Speculative Tactical Cohort (The free-for-all shared-RMT thread 20/21)

[Stemania]

Roll up, roll up! All are welcome to join the STC!

A free-for-all shared-RMT thread, for any spillage outside of personal RMT threads, and with a traditionally stats-based focus.

What's that we see before us? A strange and convoluted new thread title has appeared. Didn't this used to be called 'Statistical Thought Cave', 'Strategic Thinking Cavern', 'Specious Theoretic Chatter' or something? Well, yes, synonyms are so thin on the ground these days that we've had to outsource to a human thesaurus in the form of thebillfella (an honour bestowed after his phenomenal 6th overall finish last year). Indeed, thoughts of last season still linger strong, but in a blink of an eye the new season is upon us. Long has it been since the days Sammy The Crab formally hosted this geeky discussion (in the form of his own RMT) yet still we all trundle on, as caricatures of our past selves trying to replicate the intelligent analysis of the past. Our heads are fuzzy and our eyes are blurry. But nobly we march into the gloom of unknown points predictions, meandering socially distanced in search of that precious but elusive knowledge.




We drag our 854 page history along the seabed behind us. In a seven year period we've collected many treasures, and they continue to burden us. We daren't let go lest our efforts be lost forever in the murky sand: 2019/20, 2018/19, 2017/18, 2016/17, 2015/16, 2014/15, 2013/14.

It is, however, the time of year we can rejoice in rediscovering our beacon in the dark; our first task is to convince Mav3rick to make his return permanent.

So, welcome again all! Everything's a bit compressed this year so we probably won't have as much time as usual to fiddle with teams. I've transferred the ever-growing STC Bookshelf, feel free to suggest any recommendations for links/additions/improvements to the catalogue - Bagan is kindly on the task of updating the audiolibrary. And when the game starts, get strategising!!

---------------------------------------------------------------------------------

The STC Bookshelf


Key Texts:

Live FPL - Live gameweek rank calculator, plus other goodies (Best alternative to FPLD?)
Fplstatistics - Historically most reliable price change site, no registration
FISO's Crack the Code - CTC - making a comeback in 20/21 - no registration
FantasyFootballFix - Alternative price change site, registration required
Understat - Comprehensive xG database
Fbref - Free source of statsbomb's golddust xG data
Ben Crellin's Spreadsheet - Legendary DGW information source
FPL Statistico - Individual FPL team analyser
FantasyFootballScout - No introduction required
FantasyFootballHub - As above


Main Catalogue:
Squawka - Stats resource and comparison matrix
MathematicallySafe - Peter Blake's FPL statistical analysis blog
WhoGotTheAssist - User-friendly blog posts on psychology and probability in FPL
PLFantasy - Statistical blog by Chris Glover
MackayAnalytics - Blog specializing in xG comparison
Statsbomb - Legendary stats articles, specialists in xG
Differentgame - Paul Riley's blog on xG analysis
Eighty Five Points - Excellent blog by Laurie Shaw
FPLAcademica - FPL blog by four 'twitter enthusiasts'
FantasyFutopia - Blog by our own Tom Whelan (notably containing a great guide to Python in FPL)
Paul Riley, Mike Croom & Tim Bayer - Data resource tableaux (including fixture tracker)
Transfermarkt & Football-Lineups - Detailed player/team statistics, penalty/injury records, formations etc
Worldfootball & Soccerstats - Historical PL info
Myfootballfacts - Historic team penalty information
PremierFantasyTools - FPL tool site, inc alternative HoF
FPL Analytics - FPL analysis site
FPL Review - Points projection site based on understat xModel, inc "luck" quantifier
Teamfill - Customisable team optimiser site by our own Sneakyone
Vaastav's Github - Detailed historical FPL data (including weekly scores)
Football-data - Historic bookies odds for the PL
FPL Reddit - What it says on the tin
Spreadex market page - one of many possible sources for spread odds on seasonal g/a.



Audiobooks:
TBA - FPL podcasts of note


Online Collection:
TBA - twitter accounts of note
https://twitter.com/11tegen11?ref_src=t ... r%5Eauthor - 11tegen11
https://twitter.com/Lucid_Analytics - Lucid Analytics


Defunct ancient texts:
Fpldiscovery - Previously invaluable weekly player ownership/average scores data etc.
The3rdTurd's data - Old resource tableaux
http://11tegen11.net/
http://www.shotsontarget.co.uk/
DeepxG - Blog flirting with xG (you have to dig deep for relevant articles)
FPL Archives - FPL data from selected previous seasons, now dead

---------------------------------------------------------------------------------

The STC Workshop


Working Prototypes from 18/19:
Understat Extrapolations - Mav3rick's Solver algorithm for picking the best squad based on projections
FFS Extrapolations - Stemania's modification of the above for picking the best squad based on FFS projections (inc 6-week)
Defensive rotation sheet - FranckKessie's rotation planner
FPLAnalysis - FranckKessie's FPL data sheet

[Mav3rick]

Awesome Welcome back everybody! Even just the price reveals today have been enough to build the anticipation levels for me, a weekend of tinkering now it's not 38 degrees in my house would be great (so come on FPL release the game this afternoon for us all )

[Sutter Kane]

Whoa Stem!

Nice categorizing, informative indexing, cross-referencing, etc!

[Neath boy]

After a short break to clear my head it’s time to start again.

[DavidLloydIsAHero]

So early thoughts given the fixture schedule and Ben Crellin's analysis of it?

Early WC seems likely for many but I feel there could be huge value if you can navigate without it and use right at the end of the 1st half of the season before the BGWs/DGWs coming then with the 2nd WC being used later in that period or to exit it for the last 10 GWs.

[Smurphy Paw]

Afternoon all. Let’s hope none of us Catch a Crab

9D8B7297-8C11-4A4B-AAD5-12C19CA1FFC2.jpeg

[Mav3rick]

So early thoughts given the fixture schedule and Ben Crellin's analysis of it?

Early WC seems likely for many but I feel there could be huge value if you can navigate without it and use right at the end of the 1st half of the season before the BGWs/DGWs coming then with the 2nd WC being used later in that period or to exit it for the last 10 GWs.

Early wildcard is so powerful in those early weeks though, as the information levels increase while price changes fire fast. I'm almost certain you'd be taking some hits if not using it and needing to get in man city or man utd players for GW 2/3 which has to be accounted for in the benefit of a late use.

I see that as the critical choice though. The normal stance is to not to plan to use it, but if you have to use the WC by around GW6 then that's usually a beneficial time. If circumstances dictate that you don't need it then great save it for a better time but it usually means you got off to a flier. If you plan to use it, and therefore play a mini-game of perhaps 3 GWs, are you just going to get things massively wrong with a team full of which ever players have the best first 3 fixtures who turn out to be turkeys? Toxic price drops (caused by the high number of active managers playing this strategy), lack of points and a lost WC would be the worst of all worlds.

You could just pick the best normal team you can (including Man City players) then just sub them with one player and keep the cash in the bank, which is more like a traditional balanced, no risky players start and *may* let you save the wildcard (but equally may not).

The more I think about it the better that latter option sounds of course.

[DavidLloydIsAHero]

I used it GW3 I think last year, but I had a poor starting team (Barkley, Siggy, Zouma, Wickham) so had a lot of fires that didn't warrant hits individually. My TV really rose as a result and I made use of that later on.

I think I'd go with the latter option as well and reduce risk by taking no real punts elsewhere and hope to save the WC but it will take a fair bit of luck to do so.

[Pirlo's Beard]

our first task is to convince Mav3rick's to make his return permanent.

If that doesn't do it, nothing will.

[Mav3rick]

Don't make me look at meerkats, I still haven't recovered from the death of Shakespeare in Meerkat Mannor. #trauma

[baganboy]

yes yes yes yes yes we are back!

[Neath boy]

My first thought was to pick the best 11 including 3 Man City and Man United assets and just rely on 3 bench players in GW 1 for two pointers and save the WC for later.

Or maybe just two in case one of my squad doesn’t play first gw, and I might be better to choose a big hitter place holder for gw1 to get a third City/Man United player in.

I need all the prices and fixtures to refine my thoughts on this approach.

[Hoos01]

Hi All! Joined Fiso towards the tail end of last season. Looking forward to improving my rank with all of you in my first full season here. Cheers

[RomynPG]

[Mo Bot]

My initial thoughts are to keep the wildcard until after the transfer window is closed.

As for squad structure, I will be very much moving towards what used to be called a Widdecombe.

[Mo Bot]

And fine opening post Ste. Have you done anything else since the end of the season?

[FranckKessie]

Game on lads. Hope you all will prosper this year!

"It is, however, the time of year we can rejoice in rediscovering our beacon in the dark; our first task is to convince Mav3rick to make his return permanent. "

I can't imagine anybody chosing to pay for access to content on FPL-site without making a team Don't keep us in the dark Mav. Have you found a way of making FPL joy-full again?

[Bobby Fetta]

It seems to me that Liverpool have a massive advantage next season of not being involved in the current European games. Hence getting the 'full' summer break. Surely the non-break (or alternatively lack of pre-season) for the Manchester clubs is going to be a big disadvantage for them at some point? Solsjaer in particular has basically picked the same team for every match since the restart - with the players looking more and more knackered. I know its difficult to defend a title but this is an unusual situation.

[dino1980]

I'd like to nominate an additonal resource I found a few months ago. It's not FPL specific, but is stats baesd, and in the same vein as Statsbomb/The Double Pivot.

It's called No Grass In The Clouds, there's a weekly newsletter, sent via email, and occasional podcast. Sign up for that is here - https://nograssintheclouds.substack.com/ - if you click the 'let me read it first' button and scroll down a bit, then the 'Everything I Got Wrong (and Right) About the 2019-20 Premier League Season' post from Jul 28 is a good intro to the sort of content they do.

Quote of choice: "The Bruno Fernandes Effect is a bit overblown; their xG differential barely budged after they signed him."

[thebillfella]

Quote of choice: "The Bruno Fernandes Effect is a bit overblown; their xG differential barely budged after they signed him."

That's a big call! This is exactly why the single most important factor for me is the eye test, with stats being used to validate you views rather than driving them alone.

[dino1980]

Hey Bill, congrats on your amazing finish last season.

I think you've misinterpreted the quote. The quote is talking about United on a team level, not Fernandes on an individual player level.

Without Fernandes
25 games
xG - 37.9 (36 actual goals)
xGA - 22.2 (29 actual goals against)

1.51 vs .89 (differential of .62 per 90)

With Fernandes
13 games
xG 21.6 (30 actual goals)
xGA 13.7 (7 actual goals against)

1.66 vs .1.05 (differntial of .61 per 90)
(All figures from https://fbref.com/en/squads/19538871/20 ... ier-League)

So they just ran really cold for the first two thirds of the season agains their expected numbers and then ran like Usain Bolt for the final third. Despite running like, erm, John Regis for the entire season. This to me is why stats are better than the eye test, the eyes (well the brain) can lie, have built in biases, be looking for confirmation of a pre determined call etc)

There's no doubt that Fernandes is a fantastic FPL asset and that he helped make United a better attacking force, that much is obvious from the eye test. You felt they had better balance, had a way of attacking that wasn't just counter attacking, but of equal importance in my opinon was that Rashford, Martial and Pogba were all fit, all the time. Plus the emergence of Greenwood gave them a better balance of attacks rather than a lopsided attacking unit that focused on the left hand side via Rashford and Martial's link up play.

[thebillfella]

Thanks dino.

Yes, get all that and appreciate it the comment was about the team rather that Bruno alone. But that was my point; the team was collectively playing far better with Bruno in the team than they were without - far more pleasing on the eye and stirring other players such as Pogba and Martial to raise their game. In fact, I think the only player that got worse since he arrived was Rashford!

[raoul]

Boo hiss.

I wanted Statistics Trounced Covidly.

[dino1980]

Thanks dino.

Yes, get all that and appreciate it the comment was about the team rather that Bruno alone. But that was my point; the team was collectively playing far better with Bruno in the team than they were without - far more pleasing on the eye and stirring other players such as Pogba and Martial to raise their game. In fact, I think the only player that got worse since he arrived was Rashford!

Right, but despite playing far better (to the eye test) the metrics show that they were playing at exactly the same level.

Are they the watch-paint-dry-counter-attack-bore-offs?

Or the swash-buckling-bacalhau-fuelled-attacking-demigods?

I contest they're somwhere inbetween, which is exactly where their numbers sit.

But to get this back to FPL and Fernandes, which is the only reason I look at these numbers, if United ran hot during his 13 game spell, and Bruno ran hot during this 13 game spell (which he did), is it possible that his 10.5m price tag, which on first glance seems fair, is actually overpriced?
Bruno's xG and xA numbers p90 are almost indentical to Pulisic's numbers, yet the American is 2m cheaper. This season Pulisic varied from must have, to useless, back to must have again.

[Sutter Kane]

Bruno's xG and xA numbers p90 are almost indentical to Pulisic's numbers, yet the American is 2m cheaper. This season Pulisic varied from must have, to useless, back to must have again.

Yeah Pulisic is a steal although injury prone.

Don't we have to take into account fixtures for Man U post-lockdown when discussing xg? Playing Accrington every week is going to inflate those numbers and make Fernandes look better than he is. Being on pens is a big thing though and I'm sure he'll spend a fair bit of time in my team this year. Also he's a nice set and forget as he plays every game which makes him a viable captain sometimes, unlike many of the City lads.

[Mav3rick]

What's interesting with Fernandes price is that he's kind of fitting into that traditional "dead spot" (as it's been described) where his price is encroaching on the high side unless regularly captained.

I'll reserve judgement untill all prices are known, but my feeling is it will be difficult to have him unless he's your second super premium, which I don't think is a bad idea given the traits SK has noted above.

Of the prices revealed so far, I'm probably most intrigued with the Chelsea options in Werner and Pulisic. A traditional striker setup of super premium (Auba?), premium (Werner?), budget might be possible for the first time in a while, but ideally there will be a pickable 5.5 forward to support it.

I do wish Werner was 0.5 less though as a 9.5 premium striker and 6.5 budget striker is generally speaking 1.0-1.5 too expensive to find balance in a starting squad.

[Sutter Kane]

Auba's another set and forget type player, along with Fernandes, Salah, Pulisic, Werner, Trent, KDB...it's getting ridiculous.

A lot of people will target a midfield with the above. Budget may not allow though and you can only captain one of them. Might have to be a case of using a transfer now and again to swap between Auba+Fernandes based on form/fixtures/captaincy. First wildcard may not be that important!

Just to note, imo Ings could well outscore Werner this season, or get close.

[Aldershot Rejects]

Bruno's xG and xA numbers p90 are almost indentical to Pulisic's numbers, yet the American is 2m cheaper. This season Pulisic varied from must have, to useless, back to must have again.

Yeah Pulisic is a steal although injury prone.

Don't we have to take into account fixtures for Man U post-lockdown when discussing xg? Playing Accrington every week is going to inflate those numbers and make Fernandes look better than he is. Being on pens is a big thing though and I'm sure he'll spend a fair bit of time in my team this year. Also he's a nice set and forget as he plays every game which makes him a viable captain sometimes, unlike many of the City lads.

Pricing Fernandes must have been extremely difficult in that we only have a few games to base the price on during which, as Dino says, he had a purple patch in a team having a purple patch (don't forget Greenwood's G vs xG was completely unsustainable) against poorer side. His price allows for a 20-25% fall off which seems fair to me. I suspect he might turn out to be a tad over-priced but then again I wouldn't bet on that.

Pulisic does seem cheap by comparison and will I suspect be the standout at that price, but then again he appears to run hot and cold (as well as being injury prone) - is he a better bet than Werner at 9.5m?

[Mav3rick]

Yeah I mean it seems like the call is two super premium mids and then rotating your captain options according to form/fixtures.

Aubameyang, Liverpool mids, city mids, Fernandes can all occupy those spots at different times, but having 3 will surely be a stretch without playing 4 defenders for the most part.

Wood as the only striker maybe? Given how attacking Chelsea were last season I will still be tempted with Werner if he's played through the middle, but that's one to look at as pre-season progresses.

[Mav3rick]

Lowest team ID I've ever had - 129

Let the planning commence!