Defence or Attack - which is most important

[Im_Spartacus]

Really this follows on from the Chelsea thread, as the issue was raised about how much of an impact a good defence has on a title challenge. So being a boring cunt, I decided to look at it from a scientific point of view at the start of this thread to potentially add some data into the mix before we all head off on a big argument about whether attack or defence is the biggest predictor of a title win.

I took the premier league champions total points, goals scored, goals against and goal difference from each of the last 10 years, including the projected finish for chelsea this year.

I then ran regression analysis against to compare total number of points to GF, GA and GD to see which ones had a verifiable correlation with the number of points at the end of the season, and the results were far more clear cut than I expected.

Goals for has a 37.7% correlation with the number of points
Goals against has a 65% correlation with the number of points
Goal difference had no correlation (0.8%) with the number of points

In summary, over a 10 year period, defence has been almost twice as important as attack in terms of predicting which teams win the title.

So with this in mind, it makes both Mancini's and Mourinho's focus on not conceding somewhat more logical, given that I'm fairly sure I'm not the first one to do this and find that a solid defence has a greater bearing on the league title than a prolific attack.

So how would people prefer us to play in light of these trends? I certainly don't advocate the Mourinho style, but does this open up tactical questions about losing Yaya and his goals in favour of a less porous midfield?

[bayblue]

Really this follows on from the Chelsea thread, as the issue was raised about how much of an impact a good defence has on a title challenge. So being a boring cunt, I decided to look at it from a scientific point of view at the start of this thread to potentially add some data into the mix before we all head off on a big argument about whether attack or defence is the biggest predictor of a title win.

I took the premier league champions total points, goals scored, goals against and goal difference from each of the last 10 years, including the projected finish for chelsea this year.

I then ran regression analysis against to compare total number of points to GF, GA and GD to see which ones had a verifiable correlation with the number of points at the end of the season, and the results were far more clear cut than I expected.

Goals for has a 37.7% correlation with the number of points
Goals against has a 65% correlation with the number of points
Goal difference had no correlation (0.8%) with the number of points

In summary, over a 10 year period, defence has been almost twice as important as attack in terms of predicting which teams win the title.

So with this in mind, it makes both Mancini's and Mourinho's focus on not conceding somewhat more logical, given that I'm fairly sure I'm not the first one to do this and find that a solid defence has a greater bearing on the league title than a prolific attack.

So how would people prefer us to play in light of these trends? I certainly don't advocate the Mourinho style, but does this open up tactical questions about losing Yaya and his goals in favour of a less porous midfield?

Wow. Interesting stuff !
Wonder whether it's changed much year by year? Especially last season?
As you say it explains Maureen's approach but masochist that I am I still prefer us to go for it.
At their very best this year Chelsea often had Matic as the only holding mid. With a player like him next to him I think Yaya would still do ok.
Not sure he's going to be around anyway next season so comes back to the question of how we are going to compensate for him.

[Mikhail Chigorin]

Interesting analysis there Sparty; thanks for posting it and it's well worth pondering.

However, in spite of all such noteworthy evidence, as far as I'm concerned there's only one way to play and it's all-out attack and blowing the opposition away with a deluge of goals whenever we can.

If this means our defence is overly exposed and we concede more, over a full season, by comparison with our competitors, then so be it. However, by playing in such a way, I do believe that our back line is under less pressure by virtue of the goals we actually score and, strangely, this factor might be said to fit into the statistical evidence Sparty has so thoughtfully drawn up.

It's goals that win matches and, for me, it's a case of the more the merrier.

[Clowncrete]

A bit of both. It makes sense to take a cautious approach against teams like Bayern. But It's idiotic to put in 3 defensive players in midfield at home to Crystal Palace and grind out 1-0.

[Im_Spartacus]

A bit of both. It makes sense to take a cautious approach against teams like Bayern. But It's idiotic to put in 3 defensive players in midfield at home to Crystal Palace and grind out 1-0.

I agree, you have to select based on the opposition, but I suppose the question is about the strategy on which the squad is built rather than the day to day choices of who plays in what position in a particular game.

[nottsblue]

Surely it is a mixture of the two. One thing Mourinho and Bacon before him for that matter, both did, was try to ensure they did not lose against teams also chasing the league. Often these games would be dull affairs, settled by an odd goal from a set piece. The flair and panache would then come out against lesser sides. Score the goals to improve the GD then.

Invariably the league champions either score the most goals or are thereabouts. They also invariably concede either the least or thereabouts. Scoring goals in every game in perhaps more important than getting cricket scores against relegation fodder however. Score in every game and you will then win more games. Not conceding is important but not conceding only wins a point. Scoring a goal potentially wins three, assuming a clean sheet is kept. All out attack will win games but the dippers last year are a prime example of that not being enough.

A better tactical manager than Rodgers and we may not have won the league. Might Mourinho just simply be playing his hand to the best of his ability. He knows he can't go toe to toe with us or arse or even dippers and rags so he plays for a draw with the hope of a lucky win. To be fair a lot of us have questioned Pellegrini's tactics on occasion, especially away from home and in games close to Champions League games. Maybe, with hindsight he would have done things differently and if just three of those defeats had been wins, it would have been a different run in. Still, it wasn't and we have to hope lessons have been learned by the management and players alike for next season.

That said, a few more goals wouldn't go amiss

[DoomMerchant]

the old agage in American sports is basically "Defense wins Championships."

It's anomalous when a team that rocks out with their cocks out wins the "league/title/cup/whatever the fuck it is"...

Most teams strive for a balance, but the most of the most successful coaches and teams have been build around defending first, and layering in an attacking mindset later.

Even my NBA team, the Golden State Warriors, who are the sickest offensive threat in the league with the most possessions and highest scoring average, are based upon a solid, solid set of defending that everyone focuses upon. Think of them as the high-pressing Barca, at their best, of the NBA maybe. They turn defense into attack and are so fast in transition that you are fucked. And they shoot the lights out. Their shooting percentages are outrageous.

Teams that do what we did this year -- an absolute arrogance of attacking activity without a focus on winning the ball back when you don't have it -- don't ever win jack shit in any sport. The desire we showed absolutely merits on end results this season, sadly.

Mancini had the code and was perhaps close to a breaththrough, but he was such an Italian dictator that he threatened to break the club, and he was also full of shit about any actual support he had for the academy. It was a place to be nepotistic with his kids and their friends. I loved last season under Pellers but i'm not sure the Count can recreate it. Defensive hangover from Mancini coupled with his attacking philosophy created one for the ages.

And i'd prefer that he or another fucker keeps trying for that ambitious approach rather than getting more Mourinho like.

I want us to win in spite of the perception of "how it's supposed to be done."

Am i patient enough for that? Fuck knows. This year was trying but i'm slowing losing some anger now that i know we can't impact the results.

cheers

[zuricity]

Hmmm. Lies , damned lies and Statistics..

I think the game is won and lost in midfield.

[Ted Hughes]

So if you are 4 up in the first half. How many do you let in ?

[iwasthere2012]

Hmmm. Lies , damned lies and Statistics..

I think the game is won and lost in midfield.

And I agree, but for the sake of the argument, I would say that a winning team is built on a solid defense. For me though the balance of the overall team, the workrate of the midfield in particular and stability of the CB's partnership all add up to less goals being let in. I think like chess, if you take control of the centre of the field your chances of getting a result go up considerably.

[Blue Since 76]

There's the old saying that goals/strikers win matches but defences win titles and those stats back it up. Whilst the goal difference matters, scoring lots can just be false vanity - if last season we'd only beaten Spurs, Arsenal and Norwich by one goal, would we have still been champions?

In many ways, the only GD that matters is plus one in every game. If you can keep a clean sheet, that's not too hard to achieve. If you can score 2.5 a game, there's some room at the back for defensive lapses, but the reality is to win the league, you have to be good at both ends over the 38 games. Our defence this season is very close to last (1 goal better off with 3 to play), but we've scored a lot less, probably 15 below normal for the champions - ignoring last season, in reverse, the winners have scored 86, 93, 78. Battering QPR by 6 or 7 on Sunday may mask that, but you need to be scoring in virtually every game, not just in bunches.

So the Mourinho approach is the mist effective, but not one I'd want to watch every week. We need to learn to toughen up away from home and maybe bore the life out of the opposition and be a bit more pragmatic at times. Against Spurs we started to show that, with a more solid looking line up and a willingness to go more defensive to close it out. Take that into next season and add some goals from someone other than Aguero and we'll be a lot closer, but this season from Chelsea will be the benchmark we have to beat

[zuricity]

I think most on here would agree that Hazard has played brilliantly for Chelsea this season. Now, i would suggest that if Hazard wasn't playing for Chelsea, they would not have been so successful ( grammar ok so far Slim ?) .

Hazard is a midfield player.

Furthermore, and you will like this Carl.

Vince was a midfield player that really could not hack it in midfield. So he turned to defense. Where ... surprisingly enough, he's been found out.

As the constipated Statistician, or mathematician might say , work that one out ( with logs ).

[Beefymcfc]

Score more that your opponent or concede less, your answer is right there.

[Mikhail Chigorin]

Hmmm. Lies , damned lies and Statistics..

I think the game is won and lost in midfield.

And I agree, but for the sake of the argument, I would say that a winning team is built on a solid defense. For me though the balance of the overall team, the workrate of the midfield in particular and stability of the CB's partnership all add up to less goals being let in. I think like chess, if you take control of the centre of the field your chances of getting a result go up considerably.

Interesting comparison.

There's a saying in the world of the sixty-four squares that "he who controls the centre, controls the game". However, this applies more to those positions which are closed or semi-closed and where the need to manoeuvre is the requirement.

If, however, you're a Gambit player and go for all out attack then, in wide open positions, there may not even be a centre as such and the rapid mobility of pieces is then of paramount importance.

If this analogy is applied to football, in the loosest and vaguest sense, then it could be argued that a strong midfield is vital in those games where either a side is 'parking the bus' against you, or else a dogged midfield battle is going to be fought. In such instances it might be reckoned that you need to dominate the midfield in order to have freedom of manoeuvre.

In an open game, it might be argued that speed and mobility (and flank play) are then vital and, in such cases, some sides might even, conceivably, by-pass the midfield altogether and use longer passes to achieve desired results.

Anyway, having developed this theme to this extent, I'm now convinced I'm more than likely talking out of the back of my head and, if not completely, then almost.

Just goes to show the futility and virtual impossibility of comparing chess with football........

[Im_Spartacus]

Invariably the league champions either score the most goals or are thereabouts. They also invariably concede either the least or thereabouts. Scoring goals in every game in perhaps more important than getting cricket scores against relegation fodder however. Score in every game and you will then win more games. Not conceding is important but not conceding only wins a point. Scoring a goal potentially wins three, assuming a clean sheet is kept. All out attack will win games but the dippers last year are a prime example of that not being enough.

Your points are perfectly valid, and very persuasive, but this is the sort of observational evidence which despite perceptions (mine included), is factually incorrect.

If you took last season's league table, and did the same regression analysis for all teams, you would see that the number of points for each team is almost perfectly correlated with both goals scored and conceded (89% vs 83%). The correlation is so strong in fact, you would assume this to be true if comparing each team finishing 1st over time - and this is where our belief system about how football works, starts to unravel.

Throughout the entire league in any given season, the importance of goals scored and conceded is roughly equal. But when you look only at the teams who finish first, something different happens. You find that the further up the league you go, their points total has a significantly higher correlation to goals conceded, than it does to goals scored which means conceding less is a clearer predictor of a champion, than scoring loads.

Of course the opposite is true with lower placed teams, at the bottom of the league, its not about how many they concede (it doesn't matter whether you lose 1.0 or 5.0, you get no points), they simply don't score enough which becomes the biggest predictor of their points tally.

I know it's counter-intuitive, but a lot of things which we simply accept as a given in football are wrong when you run the numbers or even just apply conventional logic.

[Im_Spartacus]

Hmmm. Lies , damned lies and Statistics..

I think the game is won and lost in midfield.

I agree for what it's worth.

The numbers could easily mean that how the midfield interacts with and protects the defence, has a bigger impact on our points tally than Silva's magic.

If that's true should we all be quite so critical when we see the side lining up with two DMs? (assuming one of them doesn't abdicate defensive responsibility)

[Ted Hughes]

There's the old saying that goals/strikers win matches but defences win titles and those stats back it up. Whilst the goal difference matters, scoring lots can just be false vanity - if last season we'd only beaten Spurs, Arsenal and Norwich by one goal, would we have still been champions?

In many ways, the only GD that matters is plus one in every game. If you can keep a clean sheet, that's not too hard to achieve. If you can score 2.5 a game, there's some room at the back for defensive lapses, but the reality is to win the league, you have to be good at both ends over the 38 games. Our defence this season is very close to last (1 goal better off with 3 to play), but we've scored a lot less, probably 15 below normal for the champions - ignoring last season, in reverse, the winners have scored 86, 93, 78. Battering QPR by 6 or 7 on Sunday may mask that, but you need to be scoring in virtually every game, not just in bunches.

So the Mourinho approach is the mist effective, but not one I'd want to watch every week. We need to learn to toughen up away from home and maybe bore the life out of the opposition and be a bit more pragmatic at times. Against Spurs we started to show that, with a more solid looking line up and a willingness to go more defensive to close it out. Take that into next season and add some goals from someone other than Aguero and we'll be a lot closer, but this season from Chelsea will be the benchmark we have to beat

This does absolutely not prove that the Mourinho approach is the most effective. Of course it is better to have an effective defence but does anyone truly belieive Mourinho's Real Madrid would be looking for a draw vs the rags ? Or if he was current manager of Barca, they would be reliant on his style to make them competitive in the Champions League ?

The absolute truth, is that it's possible for people like Mourinho to mug their way to titles & trophies, some of the time.

For the most part, the teams who win, most of the time, will be the ones Pellegrini describes as playing like ' big ' teams. The main point is; big teams need big players. We are working towards that. Mourinho is not needed once a club achieves it

[zuricity]

Hmmm. Lies , damned lies and Statistics..

I think the game is won and lost in midfield.

I agree for what it's worth.

The numbers could easily mean that how the midfield interacts with and protects the defence, has a bigger impact on our points tally than Silva's magic.

If that's true should we all be quite so critical when we see the side lining up with two DMs? (assuming one of them doesn't abdicate defensive responsibility)

Au Contraire... Having two defensive midfielders means the manager doesn't trust the defense.

There are way too many variables that influence a football match. I need someone to convince me that michael oliver knows anything about football.

[Foreverinbluedreams]

Defence may be more important in a results context but attack is more important for the game in general.

If statisticians ran the game then there'd be little room for a cavalier approach, for mavericks, for flair - all things that make football special.

I also think there's a lot of truth in the Clichy 'attack is the best form of defence'. If you look at the three best teams in Europe at the moment they all have an attacking philosophy and in my opinion not great defences.

[Blue Since 76]

There's the old saying that goals/strikers win matches but defences win titles and those stats back it up. Whilst the goal difference matters, scoring lots can just be false vanity - if last season we'd only beaten Spurs, Arsenal and Norwich by one goal, would we have still been champions?

In many ways, the only GD that matters is plus one in every game. If you can keep a clean sheet, that's not too hard to achieve. If you can score 2.5 a game, there's some room at the back for defensive lapses, but the reality is to win the league, you have to be good at both ends over the 38 games. Our defence this season is very close to last (1 goal better off with 3 to play), but we've scored a lot less, probably 15 below normal for the champions - ignoring last season, in reverse, the winners have scored 86, 93, 78. Battering QPR by 6 or 7 on Sunday may mask that, but you need to be scoring in virtually every game, not just in bunches.

So the Mourinho approach is the mist effective, but not one I'd want to watch every week. We need to learn to toughen up away from home and maybe bore the life out of the opposition and be a bit more pragmatic at times. Against Spurs we started to show that, with a more solid looking line up and a willingness to go more defensive to close it out. Take that into next season and add some goals from someone other than Aguero and we'll be a lot closer, but this season from Chelsea will be the benchmark we have to beat

This does absolutely not prove that the Mourinho approach is the most effective. Of course it is better to have an effective defence but does anyone truly belieive Mourinho's Real Madrid would be looking for a draw vs the rags ? Or if he was current manager of Barca, they would be reliant on his style to make them competitive in the Champions League ?

The absolute truth, is that it's possible for people like Mourinho to mug their way to titles & trophies, some of the time.

For the most part, the teams who win, most of the time, will be the ones Pellegrini describes as playing like ' big ' teams. The main point is; big teams need big players. We are working towards that. Mourinho is not needed once a club achieves it

I hate Mourinho and his style of play with a passion. I hated it first time round when he bored his way past us when Pearce was in charge. But you can't make out that he's a chancer who's lucking his way to a few trophies.

Does he build for the future? No. Does he make you popular amongst neutrals? No. Does he win stuff? Absolutely, in every country he's been in. In terms of results, probably only Guardiola of all the current managers has an equivalent record.