Bristol City: Key Play Analysis Project 2007/08

[Red Zeppelin]

[Red Zeppelin]

[Red Zeppelin]

2. Attack Analysis

2.1 Overall attacking contribution %

We begin with a very broad illustration of our season’s attacking play. This is simply how frequently players feature in all 1248 passes recorded.

As we can see, Lee Johnson and Michael McIndoe between them made up over a quarter of our attacking play. I emphasise that this is a very broad window at this stage – a Lee Johnson contribution could be a simple lay-off deep in our half to Orr who launches the ball forward for a Trundle shot, and Johnson would therefore go down as the third man in that move. Equally, a Johnson contribution could also be a good thirty yard shot saved by the diving goalkeeper. However, over 1,248 contributions these will average out.

2.2 Final Two per game

Moving from our broad opening, each attack recorded breaks down into two categories. Firstly, the ‘build-up play’, which excludes the final shot and records simply all the players involved in the passing move leading up to the shot. This shall be looked at in the next section. Secondly, each attacking move can be looked simply in terms of the two players with the shot and assist (final two of the move).

The following diagram illustrates exactly this:

We can see that Lee Johnson drops down the list for end product as he is now the fourth most significant player. McIndoe heads the way as we’ll find he often does in this project, but also Lee Trundle, as expected for a striker, features highly in assisting shots and taking shots himself. Interestingly Marvin Elliott, arguably a defensive player, also ranks highly in this analysis, and it is apparent he was often in and around the box to either have a shot or set someone up.

It is worth noting that this diagram does not account for games played, hence the late additions to our squad, Adebola and Carle, are both placed low in their contributions to the attacking play over the whole season. In the next section we cover players who were at the centre of our attacks as we look at purely the build-up play.

2.3 Build-up play

This section looks at the individuals who were at the heart of Bristol City’s attacking play; who featured most frequently in moves with an end product.

Here we can see again how McIndoe was intrinsic to most moves, with Lee Johnson our other protagonist in respect to ‘building’ productive moves. Of the 754 passes in build-up play, 117 of them were from Johnson alone.

This is still a season-long analysis, and in the following segment we begin to analyse players’ performances on a ‘per game’ basis.

2.4 Overall Creativity Index

We now move into a more detailed look at the contribution of players. The reason for this is that some players such as David Noble made 16 starts whereas Lee Johnson started 39 games. This needs to be accounted for in the statistics. Therefore I began to use the ‘start and substitute’ appearances for players.

All substitute appearances were assumed to be 0.3 of a game, so for instance the 20 substitute appearances made by Cole Skuse equated to six starts (20 x 0.3).

Moreover, in order to improve accuracy further for certain players, I manually worked out the actual number of minutes played over the season. For example, Noble played roughly 1384 minutes, and McIndoe played around 4349 minutes. As a result, we can now work out on average how much a player contributed to attacking moves in every 90 minutes of football they play. The following diagram shows this creativity index for certain players:

David Noble is our most creative midfielder as could be expected in his ‘in-the-whole’ position occupied in most of his appearances. Our main strikers Trundle and Adebola occupy the other positions in the top 3 creative players which is once again normal as strikers. However, it is probably most notable that Lee Johnson contributes more to productive moves than Nick Carle who occupied his position towards the end of the season.

Now into the more revealing statistics, we now look at the breakdown of this creativity index into ‘final two’ and ‘build-up play’, as in the previous sections.

2.5 Creativity Index, ‘Final Two’

This shows the average contribution to assists or final shots per game played.

Surprisingly the forgotten man, Scott Murray contributed quite a lot in making and taking chances during his early season appearances.

It would appear that Carle played further up the pitch in attacks than Johnson as he was more frequently part of the end product rather than the build-up play in the next section. Moreover Carle was there even more than Noble who was supposed to be playing behind the striker in most of his games.

Next we look at the creativity index in terms of ‘build-up play’.

2.6 Creativity Index, ‘Build-up Play’

Once again, Lee Johnson is up there with Noble as the main instigators of build-up play. Johnson is part of build-up play almost twice as much as Nick Carle. Adebola and Trundle are up there as strikers; however Darren Byfield contributed less to build-up play of key moves in his appearances than right-back Bradley Orr.

[Red Zeppelin]

[Red Zeppelin]

4: Corners

4.1: Receiver of most corners

Overall in the highlights of our games, there were 38 corners which resulted in an effort on goal. Of these 38, Marvin Elliott ended up with 13 of them alone, equating to 34.2% of the corners.

[Red Zeppelin]

[Red Zeppelin]

[Robbored]

Impressve stuff and with limited data - you have been busy!

I guess what you have come up with is a sample version of what is available to Johnson and the coaching staff thro' prozone. Obviously prozone is able to be far more precise over a full game.

Its an intersting read and how would you feel about producing something similar at the end of each 10 game batch? I'm sure most of us would find the result very interesting.

[Johners26]

Excellent nerding. I think you have some of the attacking graphs in the wrong place though (build up play vs 'final two' when accounting for number of games played). Minor detail, but I mention it in case anybody else dets confused.

[Red Zeppelin]

Excellent nerding. I think you have some of the attacking graphs in the wrong place though (build up play vs 'final two' when accounting for number of games played). Minor detail, but I mention it in case anybody else dets confused.

Thanks for pointing that out. But the edit function seems to have gone now? TomF, anyone?

[Johners26]

Excellent nerding. I think you have some of the attacking graphs in the wrong place though (build up play vs 'final two' when accounting for number of games played). Minor detail, but I mention it in case anybody else gets confused.

[kristian444]

Bloomin well done! Interesting to see Carle's position.

[Red Zeppelin]

Excellent nerding. I think you have some of the attacking graphs in the wrong place though (build up play vs 'final two' when accounting for number of games played). Minor detail, but I mention it in case anybody else gets confused.

Corrected now, thanks to TomF.

[Red Zeppelin]

Its an intersting read and how would you feel about producing something similar at the end of each 10 game batch? I'm sure most of us would find the result very interesting.

No problem, I'll work on it.

[iWurzel]

What's interesting about this is that this shows Jamie McCombe as by some distance our best defender. Anyone surprised at this? He wouldn't have been my first choice for defender of the season...

[Red Zeppelin]

What's interesting about this is that this shows Jamie McCombe as by some distance our best defender. Anyone surprised at this? He wouldn't have been my first choice for defender of the season...

I was also surprised by this.

I perceived him little as more than a slow, lanky but towering defender who is brilliant at heading the ball clear, but not as good as our other centre-backs at all round defending. But it became clear to me that he does block a lot of shots etc and is a lot more than just a first-ball centre-half. Just goes to show...

[Red Zeppelin]

He gets more than his share of blocks in - but so does McAllister to be fair, maybe that hasn't come out in your analysis but he always seems to get in the way of crosses and shots.

Great analysis by the way.

True, McAllister also gets a lot of blocks in. Hence on the defensive 'Blocks' diagram in (3.2) it shows McCombe had 19% and McAllister had 18%.

Basically whereas we knew that McAllister made many blocks last season, I meant that Jamie McCombe made more blocks than I had imagined.

[stick]

blimey mate....impressive post !!!

must have taken you weeks !

good idea about repeating this every 10 games or so

keep up the good work (and don't forget to go out sometimes as well !!!!)

top post

[Antman]

impressive stuff, I agree with the 10 game idea that would be brilliant.

does your data show passes completed and passes misplaced per player?

a player that makes a lot of passes may not be making the right ones! (and no, this isn't aimed at anyone)

also shots on and off target

just a thought?

[Red Zeppelin]

impressive stuff, I agree with the 10 game idea that would be brilliant.

does your data show passes completed and passes misplaced per player?

a player that makes a lot of passes may not be making the right ones! (and no, this isn't aimed at anyone)

also shots on and off target

just a thought?

Pass completion - I believe that the whole 90 minutes would need to be analysed for this. Besides, the highlights usually show the beginning of moves and hence you can't see which player conceded their side's possession.

Shots on and off target - Likewise, it is best if the whole 90 minutes is analysed because highlights don't show scuffed shots, or shots that go into row Z. However, the statistics for these are already out there, and I'll reel them into this thread soon.

I'm currently looking at the 10-game analysis for last season, and it is interesting to see the hard evidence of players' form dropping and rising at various intervals. More to come.

[City Rocker]

An extraordinary piece of work, well done! Some surprising and some not so surprising revelations there, with both Trundle and McCombe showing up as less obvious high-flyers. You also seem to have proved what most - but significantly not all - of us already knew about Lee Johnson's contribution to our season.

Although speaking as a mathematical neanderthal (I am to the science of statistics what Chris Coleman is to the art of grace-in-defeat), I would nonetheless calculate that by far the most telling data you produced must be those of your bar charts based on per-game/minutes spent on the pitch? May I therefore please request that you kindly concentrate on on these criteria when producing your ten-weekly updates? If it's not too much trouble? Thanks

[Ian M]

The results go to show how much we missed Jamie McCombe at Wembley. I too was guilty of underestimating his importance to the team and we'll obviously miss him greatly over the next 6 weeks.

Excellent work and I share the interest in seeing this season's figures.

[megansdad]

Very well done - thoughtful analysis and support for those of us who value lj's contribution

[Maesknoll Red]

Interesting read, I will also be interested to see this seasons ten game analysis. I guess a project like this is what you call a labour of love.

[gizmo666]

Great stuff. Look forward to the 10 game updates (hmmm, lots of work here. Are you up for it? I hear the pay is pretty bad...)

Thanks for shuting up the L Johnston moaners... (ok, nothing will shut them up but at least some stats for them to mull over)

[stroudy]

Great Post top job fella!

[Red Zeppelin]

EXTRA

Creativity Index: 10-game blocks

This graph shows the creativity index of players every block of ten games last season.

It's very important to add that the anomaly of Noble's peak in form is due to the fact that in the '4th block' of ten games, he played just 76 minutes against Watford (H) in which he made 6 contributions i.e. that point in on the curve was his performance in just one game. I opted against removing Noble altogether, for the sake of comparison. However, his curve plummets not because of a drop in form, but because he went from 6 in one game (Watford) to an average of 3.15 in his next 7 appearances, which is comparatively pretty good at the end. In the same way, Ivan Sproule appears decent at the end of the season because he played 1.70 matches (from the bench largely) and made 6 contributions in that time, giving his creativity index as 3.52.

Nevertheless, apart from Noble and Sproule towards the end of the season, all other figures give a pretty accurate representation

Lots and lots to say about this one!

For a start, we can see how Johnson's influence plummeted when he was injured, and Trundle's resurgence in form towards the end of the season is evident.

[cityfan]

Good work.

For me this just re-enforces that Trundle should be playing

[samo]

Awesome, thank you very much Red Zep!