Author Topic: Parker - seeing out games  (Read 1572 times)

Offline Statto

  • Gentleman Jim
  • ***
  • Posts: 8213
Re: Parker - seeing out games
« Reply #20 on: December 04, 2019, 05:37:04 PM »
Another stat for the last part of the game

0-75 mins we are ranked 2nd in avg number of shots with 11.3 (Leeds are top with 13.4)

75-90 minutes we drop to 18th in the table (Leeds are 7th)
I'd add that only Wigan have less shots in the penalty area than we do in the last 15 minutes of a game.

But again, totally disregards the elevated effort levels from the opposition when they're 1-0 or 2-1 down in the final 10-20 mins versus the effort level when they're cruising at say 2-0 up versus Wigan.

Would also be interested to see what the outcome was of a Jokanovic's approach - so if you take those games where we had a 1-0 lead in the final 20 minutes, how often did we convert that to a win (and ultimately, was it better than Parker's 75% conversation rate?)

Offline MJG

  • MAESTRO
  • **
  • Posts: 2566
  • Quis custodiet ipsos custodes?
Re: Parker - seeing out games
« Reply #21 on: December 04, 2019, 05:41:02 PM »
Another stat for the last part of the game

0-75 mins we are ranked 2nd in avg number of shots with 11.3 (Leeds are top with 13.4)

75-90 minutes we drop to 18th in the table (Leeds are 7th)
I'd add that only Wigan have less shots in the penalty area than we do in the last 15 minutes of a game.

But again, totally disregards the elevated effort levels from the opposition when they're 1-0 or 2-1 down in the final 10-20 mins versus the effort level when they're cruising at say 2-0 up versus Wigan.

Would also be interested to see what the outcome was of a Jokanovic's approach - so if you take those games where we had a 1-0 lead in the final 20 minutes, how often did we convert that to a win (and ultimately, was it better than Parker's 75% conversation rate?)
a quick glance of the image below looking at where the goals were show how we carried on the whole game.
It's the dramatic change we seem to show that worries me, it's just just reduced to a trickle of chances, it's almost like we switch the tap off.

Offline Bokonon

  • Lakey/Dark Room
  • *
  • Posts: 84
Re: Parker - seeing out games
« Reply #22 on: December 04, 2019, 11:32:00 PM »
Another stat for the last part of the game

0-75 mins we are ranked 2nd in avg number of shots with 11.3 (Leeds are top with 13.4)

75-90 minutes we drop to 18th in the table (Leeds are 7th)
I'd add that only Wigan have less shots in the penalty area than we do in the last 15 minutes of a game.

But again, totally disregards the elevated effort levels from the opposition when they're 1-0 or 2-1 down in the final 10-20 mins versus the effort level when they're cruising at say 2-0 up versus Wigan.

But then again, aren’t all losing sides going to raise their game and open things up in the last 10-20? If so, why does Leeds have a better shots taken stat versus ours in the same part of the game?

I’m not saying there’s an obvious answer there, but I think you are relying too much on this point as if it uniquely affects our club. And if that is the case (we are uniquely afflicted with teams who raise their effort at ends of games when down to us), I think it begs the question of “Why?”


Offline Statto

  • Gentleman Jim
  • ***
  • Posts: 8213
Re: Parker - seeing out games
« Reply #23 on: December 04, 2019, 11:42:59 PM »

But then again, aren’t all losing sides going to raise their game and open things up in the last 10-20? If so, why does Leeds have a better shots taken stat versus ours in the same part of the game?

I’m not saying there’s an obvious answer there, but I think you are relying too much on this point as if it uniquely affects our club. And if that is the case (we are uniquely afflicted with teams who raise their effort at ends of games when down to us), I think it begs the question of “Why?”


Absolutely not “relying... on this point as if it uniquely affects our club" - never suggested that at all

Leeds may be higher than us but they're still only 7th, ie, much worse in the last 15 minutes of games than in the first 75. So clearly this things affects them, in the same way

The thing is, to do a proper, useful comparison that tests the specific point being made in this thread, we'd need to isolate the games where Leeds were defending a single goal lead in those final 15 minutes 

Offline colinwhite

  • Jimmy Hill
  • *
  • Posts: 2174
Re: Parker - seeing out games
« Reply #24 on: December 05, 2019, 06:05:54 AM »
Some very good points made on this thread. Maybe the best defence in the last 20 minutes of games is to start better and be ahead in a fairly safe lead (derby at home ) when teams start throwing the kitchen sink at you .We backed off went defensive but dared to score the third ,in the home game against Derby.
If we had only been 1-0 up Cairney probably wouldnt have made  that run in support of Mitro that led to our third goal.

Offline RaySmith

  • Gentleman Jim
  • ***
  • Posts: 6076
Re: Parker - seeing out games
« Reply #25 on: December 05, 2019, 08:14:27 AM »
You're right Colin - it's always going to be very tense defending a narrow lead, and teams are always going to come at you in an all or nothing attempt to save the game.

Parker, or any manager, is always going to be  challenged for his decision - either to bring on extra defenders, or try and score yourselves - if the outcome is that we  concede in the final minutes, and only draw, or even lose, a game we were winning.

Personally I would go for bringing on extra defensive  support, if we only have a narrow lead - I've seen us hit on the break when we've gone for it near the end of games, with everyone  up the field -so many times.


Offline Jim©

  • tjl
  • *
  • Posts: 203
Re: Parker - seeing out games
« Reply #26 on: December 05, 2019, 10:23:22 AM »
I think this thread highlights the problems that managers face from the fans too.
Damned if you do, damned if you don't.
As stats show, we've been pretty good at holding onto leads in many of the games and the changes Parker has made have certainly helped. However it's nervy and there's no one more nervous than us, the fans.

Say for example that Swansea had scored with one of their late chances- SP would be getting a load of stick about making the wrong decisions, not knowing what he was doing etc. It's fine lines and the disagreement amongst us all show how hard it must be for him to make those decisions!

Offline Southcoastffc

  • The Bard/Corked Hat
  • *
  • Posts: 3023
  • Mr Khan said "Whatever It Takes" That's ok then.
Re: Parker - seeing out games
« Reply #27 on: December 05, 2019, 10:52:48 AM »
Its a good informative post but as we do not have control group outcomes if we continued to press (which I believe is more effective with our squad) we have nothing to compare with.
I do not like the idea of 'seeing out a game.' I tend to think it goes wrong more often than pays dividends. You do not have to go crazy if winning by one nnear the end but the mania to bring on defenders and time waste often back fires.  I think if you are playing well, stick with it, do not invite pressure, if the ball is allowed to be launched into your box to often then percentages will kick in and the opposition will get a goal... better to get down their end and either play keep ball or better still try and score another exploiting the spaces left by an over exposed opposition.

Agree with both of these - sometimes we revert simply to kicking the ball away and thus keep inviting pressure.

Online toshes mate

  • Gentleman Jim
  • ***
  • Posts: 5066
  • Vitam Impendere Vero
Re: Parker - seeing out games
« Reply #28 on: December 05, 2019, 11:28:56 AM »
The big problem within this thread is that statistics do not deal with what is to come they only deal with what is past.  In other words the whole notion that you can use a 'control' within what is known becomes flawed unless you understand the difference between probability (i.e. chance) and the randomness of reality. 

Probability is that a class player will instantly control a good pass from a team mate whereas the reality is that he may, even for the one and only time in his career, make a hash of it leading to an opponent scoring a goal.  Randomness is about everything being an anomaly (i.e. a recorded unique event) as compared to the mean which is an average or mean of what has happened before that anomaly became fact.  The class player I referred to earlier could steadily become a not so classy player if sufficient recorded anomalies change his factual mean or average. 

The key here is that outcomes may have absolutely no connection with any statistical material available which is why so many businesses use the risk element in creating their operational margins.  Those who are unrealistic in doing this generally fail.  For a football management team the risk is complex to the extreme (randomness in huge numbers e.g. players, offcials weather, and so many others) and they earn their money from results which are, at the very least, concrete and factual.   If there were a fail safe and totally reliable method then everybody would copy it and every match would have the exact same outcome - i.e. stalemate.  We look at statistics and data in any shape or form at our peril - is the data even reliably accurate is the first thing on the list.


Offline epsomraver

  • Gentleman Jim
  • ***
  • Posts: 9232
Re: Parker - seeing out games
« Reply #29 on: December 05, 2019, 03:51:26 PM »
Not sure those subs confirm the point being made. Is Onomah a dm, or am. The thing that SP has been doing is going to a back three, with 2 fullbacks and at least one dm to close out games

Didn't want to bore people by getting too deep into the weeds but if you focus on the late changes, say those after 70 mins, they're overwhelmingly defensive (irrespective of what you call Onomah) and actually two of the exceptions are Sheffield Wednesday and West Brom, and we all know what happened there, which strengthens the point being made.

Agree regarding the formation changes, although I think it pretty obvious that if MLM comes on for Knockaert, for example, it's not so we can carry on playing 4-3-3 with MLM on right wing.

Sorry don't consider either of the Reeds as defensive so your comparison is flawed if you are suggesting we put on defenders only to see out a 1-0 lead

Offline Statto

  • Gentleman Jim
  • ***
  • Posts: 8213
Re: Parker - seeing out games
« Reply #30 on: December 05, 2019, 04:03:24 PM »
Did a comparison against Jokanovic 17/18 season

By my calculations there 14 games where we had a 1 goal lead after 70 mins, a further 6 where we took the lead between 70 mins and 86 mins, and 2 more where we took the lead in injury time

If you exclude those last two (seems a bit unfair to include them as we only had to defend the lead for a couple of minutes, and in both cases we'd also already used all our subs) then Jokanovic's conversion rate was 80% (16 wins out of 20) versus Parker's rate of 75% so far.

So a bit better but not significantly

Offline Statto

  • Gentleman Jim
  • ***
  • Posts: 8213
Re: Parker - seeing out games
« Reply #31 on: December 05, 2019, 04:05:45 PM »
Not sure those subs confirm the point being made. Is Onomah a dm, or am. The thing that SP has been doing is going to a back three, with 2 fullbacks and at least one dm to close out games

Didn't want to bore people by getting too deep into the weeds but if you focus on the late changes, say those after 70 mins, they're overwhelmingly defensive (irrespective of what you call Onomah) and actually two of the exceptions are Sheffield Wednesday and West Brom, and we all know what happened there, which strengthens the point being made.

Agree regarding the formation changes, although I think it pretty obvious that if MLM comes on for Knockaert, for example, it's not so we can carry on playing 4-3-3 with MLM on right wing.

Sorry don't consider either of the Reeds as defensive so your comparison is flawed if you are suggesting we put on defenders only to see out a 1-0 lead

So you don't consider Reed for Cavaleiro as a defensive change?
Fine, that's your prerogative

(but you're wrong)



Offline Craven_Chris

  • Michael Jackson
  • *
  • Posts: 1
Re: Parker - seeing out games
« Reply #32 on: December 06, 2019, 12:12:13 PM »
Hi first time poster here...

I sent some data into the Fulhamish podcast recently on this topic so thought it might be worth adding here: the summary of that data is that, under Scott Parker, there are 18 games where we have gone into the 75th minute with a good scoreline that we would want to defend (basically I have defined this as a lead, or a draw in difficult circumstances such as being down to 10 men or playing away in the premier league or against a top 6 premier league team).

In 9 of these games, we have switched to a back 5 (using substitution to replace a midfielder / attacker with a defender) and in 9 games we have not. So the definition of the move to a more defensive formation is slightly different to that used above by Statto, Im specifically referencing moving from 4 to 5 at the back (so not including bringing on a CDM for a winger).

 The basic summary is that in the 9 games with the defensive switch, we conceded zero goals and obviously therefore held on to the result each time.

In the 9 games without a defensive switch (which basically serve as a sort of control group), we conceded late goals in 6 of them (7 goals in total) and this turned the result against us in 5 of the 9 games.

There is a slight caveat in that I have included Everton at home last year as a non-defensive switch game although we did go 5 at the back in injury time, but we had already extended the lead to 2 by then so thought it a better example of a non-defensive switch success!

For info the defensive switch games were Bournemouth away, Cardiff home, Blackburn home, Huddersfield away, Cardiff away, Middlesbrough away, QPR home, Derby home, Swansea away. We also only scored one late goal in these games, that being the recent Cairney goal against Derby

The non-defensive switch games (control group) were: Leicester away (2 late goals conceded), Liverpool home (1 late goal conceded), Everton home (no goals conceded, one late goal scored), Wolves away (1 late goal conceded), West Brom home (1 late goal conceded), Sheffield Weds away (1 late goal conceded), Wigan home (no goals conceded), Luton home (1 late goal conceded but won anyway), Birmingham away (no goals conceded).

I dont think this result is actually conclusive yet at all, one could point out quite a few of the non-switch outcomes were against elite opposition in the premier league for example - and indeed there were moves towards a more defensive set up in the control group (just not a back 5). 

But it suggests for now that the tactic has been working, even if its too soon to say where its just attributable to chance. I suppose an analyst with more data would look at the xG against different formations in the late game from matches around the world, and I would hope Fulham are doing this!

A relevant point though is that it has become predictable, and other teams may now be designing tactical changes specifically to break down a Fulham back 5, given its very likely we will use that now to defend a lead.




Offline Statto

  • Gentleman Jim
  • ***
  • Posts: 8213
Re: Parker - seeing out games
« Reply #33 on: December 06, 2019, 02:07:09 PM »
Great post Chris

Offline The Rational Fan

  • Jimmy Hill
  • *
  • Posts: 2293
Re: Parker - seeing out games
« Reply #34 on: December 06, 2019, 11:28:20 PM »
Very Good post Statto, Great post MJR and Great post Chris.

We have a decent team to keep teams out. We kept three clean sheets in a row last season, Parker focused pre-season on being able to defend against good attacks and statistics say it looks like working now.

We could of attacked against WBA and Sheffield Wednesday, but my guess is we wouldn’t have scored and we would have still conceeded. I am sure many of the other games we won, attacking more would have got more goals but would that make a difference.

Those last 20 minutes other teams will throw everything at us no matter what, if we attack too the only certainty is more goals. If your one nil up, maybe it’s better to play for 75% for not conceeding than a 62% of scoring next (as we have scored 31 goals of the 50 goals in our games). .

As those that think Reed for Cavaleiro is not a defensive change, is it defensive or attacking? I would point out that every substitution is either slightly defensive or attacking (eg Reid for Mitro is marginally defensive and MLM for Bryan is more defensive/less attacking. If Reed for Cavaleiro is not an attacking move, which is surely the case then it is at least a little defensive.
« Last Edit: December 07, 2019, 01:14:54 AM by The Rational Fan »


Online toshes mate

  • Gentleman Jim
  • ***
  • Posts: 5066
  • Vitam Impendere Vero
Re: Parker - seeing out games
« Reply #35 on: December 07, 2019, 08:30:21 AM »
I don't think this result is actually conclusive yet at all, one could point out quite a few of the non-switch outcomes were against elite opposition in the premier league for example - and indeed there were moves towards a more defensive set up in the control group (just not a back 5). 

But it suggests for now that the tactic has been working, even if its too soon to say where its just attributable to chance. I suppose an analyst with more data would look at the xG against different formations in the late game from matches around the world, and I would hope Fulham are doing this!

It is good to learn about your endeavour and enthusiasm for data and statistics connected with FFC.  The thing about theories and accepted 'proof' of aunthenticity via review is that these are also random incidents with a strong probability (as close to certainty as you can get i.e. one) that progressive study will find a flaw in the data, the processing or the analysis.  That is how progress has been made in all things since ancient times before, long before we got so caught up with computers and statistics.  The lowest common denominator in progress is that something works.  Making a process more efficient is never ending since chance cannot be solved mathematically because there is always infinite possibility even in matters which look to have certainty.  An interesting place to find out more about this is to study gambling theory where some people have made lots of money because they were able to crowd fund attempts to beat the odds.  Search for people who have won the lottery more than once.