Stu255 RMT 19/20

[Ironfist]

He can't do that unfortunately as he already has 3 LIV

You're right... I completely missed it seeing as Alisson is a rare thing nowadays.
I'd certainly get rid of one of the Liverpool defenders and bring in Mane.

[Stu255]

He can't do that unfortunately as he already has 3 LIV

You're right... I completely missed it seeing as Alisson is a rare thing nowadays.
I'd certainly get rid of one of the Liverpool defenders and bring in Mane.

Had a good look at selling Sterling and Alisson to bring in Mane.

But not sure it’s the right move?

Exhibit A

B0DB2C43-3D34-4667-90C6-3160B3F3F635.jpeg

I know Sterling has some tougher fixtures, but I still think he is the stronger asset. Not sure it’s worth 2x transfers.

I think Mane has flattered himself, and the clamour of so many others to bring him in, and the ensuing bandwagon, might be an opportunity to steal a march.

[Stu255]

I've been looking at PPDA and OPPDA recently and it's starting to look very interesting when taking alongside other models...

GW13 chart.png

PPDA (passes per defensive action) is effectively a measure of how good a team is at pressing. The lower the better. One way to think about it is 'how good a team is at breaking up their opponents possession'.
If you plot PPDA against xGA (both red here) you get fairly good correlation across the Premier League.

That is teams that are good at pressing tend to have lower xGA (I think of xG and xGA as a measure of 'opportunity to score').

Interestingly OPPDA (opposition passes per defensive action), is the opposite it is effectively a measure of how good a team is at beating the press. The higher the better. One way to think about it is 'how good a team is at resisting having their possession broken up'.
OPPDA seems to have excellent correlation with xG (both green above). This is interesting, as it suggests that teams that can string lots of passes together tend to create more xG (again xG is 'opportunity to score')

I think these 4 metrics begin to represent the 4 phases of play...

where:-
xG = in possession in opposition half (ish)
OPPDA = in possession in own half (ish)
PPDA = out of possession in opposition half (ish)
xGA = out of possession in own half (ish)

These 4 abilities could then map onto the other team and the big mismatches might hint at where a particular player is likely to find an unusual competitive advantage.

Seems like the next step is to build such a model and backtest it.

But first, how about a radar diagram as way to display the 4 stats against the 4 phases between two opposing teams?

[Stu255]

OK, have done a bit of work with this now and have the following charts for the upcoming GW13 matches...

GW13.png

Basically a horizontally skewed diamond favours the home team and a vertically skewed diamond shape favours the away team.
The labels on the four corners of the diamond describe the phase of play for the HOME team.

It's intended to illustrate the strength of each team for each of the four phases of play. I haven't bothered to look at teams in terms of home and away just their overall data. I know plenty of teams play different tactics home and away so could look at this next.

But it gives a good illustration as to where the big battles are likely to occur on the pitch and therefore which individual players might benefit from various teams lining up. It's a bit more nuanced than xG alone.

[Stu255]

OK have redone this now based on home / away performance. I recognize that most teams have different styles of play for home / away and so hopefully this captures that now.

GW13 HA.png

The radar labels describe the game phase from the home team point of view, a horizontally skewed diamond suggests a home team advantage a vertically skewed diamond suggests away team advantage. Having Man CIty and Liverpool in the dataset tends to make all the other teams look relatively bad. Will properly normalise the four dimensions at some point to make it all a bit more accurate, and also maybe at some later time look at a two tier model because of MCI and LIV.

NB:- Can already see a few tweaks that will make these phase-of-game radars (POGR's to coin a term) a bit more readable / intuitive, such as flipping the horiztonal axis so the home team is carrying the ball from left to right.
But I think this is a worthwhile piece of work and will expand on it with some comments in due course.

[Stu255]

Hmmm,

As an FPL tool it might be better to have a dial for each team (rather than each match), and plot the next 5 or 6 fixtures onto one radar for each team. Can then normalize the opposition data relative to the team in question and should clearly show which phases of play teams would be expected to enjoy a competitive advantage.

[Stu255]

OK, here are some phase of game radars (POGR's) for a few select teams over their next 6 fixtures...

In the POGR a horizontally skewed diamond shows a more favourable fixture, a vertically skewed diamond shows more difficult fixtures.

MCI POGR

MCI13POGR.png

Pretty clear where MCI strengths lie, when they have possession in their own half they are incredibly good. They are also excellent at bringing the ball forward and creating lots of xG. City are happy to win the ball deep and build up play patiently.
In terms of their high press, they are slightly above average but defending in their own half they give away very few chances. Their next run of fixtures looks excellent with MNU looking like their toughest fixture based on the underlying stats alone. Also interesting is that Leicester might disrupt them defensive possession with their excellent high press but the data still favours city for this home fixture.


ARS POGR

ARS13POGR.png

Arsenal look likely to string some good form together based on this, they should also get some clean sheets from these fixtures with the exception of MCI. Unusually at home v's Brighton is their easiest fixture here, looks like a clean sheet and plenty of goal threat from Arsenal. Arsenal whilst looking good defensively for most of these fixtures look like they will struggle to get goals. Even at home to SOU looks like a struggle for a load of goals.

NOR POGR

NOR13POGR.png

Norwich look fairly comfortable on the ball in their own half but are completely toothless going forward. The teams they are playing look likely to get clean sheets with the exception of the Southampton away fixture. Norwich look suceptible to counter attacks v's wolves and Leicester (obvs). Blades will enjoy plenty of possession but might not get great xG.

[FranckKessie]

Always great to see new stuff! 6 matches per radar is too much for me and I find myself concentrating more looking at getting the colors/team right. It has to be said though that I am more of a table person, so may simply be me.

What if you have a 6 windows in the graph(one per fixture) for a given team?

[Stu255]

Always great to see new stuff! 6 matches per radar is too much for me and I find myself concentrating more looking at getting the colors/team right. It has to be said though that I am more of a table person, so may simply be me.

What if you have a 6 windows in the graph(one per fixture) for a given team?

Yeah maybe 6 fixtures is a bit busy.

The table format is a bit unwieldy that’s where I was before. It requires home and away stats for each team in question then you need to cross reference that with the opposition home/away.

The data is from the new understat tables.

If you add PPDA and OPPDA columns and then flick back and forth between home/away tabs to cross reference the 4 pairs of stats (xG/xGA, xGA/xG, PPDA/OPPDA, OPPDA/PPDA) for each fixture.

For 6 fixtures that would be 48 table look ups, then for 20 teams thats almost 1,000 look ups.

I CBA doing that every week!

So trying to find a timesaver visual method. I know different people prefer reading different data methods.


I guess the radars are there so you can identify teams with a coherent shaped radar as this suggests that certain of their players are likely to enjoy/suffer a good or bad run.

eg Norwich do not look like scoring many in their upcoming games.
City look like their defenders will get a huge amount of time and space on the ball and their deeper players may get more assists/goals than usual.

The radars are a higher level scan of fixtures, each radars is 48 table look ups. If you then want to explore individual fixtures the tables might be better?

But yeah personal preferences and all that.

What do you think of the generally principle of mapping xG, xGA, PPDA and OPPDA onto the 4 phases of play?

Not perfect of course, but as a crude high level yardstick does this seem plausibly interesting? Or illuminating in any way?

[FranckKessie]

What do you think of the generally principle of mapping xG, xGA, PPDA and OPPDA onto the 4 phases of play?

Not perfect of course, but as a crude high level yardstick does this seem plausibly interesting? Or illuminating in any way?

What you have shown so far I think could be a good way to get an idea of the style of play each teams have. How useful it is to assess FPL-points potential I am don't feel to be in a position to tell. The simplicity of xG and xGA (each being a single parameter) makes them very useful, but it may be that your 4 phases of play can highlight additional stuff.

Take Leicester of the last couple of seasons. Can your 4 phases of play highlight why they have, at least in FPL for Vardy, scored similar against better opponents as weak opponents. Our idea is that it is because better opponents wants to dictate the match, which fits nicely into Leicesters counter-attack style? While they historically have been limited by the quality of their central mids, this years is perhaps different because the quality of Maddison and Tielemans is greater than those of the past?

Personally I'd be interested in simply getting an overview of the characteristics of each team? I dont watch many games and have access to limited data, so my knowledge about each team is very limited I'd say. Especially teams I never watch on TV, which means more or less all teams but the traditional topSix.

[Stu255]

Personally I'd be interested in simply getting an overview of the characteristics of each team? I dont watch many games and have access to limited data, so my knowledge about each team is very limited I'd say. Especially teams I never watch on TV, which means more or less all teams but the traditional topSix.

To be honest I’m in the same boat and that’s basically the purpose of it.

Nobody can watch all the games, it’s so you can get a feel for how teams play and how that matches up with their next few fixtures.

You can see where teams strengths and weaknesses match up, or where one teams strengths match up with another’s strengths and a game is likely to be more draining on the otherwise better players in a team.

eg a team with a great press that plays 3 teams with high OPPDA in a row are likely to be a bit more tired than normal for the 4th game.

The POGRs try to pull all the cross matched table look ups (home/away, etc) into one glimpse.

[Stu255]

Have 2FT’s and considering some major surgery now.

Currently have...

DD426E75-574E-40BE-981B-77B24193F211.png

Thinking:-
Alisson to Gazzaniga
Sterling to Salah
Mount to [Loucas Moura] or A.N Other £8m mid

This would be a -4 hit for me.

Mane being parity with Salah just feels like a quirk of form and injuries and the current ownership balance between them seems unsustainable for any great length of time.

[Ironfist]

You are over complicating things...
Get rid of Alisson & TAA/Robertson for another Fox. Lose Mount & Saka for a Spurs mid. And when possible get Mane in.
Salah will mever be as good as before. He'll get goals but won't ever replicate his former form. In a couple of seasons he'll be playing for Herta Berlin or Fenarbache.

[Smurphy Paw]

The front six looks great and not worth taking a hit to tweak
Over the next couple of weeks I’d also look to remove Alisson. Not, in your case, for Gazzaniga though. You need him for GW18 which is when Spurs play Chelsea.

[Stu255]

You are over complicating things...
Get rid of Alisson & TAA/Robertson for another Fox. Lose Mount & Saka for a Spurs mid. And when possible get Mane in.
Salah will mever be as good as before. He'll get goals but won't ever replicate his former form. In a couple of seasons he'll be playing for Herta Berlin or Fenarbache.

I have 3 foxes, Pereira, Mads and Vardy.

Looking to get into Spurs, but not quite sure how/where? Also slightly concerned Spurs will be another Mourinho troll fest, but surely not? Surely?

Aren’t you a bit quick to write Salah off? He’s only 27. He’s actually younger than Mané and he doesn’t need to replicate past form, he is cheaper than Mané.


Was thinking of swapping Connelly for Mousset until this weekend.

[Stu255]

My squad is currently looking like this;

4CA4AFD4-DF87-4AF5-86EF-6E42F128317B.png

I am waiting until after tonight’s game to trade Mount for Pulisic, but then I’m actually very happy with the squad.

This is making me worried, because it is usually just when I feel happy with my squad that I am blindsided by a string of suspensions and injuries and unforeseeable benchings.

Probably best to roll up a spare TX as prep.

[Smurphy Paw]

We have near identical first elevens. The differences are personal preferences from the same teams (Mané/Salah, x2 different Foxes)
How are you looking for GW18? That’s increasingly what my forward planning will be focused on

[Ironfist]

Get Mane !

[Stu255]

Get Mane !

Phaaaark!

15pts the difference tonight. I have Mousset coming off the bench for Salah so hopefully he can get 5 goals against... oh Newcastle (could happen).

[Stu255]

Looking ahead at form and fixtures...

I’m actually thinking of this...

2B0E4A88-94DA-4BC4-98BF-A9050BC71C80.png

Remove Salah and Mousset and bring in Alli and Rashford.

This gives a very flat squad with no £12m players.

I’m not sure I need one with Vardy bailed as capt just now.

[Stu255]

What on earth should I do with this lot?

73DD87A7-1DF1-46B0-A633-A4188F9F0636.png

I’ve had 5 small red arrows in a row and am treading water around the 1m OR. Thought I was in good shape a while ago but it hasn’t been a great month for my team.

[From4corners]

I'd keep patience. Perhaps look at getting rid of Saka if two free transfers or yeah wait to see what Arteta has in store for him

[Stu255]

Not too worried about Saka as he is 15th man. It’s an empty shirt until BB week.

Currently thinking Mount Dilva is my next move.

[Stu255]

Currently sitting with this team and quite liking it.

2FT’s yesterday which were...
Mount Salah
Rashford Greenwood

Only issue is if Aguero explodes back onto the scene it is structurally very difficult to bring him in now. Hopefully Vardy keeps chugging a bit longer and Leicester don’t give up on the chase.

2A5B587B-7FBE-46B2-A811-67BB7706FC4E.png

[Smurphy Paw]

Looking good

[Gambit]

Currently sitting with this team and quite liking it.

2FT’s yesterday which were...
Mount Salah
Rashford Greenwood

Only issue is if Aguero explodes back onto the scene it is structurally very difficult to bring him in now. Hopefully Vardy keeps chugging a bit longer and Leicester don’t give up on the chase.

2A5B587B-7FBE-46B2-A811-67BB7706FC4E.png

Alli has some fitness issues (maybe hamstring!) according to JM so you may need to look at that. I also think Maddison now has real competition for less/similar money in Martial/Grealish/Antontio/Richarlison, I know his fixtures are good but still playing too deep for me (I recently used him as the makeweight to get Salah).

[Stu255]

OK, it now seems increasingly obvious that structurally I need to move towards Salah, Mane, TAA.

Robo needs to be liquidated to make room for Mané, and additional funds need to be found (probably from Tammy) in order to upgrade Alli or Maddison to Mané.

Mane and Salah both have insane xFPL over the next 5-8 weeks. Looks like there will be some big hauls.
GW24 looks like a potential TC lottery between Mane and Salah, with BB coming later in the season.

[Smurphy Paw]

As you have x3 Liverpool already you can take your time to do that, I’d suggest.
Whilst I have just doubled up on Mané & Salah I had to make the choice now, the alternative to bring in Robertson and have a defence like yours. It’s a good setup.
I’d agree by GW24 though

[Hotpot]

OK, it now seems increasingly obvious that structurally I need to move towards Salah, Mane, TAA.

Robo needs to be liquidated to make room for Mané, and additional funds need to be found (probably from Tammy) in order to upgrade Alli or Maddison to Mané.

Mane and Salah both have insane xFPL over the next 5-8 weeks. Looks like there will be some big hauls.
GW24 looks like a potential TC lottery between Mane and Salah, with BB coming later in the season.

My team look very similar to yours prior to this weeks Salah/Greenwood move. Assuming Greenwood keeps his place I plan to mirror the Salah/Greenwood move in the coming weeks.

I was however thinking of then sticking with the 2 defs and 1 of Salah/Mane for a while. Hoping that Robertson + the £5m elsewhere would make up for any hauls from Mane. It just feels like the rest of the team will be compromised to squeeze both Salah and Mane in.

How close was this call for you to go for both Salah and Mane over the extra def?

[Hotpot]

Pinched from the transfers thread but I think this illustrates the point I'm making above -

Anyone watch Calvert Lewin today?

I ask because if I do Tammy Calvert-Lewin tonight then it opens up Robo £4.3m def plus Maddison Mane.

You are contemplating taking out 3 of the best performing players in the game in Robertson, Maddison and Abraham (each of whom is the 5th best scoring player in their respective position) to accommodate Mane.

For me that is not worth it, especially as you already have Salah as a captaincy option.