Behind The Parker Stats: A Theory of AVB

Hey all, A1 here. Quickly becoming the site's Parker-apologist, after having slated him in the early season...


So MCofA recently did an analysis of the Parkerbele pivot and a number of users registered surprise that it outperformed the non-Parkerbele pivot systems when they were tried.

I'd like to firstly point out that I don't think the difference is quite as significant as the stats might imply because the "Other" subset will have suffered from a lack of Dembele. There are doubtless numerous other confounds. That said, the Parkerbele seemed statistically comparable with the Sandrobele or Moosebeast pivot. Why could this possibly be the case? Was it anomalous?

Damning counterevidence

Brett Rainbow took a further look at the stats and found that our goal difference (so not just goals scored - the balance of goals scored versus conceded) with Parker on the field came out at 0, and our goal difference when he was subsequently removed (presumably for Thudd/Holtby/Siggy/Carroll etc.) was +15. This remarkable statistic seems completely at odds with the findings from MCofA's power analysis and with the notion that Scott Parker should still be playing professional football...


However, I'd like to explain how both MCofA and Brett Rainbow can be right, and hopefully explain why AVB isn't totally crazy to keep playing someone who so obviously should be substituted in every game.

To do this, I'm going to use some hand made graphs.



They'll look like this. "Performance" will be used to denote the performance of the team in quite a fluid way, and you'll have to forgive that if this is going to make my point at all. The broad interpretation and use of graphs is really only trying to make conceptual points easier to see, and I fully admit that the simplistic crayon view of performance is not one which stands up to much scrutiny; there will be a trillion confounds and it may not even match up to actual stat data measuring performance as a result. The principles though, I hope, do stand up to scrutiny. So, I guess I'm saying... suspend your skepticism for a bit.


So here I'm representing basically performance in a vacuum and as a function only of energy levels. Again, this isn't the only factor obviously, but because fitness is quite important to the point I'm making, I hope you can allow me to run with it. Bit of a boost at half time, otherwise a gradual tailing off. Improved stamina will shallow the gradient and start you from a higher position. This will be my basic slope.


Another major factor influencing performance is the attacking impetus. We've speculated about teams with nothing to play for (affecting the whole game performance; lowering the entire graph) but there's also a big question of mentality. Man United historically and Spurs recently have both had a knack of scoring "big late goals". I put it to you that this is less about how great the clubs are than it is about how both the impetus of the side needing a goal and the psychology of the situation affect the game.

We saw that there were games earlier in the season where Spurs had a lead against inferior opposition but sat back to prioritise defence and eventually (inevitably?) conceded. It's something England have always been criticised for too. It's almost a narrative-drive, and it disproportionately affects teams (or people in general) more as they get closer to the (a) deadline, as one side totally prioritises attack and the other totally defence. The threat of the counter disappears - and in the Premier League more often than not the attacks involved are good enough to stick one away. So this is a big deal.


Subs (even by terrible managers) inevitably give some kind of boost because they're fitter than the players they replace. This is one part of the Longkaku Effect.


The effect of the timing of the substitution is also clearly key. The later it is, the bigger the disparity in fitness but also the smaller the amount of time to have an effect. AVB clearly thinks of 70 minutes as being the ideal point, and honestly I think having seen the impact of our late shows this year, I'm inclined to agree. I'll come back to this later.


More subsitutions, more offensive substitutions (striker for midfielder, midfielder for defender), and a situation-dependent change of system (switch to 3 at the back etc) magnify this effect further. We saw against Arsenal - in what the history books record as a convincing 5-2 loss but memory recalls as a surprisingly close game - that the impact of a change of system can be very dramatic. AVB has taken a switch from what I'm going to call "Control Mode" to "Creative Mode" as being very important. We've seen it a number of times, particularly effective against City where Control Mode was leading to a tight, edgy game and then suddenly City were blown away in about 10 sensational minutes.

I would argue at this point that it is crucial to include Control Mode in our gameplan in order to enable us to win games, though. Against City under HR, we were forced to use Creative Mode from the outset due to a number of injuries and suspensions. We ran them close but ultimately we lost 4-3, conceding a late goal in a bizarre game that could have ended 10-0 to us or to them. It's not a system that produces reliable results over 90 minutes, even if it smacks of Spurs' commendable attitude toward echoes of glory in failure and 5-4 defeats and 4-4 draws and all that kind of shenanigans. We haven't seen those this year at all (although we also haven't seen Liverpool-style high-scoring demolitions of poor opposition) and we've not once been beaten soundly. This is a direct consequence of the Control Mode.


It's also worth saying that the key part isn't just the wiggly line, it's fundamentally the area underneath the graph. So the gradients and the starting point and the size of the "70 minute change" gap and all of these things have important parts to play and ultimately the real crux is what the total coloured area ends up being.


So this is how I'd represent our system this year. The initial performance I've taken down quite significantly, reflecting Brett Rainbow'sstats regarding the difference between Parker and post-Parker.


This is how I'd represent the changes between the "Control Mode-Creative Mode Transition System" versus the "Creative Mode System" where we start without a Sandro or a Parker and relentlessly try to break down an opponent. I'm using "system" here to describe a whole-game approach, where a team can have several Modes within a System. The darker line is Creative System and the paler line is the Control-Creative Transition System. The red area is where the Creative System would be more effective (expect better goal difference) and the blue is where Creative-Control System would be better.

As can be visualised, we could still make an effective 70 minute substitution for a boost (though this boost is clearly less major than switching from a more containing system to a more rampant attacking system) and it's entirely conceivable that either system could be better, depending on the difference in sizes between the blue and the red areas. MCofA's stats suggest that this year the blue area won out.


My theory is that the key is that the post-70 minute area is significantly more important psychologically than the pre-70 minute area. Remember that the opposition will tend to hang on in these periods, multiplying any attacking change's effectiveness. Think late Javier Hernandez subs, or Aguero's last minute title winner for City, or... every goal Bale has scored this year. Or the late goals Defoe scored. The probability of the team needing a goal scoring in this period is significantly higher than at any other point. It categorically is not luck that has found us scoring a lot of goals late on, resulting in the goal difference discrepancy Brett Rainbow identified between our Control Mode and Creative Mode. It is by design, as part of a system which is more reliable than all-out attack from the outset.


The fact that the Control-Creative Transition System selectively targets our strongest personnel at their fittest against the opposition's tired defence makes it a more effective system than the Creative System.

Thus, Parker (or Sandro, who would raise the pre-70 minute level higher) should start, and both MCofA and Brett Rainbow's stats are consistent with this.

If you're still not convinced, then I refer you back here.

Thanks for reading!

Before you write a FanPost: Lurk! Comment! Talk with your fellow members! If you've been here for more than a couple of days and you know what's up, write away!

Log In Sign Up

Log In Sign Up

Please choose a new SB Nation username and password

As part of the new SB Nation launch, prior users will need to choose a permanent username, along with a new password.

Your username will be used to login to SB Nation going forward.

I already have a Vox Media account!

Verify Vox Media account

Please login to your Vox Media account. This account will be linked to your previously existing Eater account.

Please choose a new SB Nation username and password

As part of the new SB Nation launch, prior MT authors will need to choose a new username and password.

Your username will be used to login to SB Nation going forward.

Forgot password?

We'll email you a reset link.

If you signed up using a 3rd party account like Facebook or Twitter, please login with it instead.

Forgot password?

Try another email?

Almost done,

By becoming a registered user, you are also agreeing to our Terms and confirming that you have read our Privacy Policy.

Join Cartilage Free Captain

You must be a member of Cartilage Free Captain to participate.

We have our own Community Guidelines at Cartilage Free Captain. You should read them.

Join Cartilage Free Captain

You must be a member of Cartilage Free Captain to participate.

We have our own Community Guidelines at Cartilage Free Captain. You should read them.




Choose an available username to complete sign up.

In order to provide our users with a better overall experience, we ask for more information from Facebook when using it to login so that we can learn more about our audience and provide you with the best possible experience. We do not store specific user data and the sharing of it is not required to login with Facebook.