Roberts would do well to consult Sinclair about being a young Englishman at Manchester City. He ended the imaginative 18-month long deal at the end of the season. He was never left out of the match day squad when available. Only completing 11 full 90 minutes, he was used off the bench at the beginning of the season with Forrest preferred and an early season injury stalling his momentum in August. Roberts was Celtic most used sub, appearing 18 times from the bench. Injuries to the strikers at the turn of the year resulted in him occasionally filling in a central role. By the end of the season he was back to being a first pick.

Although suffering a long injury absence, this season may have seen the swansong of the Mackay-Steven Celtic career. Never completing 90 minutes, he didn’t even make the squad on 23 occasions. Another strongly recommended read “Give Me Wings” illustrates the challenge facing Mackay-Steven competing with the current crop of wingers. A characteristically mixed performance at home to St Mirren in the Scottish Cup in March, giving the ball away 8 times from 18 attempted passes, saw him hooked at half time. Although 11 minutes action followed as a substitute against Hearts, I’d be surprised if we see him in first team action again.

The season saw a welcome debut for a Celtic youth player, Johnston, given a start in May against St Johnstone. It proved a highly successful debut with an assist. He also completed 2 dribbles and completed 22 passes from 24 attempts in his 57 minutes. After 14 goals and 8 assists for the development squad, I hope we see more of him in 2017/18.

Finally, I have added in Hayes appearance data with Aberdeen last season. Another who was always picked when available, he got through a lot of work racking up 3,823 minutes. It is perhaps coincidental that all four first choice wingers averaged around 7 matches lost to injury.

I don’t have further performance data for Hayes apart from goals and assists. Johnston and Mackay-Steven are excluded from further analysis due to insufficient data size.


Although not their core role, within Rodgers pressing system, all players are expected to challenge from the front. A reason Forrest is often preferred is that he is the defensively most sound of the wingers, by some distance. He wins the most challenges and actions the most intercepts, whilst losing the least number of challenges.

Given Sinclair provides the least defensive cover, but is a first pick for his scoring and pace, not to mention experience, in tight or difficult matches, playing Roberts and Sinclair is a risk. Roberts only started 4 of the 12 Champions League matches, lasting 90 minutes against only Lincoln Red Imps at home, FC Astana away, and Manchester City away, which is only polite being his home ground.

Roberts is more robust than his stature would suggest, his powerful upper body allowing his to dribble effectively and with balance, whilst withstanding challenges. His speed and control does allow him to win possession from interceptions at the highest rate (47%). Sinclair is taller but slighter in build and can be brushed off the ball quite easily. The more wiry and muscular Forrest wins nearly 1 more challenge per 90m than Sinclair.


Roberts led the squad with 18 assists at a rate better than 1 every 2 games. No one came close to his 1.17 chances created per 90m (for example Rogic is 0.78 and Griffiths the next highest on 0.92). Remarkably, Griffiths exceeds his assist rate with 0.66 per 90m. In his 18-month spell, Roberts finished with 20 assists for the club at the rate of 0.49 per 90m.

Beyond the numbers there were moment of beauty. The assist of the season was undoubtedly Roberts taking out nine Motherwell players with one pass, one minute after going 2-3 down, to set up Armstrong’s equalising goal in perhaps the most exciting match of the season.

Forrest completed his most productive season for the club and ranked second for assists on 16. He completed the most passes of the three and is most accurate. He ranks third in the squad in terms of assist rate but only Roberts attempted to create more chances per 90m.

Sinclair was involved in a lot less passes that the other two and is also the least accurate. Playing more a forward role than an out and out winger, his assist rate, chances created and attempted lags behind the other two. He ranks 5th in chances created per 90m and 5th in assists per 90m.

Hayes for Aberdeen recorded an impressive 24 assists in all competitions last season. This was the most productive season of his career and resulted in an assist per 90m rate of 0.56, behind only Griffiths and Roberts for Celtic. He recorded 12 in season 2015/16 (0.30 per 90m) and 14 in season 2014/15 (0.34 per 90m). Purely based on the numbers, last season was a stand out by Hayes and well above his norm. Was last season a rich run of form, and will he subsequently return to the less productive numbers shown throughout his career? Or, at 29, is he a late developer and will flourish further playing with better players in a more attacking team?

General Attacking

Roberts is the master of the dribble within the squad. He attempted 6.57 per 90m, the next nearest being Forrest at 4.25 per 90m. Noticeably, the more experienced Forrest and Sinclair both have better dribble success rates than the younger player. Forrest succeeds with 60% of his, and Sinclair is at 50% compared to Roberts at 44%. Forrest and Sinclair tend to rely on raw pace rather than trickery, hence are likely to attempt less such explosive actions. Roberts is also dispossessed more often which is not surprising as his running style engages defenders at close quarters, inviting challenges before deploying skill and quick feet. Sinclair and Forrest tend to blast past players with less engagement. Therefore, it is perhaps not surprising Roberts draws the more fouls, at 1.41 per 90m. Only Rogic draws more, at 1.8 per 90m.

Roberts is one of the most thrilling dribblers Celtic has seen. From the beginning of the season when he wriggled into the box and passed to Griffiths to score the goal that calmed nerves against Lincoln Red Imps, the season has been punctuated by productive brilliance. In the same game his run, 1-2 with McGregor and goal killed the tie. Showing tenacity, he chased down Tierney’s long diagonal ball in Astana to win a challenge and cut back for Griffiths to score that vital away goal with the only shot on target in that hot, distant match. After a slow start to the season following injury, Roberts exploded into life away to Ross County at the end of October, slaloming through challenges and ending with a trademark curling left foot finish after 3 minutes. Top and tailing that match, his drive down the wing allowed Dembele a tap in. Showing an impeccable sense of occasion, his 4th minutes hustle into the box and scoring shot in December reminded Manchester City what they may yet have to come. Roberts helped secure league invincibility beating his man and crossing or Griffiths to head the opener against Hearts in the last league game of the season in May.

Goal Scoring

Under Rodgers and Deila Celtic have overwhelmingly played with a single striker. Possessing a goal scoring number 10 in Rogic partly mitigates this. There is expectation, however, that the wide attackers also provide a goal threat. In total, the three regular wide forwards provided 38 non-penalty goals. Allied to Rogics’ scoring rate of 12 goals at 0.49 per 90m, Celtic do have considerable threat supporting the lone striker.

Sinclair had shown at Swansea City, under Rodgers, an ability to score. Before joining Celtic, 64% of Sinclair’s career goals had been under Rodgers in two seasons at Swansea City from 2010 to 2012. Now 75% of Sinclair’s career goals have been under Rodgers. Before this season, he had never reached higher than 6 goals other than those two seasons in Wales (scoring 27 in 2010/11 and 8 in 2011/12). Before joining Celtic, 17 of Sinclair’s 55 club goals had been penalties. Removing those from the calculations results in a pre-Celtic Non-Penalty Goals per 90m of 0.22. His overall scoring contribution (goals and assists) removing penalties is 0.31 per 90m.

Two seasons heavily skew Sinclair’s goal scoring record – 2011/12 in the English Championship with Swansea City and last season in Scottish football (he did not score in Europe). His scoring rate in those two seasons is 37 non-penalty goals in 8,180 minutes, giving him a Non-Penalty Goals per 90m of 0.41. Over the rest of his career, his scoring record is 21 non-penalty goals in 11,386 minutes – 0.16 Non-Penalty Goals per 90m.

Another relevant factor is that Sinclair had an xG of 14.09, removing penalties again. He out performed this with 19 goals although only 1 goal was from outside the box. Normally players will regress to the mean, that is, he would expect to have a scoring tally closer to his xG over time.

I have questioned above whether Hayes’ 2016/17 performance numbers were an upward trend or a one off and at 29 can he continue to improve. It is only fair, therefore, to question whether Sinclair, at 28, will maintain last season’s scoring form, or revert to the norm of the rest of his career. If Sinclair can score double figures next season it will be the first time in his career he has achieved that in consecutive seasons. Given he will be playing in the same team in the same division, and given the evidence of our eyes that he is a stand out at this level, Celtic fans can have justified hope in this regard.

Forrest, being a more traditional winger who stays wide and attacks the right as a right footed player (he is less likely to be cutting inside onto his strong foot like Sinclair and Roberts), had the lowest goal tally of 8. Nevertheless, this represents his second highest total, scoring 9 in 2011/12. Before this season, Forrest scored 0.21 NPG per 90m, compared to 0.24 this season. With an xG of 7.24 this was around what was expected. Forrest is quite a careful player in that he has a high pass completion rate compared to his peers, and his dribble success of 60% highlights good decision making around when to take players on. His shooting, however, is less accurate than peers, and 35 of his shots were from outside the box, resulting in 1 goal. For comparison Sinclair scored 1 from 34 and 3 from 27 outside the box. Bear in mind Forrest played a lot less minutes that Sinclair. Therefore, a good decision maker as regards passing and dribbling, he can continue to improve his shot selection.

Roberts has the highest goals per 90m rating behind Sinclair, with 0.37. He has the least shots per 90m of the three, and is only 1% less accurate than Sinclair. Whilst Sinclair predominantly plays inside left, and Forrest wide right, Roberts is the one of the three with most variety in his play as regards where he appears on the ball. Comfortable as a number 10, he was effective as a withdrawn striker as shown against Hearts in January, Aberdeen in February, Hearts again in April and the following two home games against Partick Thistle and Kilmarnock. Roberts leaves the club scoring 17 goals in 3,676 minutes. He averages 0.42 NPG per 90m over his Celtic career, the same as Sinclair this season. A scorer of spectacular goals, he is also an entertainer. His strike in the title clinching victory at Tynecastle was clinically final, possessing power, accuracy, and aesthetic beauty.

Roberts was another who out performed his xG of 6.8 with 11 goals including 3 from outside the box. Either Celtic have exceptional finishers across the squad or we must be prepared for the scoring rate to decrease somewhat next season!

Overall On-Target conversion rates (% of goals from on target shots) was:

Despite attempting 27 shots from outside the box, 9 of Roberts’ were on target for 3 goals. Furthermore, his overall conversion rate was considerably higher than the other two wingers.

Some context (top converters for on target shots in Europe’s top five leagues):

WARNING – I have looked at Squawka’s on line database and have been unable to reconcile these numbers and it is not clear to what extent penalties have been included.

Neither Sinclair nor Roberts would qualify for comparison on the basis that neither scored 20+ league goals (no penalties) nor play in one of the top five leagues. Nevertheless, Sinclair’s conversion rate of 37% is excellent. And Roberts, well, I’d love to know which players in Europe in any league with more than 10 goals had a better conversion rate.

Scoring Contribution

Scoring Contribution is the average number of non-penalty goals and assists per 90m. It is a useful measure of overall attacking productivity. Remember that penalties are excluded as, irrespective of who takes it, you have around a 78% chance of success. So adding them in provides nothing other than artificially inflating the numbers for penalty takers.

Over his Celtic career, Roberts has a Scoring Contribution per 90m of 0.9, scoring 17 goals and 20 assists in the equivalent of 40.8 matches. Dembele’s scoring contribution is 0.99 and Griffiths 1.36 last season. Replacing this productivity will be a huge challenge for the management team. After that, Roberts leads the attacking midfielders by quite some distance. Assuming Ciftci is not part of the Celtic future, the Number 7 shirt continues to gather dust!

Looking back over Forrest’s Celtic career it is remarkable the extent to which he has had to carry the burden of providing the attacking width. Since breaking into the team in 2010/11, the following players have been played in wide positions without making the position their own: Maloney, McCourt, McGinn, Samaras, Wakaso, Mackay-Steven and Armstrong. I have missed out many who have really been bit part (I’ll add Boeriggter’s name which will probably be the only time he ever appears in one of my articles). Commons has lined up as the team “7” or “11” on many occasions but no one would claim he is a winger. Lennon did not seem to value true width, whilst Deila set the team up for width but never managed to find a consistent solution to the wide areas. Therefore, since Forrest became a Celtic regular, this season has been the first he has had a strong cadre of peers with which to compliment and compete. Whether it is a coincidence, due to the manager, or due to increasing maturity and experience, this has been his most productive season:

There is a trend here of players having their most productive seasons. Forrest has been at the club longer than most and after his least productive season under Deila has really kicked on. Like Hayes and Sinclair, the challenge is to prove this wasn’t a one off and that he is redefining a new and higher “average”.

Hayes data is included above and illustrates what a stellar season he had, with 0.82 goals and assists per 90m he out performed Sinclair and Forrest. Whilst hugely encouraging, this remains an outlier compared to the rest of his career, save for a season in the Scottish Championship with Inverness Caley Thistle.

Sinclair is only 0.02 more productive than Forrest. As a comparator, in English Premier League football, he has played 6125 minutes for 13 goals – 0.19 goals per 90m. Adding in 6 English Premier League assists, his total scoring contribution in England’s top flight is 0.28 per 90m.

C:\Users\Alan\Documents\Football\Celtic Stats Analysis\Images\sinclair hattrick.JPG

Concerned at the fact his scoring contribution drops dramatically when playing at a higher level, I looked at the relative performances of the four wingers in European football.

Roberts overall scoring contribution holds up well, albeit either he nor Sinclair have many minutes at this level. Hayes numbers are again encouraging but mainly achieved against teams in the early rounds of Europa League qualifying rather than Champions Leagues. Forrest has this experience and his overall contribution in European matches tends to be at the lower end of his overall numbers. Sinclair’s contribution at this level reinforces slight concern, registering 1 assist in total.


Rodgers appears to have worked his motivational magic to raise the performance levels of Forrest and Sinclair well above their career norms. Intriguingly, Hayes displayed the same trend last season at Aberdeen and it will be fascinating to see if Rodgers can further improve him. With Mackay-Steven drifting out of the picture, and Roberts leaving, the wing positions will require further replenishment. Youngsters like Nesbitt, Miller and Johnston may get game time.

There is a gaping hole, however. I showed in “Give Me More Wings” that Roberts is one of the top talents in his position and age group in Europe today. I have rarely enjoyed watching any Celtic player more. Hugely entertaining as well as gifted, he makes great decisions and is prodigiously productive. I struggle to maintain objectivity when analysing him, and believe he may be one of the most talented players I will ever see at Celtic.

“A lament for Roberts”.

“What do they say about him?”

“I have not the heart to tell you. For me the grief is still too near.”

One thought on “Goodnight, Sweet Prince

Leave a response

Your email address will not be published. Required fields are marked *