Armstrong has been a standout for Celtic over the last few games. This has coincided with his move to a more central orthodox M(idfield)C(entre) position after being mainly used as an attacking wide midfielder. What does the performance data say about his contributions in different positions and will his current form last?
Positional Changes – A Summary
- 16 different position combinations in 79 appearances;
- 62 matches played in 1 position during time on pitch;
- Appeared in 6 positions – D(efender)R(ight), MC, M(idfield)L(eft), A(ttacking)M(idfield)C(entre), AMR(ight), AML(eft);
- Starting positions:
- AML 44
- MC 19
- AMC 10
- ML 4
- AMR 2
- Only played 90m in consecutive games on 3 occasions in Celtic career – until 26/10/16 and now played full 90m for last 10 appearances in a row;
- Played 90m 27 times (out of 79);
- For the purposes of the analysis aggregating appearances stats for AML and ML and AMR as such small samples but similar roles.
So, I am comparing Armstrong performances as orthodox MC; as Number 10 and as left sided wide attacker.
Analysis
To start some words of warning regarding the data (which is all below):
- The AMC stats are based on a very smaller sample – equivalent to 6 games.
- Many stats (passing, possession effectiveness, tackling, defensive effectiveness) are usually much lower for attacking players than defensive players given forward players are under more pressure, have less time on the ball and have less of the field to move into. Therefore higher / lower numbers for the MC position must be seen in that light. At some point, I will do a comparison between players in the same position.
- Finally, Armstrong moves position during games more than most players – therefore the stats reflect starting position/position played for most minutes and therefore include numbers whilst playing in a different position.
Armstrong has more touches, more possessions (83 possession events per 90 mins) significantly better passing stats, better tackling and more defensive actions as MC. He has less dribbles as would be expected compared to a wide attacker. As mentioned above, all that could be because of playing as MC as opposed to playing better because of being a MC – i.e. I’d expect a MC to have better stats in those areas compared to an attacking wide player.
However, the real bonus from playing Armstrong at MC is he is getting more shots and more goals, more assists and more chances created in that position. Despite playing in a deeper role with more defensive responsibility, Armstrong is more productive in an attacking sense in the central role. He is averaging nearly 3 shots per game and 0.32 goals per game which is 1 goal in 3 games. He is shooting from better positions needing 9.2 shots per goal. The passing and tackling stats are significantly (10%) better in the MC role. This suggests Armstrong is a player who needs a fraction second longer to decide next action. Also, playing deeper with the game in front of you does suit some players.
His Possession Effectiveness Indicator (PEI) is 10% higher than when playing as a winger – this means that across all his on-field actions, he keeps possession 10% more of the time.
A caution is number of defensive errors has increased and these have largely been due to misplaced passes and being caught in possession, 2 traits of Armstrong’s game in all positions. When playing MC giving the ball away in a central area either side of the halfway line, or being caught in possession will often lead to an opposition chance due to the possibility of a central transition opportunity. Basically, the opposition can break onto a disorganised defence from a position on the pitch to go left, right or central. Celtic are prone to this due to the high position the full backs take.
Armstrong is not going to be the “Number 1” in any position and if you compare just passing stats to illustrate for 2016/17 (passing accounts for 50% of all on field actions):
- MC passing exemplar – Brown 94%, Bitton 92% (Armstrong 84%)
- AML passing exemplar – Sinclair 81% (Armstrong 74%)
- AMC passing exemplar – Rogic 83% (Armstrong 80%)
But he has found a niche as the “box to box” player in the squad. McGregor can play this role but is not as defensively solid nor as physically strong.
One of my criticisms of the Deila team, and Rodgers has settled on the same 4-2-3-1, is that there is no need to have 2 sitting defensive midfielders against any domestic opposition. Rodgers has settled on Brown being the deep lying MC and the 2nd MC has license to get forward and join the forward rotations. Armstrong has so far proven the most adept in this role.
Conclusions
If you play a player regularly, give him minutes on the park in his favoured position, you will get better results than if you don’t. That is basic common sense management.
Secondly, Armstrong offers an attacking option allied to MC solidity and responsibility that no one else in the squad can currently offer.
He has had a great run, but Brown has been at the coal face for Celtic at this level for 9 years. That is the standard and consistency Armstrong will need to achieve to maintain his position in this team long term.
The Data – Summary
| KPI | MC | AMC | AML |
| Minutes | 1390 | 535 | 3067 |
| Effective Full Games | 15.44 | 5.94 | 34.07 |
| Goals | 5 | 3 | 6 |
| Assists | 5 | 1 | 10 |
| YC | 2 | 3 | 2 |
| RC | 0 | 0 | 0 |
| PEI | 76% | 71% | 66% |
| Poss Events per 90 | 83 | 71 | 66 |
Attacking
| KPI | MC | AMC | AML |
| Passes Complete | 855 | 259 | 1249 |
| Passes Incomplete | 160 | 64 | 444 |
| % Passing | 84% | 80% | 74% |
| Passes Complete per 90 | 55 | 44 | 37 |
| Shots On Target | 11 | 8 | 22 |
| Shots Off Target | 35 | 3 | 51 |
| Shots per 90 | 2.98 | 1.85 | 2.14 |
| Goals per 90 | 0.32 | 0.5 | 0.18 |
| Shots per goal | 9.2 | 3.67 | 12.17 |
| Chances Created | 11 | 3 | 23 |
| Chances Attempted | 17 | 8 | 42 |
| Chances Created per 90 | 0.71 | 0.5 | 0.68 |
| Assists per 90 | 0.32 | 0.17 | 0.29 |
| Fouls Suffered | 19 | 11 | 36 |
| Fouls Suffered per 90 | 1.23 | 1.85 | 1.06 |
| Dribbles Complete | 10 | 9 | 35 |
| Dribbles Incomplete | 18 | 8 | 41 |
| Dribbles Comp per 90 | 0.67 | 1.52 | 1.03 |
Defending
| KPI | MC | AMC | AML |
| Tackles Won | 71 | 19 | 103 |
| Tackles Lost | 62 | 27 | 173 |
| Tackles Won Possession | 69 | 18 | 135 |
| Tackles Lost Possession | 64 | 28 | 141 |
| Tackle Won % | 53% | 41% | 37% |
| Tackles Won per 90 | 4.6 | 3.19 | 3.02 |
| Clearances per 90 | 0.84 | 0.67 | 0.41 |
| Intercepts per 90 | 2.4 | 1.68 | 2.05 |
| Defensive Saves | 0 | 0 | 0 |
| Defensive Errors | 8 | 0 | 8 |
| Def Errors per 90 | 0.52 | 0 | 0.23 |
| Fouls Committed | 21 | 7 | 39 |
| Fouls per 90 | 1.36 | 1.18 | 1.14 |