THREAD:

Today I’ll breakdown #DeGea’s performance in the #PremierLeague using my bespoke GK models!

The headline is for the first time in 3 seasons #DeGea’s shot stopping skills are outweighing his weak shot prevention skills & thus far he has saved #MUFC ~5 additional goals! Image
Shot Stopping:

#DeGea’s shot stopping has been world class, my model predicts he has saved #MUFC ~5 goals more than an average #PL shot stopper would’ve thus far!

The goalmouth map highlights just how hard #DeGea has been to beat, if you don’t hit the corners you won’t score! Image
Shot Stopping:

His Best 3 Saves (Lowest ExSave Saves Made) are below 😮:

1. 24% Save Probability vs Jesus

2. 34% Save Probability vs Almiron

3. 42% Save Probability vs Pukki

They highlight #DeGea big strength, his reflexes & ability to cover huge amounts of the goal area. ImageImageImage
Shot Stopping:
His Worst 3 Goals Conceded (Highest ExSave Goals Conceded) are here 😔:

1. 91% Save Probability vs ESR

2. 84% Save Probability vs Pedro

3. 63% Save Probability vs Salah

They highlight #DeGea’s issues if he fails to sort his feet out, especially at the near post ImageImage
Handling:

#DeGea’s handling has presented the opposition with 0.5 more ExG than the avg #PL GK would be expected to given the shots he’s faced, but his handling has been solid & if the poor parries vs #WWFC & #LCFC are removed his handling has prevented #MUFC 0.8 ExG in rebounds Image
Handling:

I don’t think #DeGea’s handling is something he needs to work on & it hasn’t been a consistent problem in the past few years so I think with a larger sample size his numbers should improve given they are driven by just 2 poor parries! ImageImageImageImage
1v1 Stopping:

#DeGea’s 1v1 stopping has been solid!

My model predicts he has saved #MUFC ~2 goals more than an average #PL shot stopper would’ve thus far!

His calmness when facing long range 1v1s has been excellent & he has consistently given the striker the hardest finish! Image
1v1 Stopping:

I do worry that #DeGea’s close range 1v1 level isn’t sustainable & could drop off & given it is only average at the moment that could cost #MUFC goals!

DDG has relied on his reflexes to make 4 of the saves rather than engaging which my model finds is suboptimal! Image
Cross Claiming:

As usual #DeGea’s cross claiming performance has been well below average.

My model finds that while he often catches & rarely misses the ball he is far too inactive & thus has only prevented ~0.15 ExG per game due to claims (0.10 ExG lower than an avg #PL GK!)! Image
Sweeping:

#DeGea’s sweeping is also below par!

Again while he is very tidy when he does come off his line my model finds he has prevented just ~0.05 ExG per game due to sweeps (0.05 ExG less than an avg #PL GK), so again his poor output is driven by a lack of activity! Image
Conclusions:

Overall #DeGea has been worth +0.33 goals per game for #MUFC this season

While his shot prevention remains down their with the worst goalkeepers in the #PremierLeague his excellent shot stopping has meant he has had a positive influence on #ManchesterUnited Image
Conclusions:

This is the first time in 3 seasons where #DeGea’s actually been an overall above avg GK:

Shot stopping 19/20 = +0.02
Shot preventing 19/20 = -0.16

Shot stopping 20/21 = +0.06
Shot preventing 20/21 = -0.15

Shot stopping 21/22 = +0.51
Shot preventing 21/22 = -0.18 Image
Regarding the claim that #DeGea is the GK of the season:

Not for me (not just yet anyway).

It is important to note that GKing is more than shot stopping & here are #Ramsdale’s numbers a GK who is performing at +0.44 goals per game above average thus +0.11 more than #DeGea! Image
Thank you everyone for taking the time to read all this & a happy new year!

I will do a similar thread to this for #Alisson in January, I think Ali is having an excellent season & his performances aren’t talked about enough due to a recent drop in form over the past month!
Also what is going on with the left side of #MUFC’s defence?

A huge fraction of the 1v1s #DeGea has faced this season have come from the space between the left back & the left centre back, I am not sure if teams just attack that side more or if that side is just much weaker? Image
If you are interested in the shot prevention model here is a thread that outlines the model in full with lots of handy breakdowns & match footage!

#GKUnion #GKAnalytics
Additionally if you are interested in the 1v1 shot stopping model here is a thread which breaks down that model & analysis in full!

• • •

Missing some Tweet in this thread? You can try to force a refresh
 

Keep Current with John Harrison

John Harrison Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @Jhdharrison1

24 Dec 21
THREAD:

How good has #Ramsdale been for #Arsenal so far?

I’ve seen him described as both “the signing of the season” & “a camera save merchant” but what does the data say?

In this thread I will breakdown his performance in all areas of goalkeeping using my bespoke models

#AFC
Shot Stopping:

#Ramsdale’s shot stopping has been excellent, my model predicts he has saved #AFC ~3 goals more than an average #PL shot stopper would’ve thus far!

He’s basically only conceded vs close range shots in the corners & has pulled a few shots out of the top corners!
Shot Stopping:

His Best 3 Saves (Lowest ExSave Saves Made) are below 😮:

1. 31% Save Probability vs Bowen

2. 52% Save Probability vs Maddison

3. 57% Save Probability vs Moura

They highlight #Ramsdale big strength, his agility & ability to cover huge amounts of the goal area.
Read 16 tweets
3 Nov 21
Last night was rough for #DeGea

The 1st goal was a textbook shot stopping technique selection mistake

The shot went next to DDG’s right foot but rather than go with his foot he went with his hand

It was odd to see as DDG’s shot stopping technique selection is usually flawless!
I actually thought #DeGea’s footwork before the shot occurred was excellent & it got him into a great position to make the save but his shot stopping technique selection let him down at the last moment & meant he couldn’t make the save!

#MUFC #UCL #ATAMUN
The 1st goal was a mistake but I am way more concerned about the 2nd.

Even though #DeGea had ample time to judge the 1v1 would be from close range he failed to get close enough to the ball to maximise the goal area covered & snuff out the shot.

He needed to be on top of the CF!
Read 11 tweets
1 Nov 21
My models view of #Ramsdale’s FK save:

He was well positioned as he was stood on the line between the ball & the outside of the wall!

It was a good, but not unbelievable, save. My model finds such a FK gets saved ~50% of the time due to its low speed & distance from the post! Image
I’ve seen the save described on Twitter as both the save of the season & poor positioning followed by a slow recovery!

As usual the truth is between the two & I should make it clear here that saving a 50/50 shot is a lot closer to the save of the season than poor GKing!

#AFC Image
I should add that #Ramsdale had a phenomenal game, saving 4 out of 4 1v1s!

Which is unreal!

If you run all the shots he faced through my model you find #Ramsdale saved #AFC in total 2.38 goals above average which in a 2-0 win means he literally won them the game! Image
Read 4 tweets
29 Oct 21
#DeGea’s ability to deal with crosses has quite rightly been criticised.

In the #PremierLeague #DeGea has only prevented 6 opposition chances occurring by claiming crosses!

My model finds this lack of activity means #MUFC concede on average an additional 0.14 ExG every game!
In comparison #Alisson of #LFC has prevented 18 opposition chances occurring by claiming crosses & #Alisson has played 90 fewer minutes!

This difference in cross claiming ability corresponds to #LFC facing on average roughly 0.36 fewer ExG per game in comparison to #MUFC!
Additionally #DeGea’s sweeping numbers have also been low this season.

In the #PremierLeague #DeGea has only prevented 4 opposition chances occurring by sweeping throughballs!

My model finds this lack of activity means #MUFC concede on average an additional 0.07 ExG every game!
Read 9 tweets
23 Oct 21
I’ve seen this describe as a wonder save but it’s a decision making error followed by a good recovery

My 1v1 model predicts this to be scored 2% of the time if the GK holds deep while if the GK rushes, like #Mendy, it is scored 64% of the time

He made it way harder for himself!
I think #Mendy is a top GK by the way & he’s made many great saves this season but just because a GK is good doesn’t mean we should praise them regardless of how good their goalkeeping actually is!

Outcomes shouldn’t dictate analysis, processes should!
If anything this save should actually worry #Chelsea fans a bit as it is reminiscent of an issue #Mendy had last year when he gave away soft goals vs #LUFC & #EFC

Hopefully for #CFC fans it was a 1 off bad decision & it won’t happen again unlike in the middle part of last year.
Read 4 tweets
23 Oct 21
#Ramsdale’s distribution last night was unreal!

All 12 of his open play passes resulted in unpressurised #AFC possession & made it more likely that #AFC scored next!

Additionally 7 of those passes successfully broke #AVFC lines & took multiple players out of the game!

#FNF
What a ball this was

It’s unbelievable awareness to spot the run, then exceptional technique to pick it out!

Being able to set up 2 v 3 situations from the GK is so useful

This pass was certainly the standout moment for me but importantly it was not an isolated incident at all
This pass map is genuinely outrageous!

I don’t think I have ever seen a GK record a 100% successful & positive open play game where they were pressed on 40% of their passes & where they attempted to break the oppositions formation with 67% of their passes!
Read 4 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!

Ethereum

0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy

Bitcoin

3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us on Twitter!

:(