Thursday, 4 October 2012

GW7 Select XI



 

This is my statistically optimised team for GW7.  This is determined based on current team form and the players underlying stats.  Usually my team is very similar to those picked for teams around other FPL websites.  

This time I think there may be a couple of noticeable omissions, most notably Bale (or any Totteham) at home to Villa, Hazard vs. NOR (H) who is currently leading The Scout's captain poll, and one of Tevez or Aguero up front I suppose.  The absence of Bale in particular made me reassess the underlying stats but I'll stick with it (I think Spurs will score 2 past Villa whilst City will get 3 against Sunderland with Silva more likley to be involved overall).

This team costs a rock bottom £66m and I am predicting somewhere around the 80 pt mark.  I hope Lukaku and Oscar get to start!  If they don't then Hazard would come in as next best pick and I'd probably just be "safer" going with him, but I won't.
 

14 comments:

  1. Lukaku C is an interesting choice considering his limited gametime.

    ReplyDelete
  2. And no Tott players vs Villa at home?

    ReplyDelete
  3. I'm not particularly bothered about the (C) pick, but, I think Lukaku will get more games at home and if he does he could tear QPR a new one. I explained why I had no Bale - I fancy Silva and Oscar over him. I should probably swap Ba for Bale and have the funds to do so. But I will stick with my guns. Are you doing a Team of the Week? Link it!

    ReplyDelete
    Replies
    1. Yes I am, just waiting for the press conferences to see on the fitness of some players. Berba would be first name on my teamsheet if fit. So sometime tomorrow. I will have some similar picks for sure- probably an identical defense at least in terms of defenses represented. And Morrison instead of Lukaku for Brom offense. Hazard will also get a shout (I think he will get his first open play goal this GW)
      Keep up the good work.

      Delete
  4. I agree the Lukaku choice is interesting. I would guess you are not factoring in rotation risk into your picks?

    One thing I do notice is that you are actually picking a team that fits in a budget. The FFS team is great and all but usually expensive as all Hell. At least this is realistic.

    ReplyDelete
  5. FFScout's aren't so pricey, take a look at posts labelled 'Dream Team' where I have compared various site's picks. http://www.shotsontarget.co.uk/2012/09/around-web-gw6-dream-teams.html

    Lukaku though, yes an odd and audacious pick. I am not factoring in rotation risk or reduced minutes. This pikc is based purely on him getting 90 mins this weekend, and not a long-term suggestion at all. But his stats are great. Look at WBA's page and see how he monopolises their goal efforts despite less mins than his ltd mins. http://www.shotsontarget.co.uk/2012/08/west-brom.html One factor I might have overlooked is that when he is used as a late sub he comes on to a different style of game than players who start the 90.

    I have some ideas on how to factor this in but not really worked it through fully. My 1st thought would be to only count his stats when he plays 60+ mins. He's done this once this season against Reading and in that game had 1 of WBA's 5 Shots in the Box and one of their Shots on Target, scoring the game's only goal. His stats for this single game are worse than when he makes a sub appearance.

    Any suggestions on this aspect would be most welcome. It might be best to just ignore him completely til he has more "reliable" numbers.



    ReplyDelete
  6. One other thing...really surprised not to see Suarez here instead of Ba. MUN hasn't been great defensively, but they have played a fairly tough attacking schedule and NEW hasn't been generating a ton of offense. I would be pretty surprised to see Ba get more than one. In addition, bonus points might be difficult to come by unless NEW can win (which I don't think they will). I would be more confident with Suarez as Liverpool are huge favorites at home and Suarez has been in superb form.

    ReplyDelete
  7. I was suprised to see Suarez not in the lineup too. However I am selecting based on my model. I cannot bias it with my own opinion. However, when a clearly valid somission like Suarez or Bale raise's it's head it's worth a closer look.

    Right now my model has Liverpool performing offensively at home at about "average" ~ i.e 1.6 GF / 5 SoT per game with results so far of (goals:SoT) MCI 2:3, 0:5 ARS, 1:6 MUN. Avg=4.6 SoT/g. Stoke's away defence so far has been half a deviation better than average, conceding RDG 1:2, Wig 2:3, Che 2:4. Avg SoT against/g = 3. for both team's there is some baseline form included from last season too.

    So Liv Att are 0, Sto Def are -0.5 yielding an expected SOT for Liv of 4.2. From this I'd expect 1-2 goals really, and of those Suarez stats would say he'd be involved in 48% (best in league), so I have down for 1.3 goals + 0.4 assists, or about 7 FPL pts.

    For Ba, United's poor defensive form has New expected to have 5 SoT with Ba at about the same level of goal potential as Suarez (42%) meaning he comes out on top, just, with an expected 8 pts. Bonus Points have not been included very well if I'm on honest and are just based on a study IU did last year where on average a goal is worth 1.2 BP. There's work to be done here but it does get quite complex.

    I hope this explains my decisions in a little more detail and I stnad by choice of Ba > Suarez for GW7. HOWEVER (again) we are talking about such small margins of difference here - it's the difference between rolling two people rolling a dice each, one has to get a 6 to win, the other a 5 or 6. On one roll - eiter can win, but over 100 rolls there will be only one winner.

    Adding expert opinion into the mix to suppoirt the numerical analysis is of course the ideal but it's then gets hard to incorporate factors such as these teams involvment in Europe this week.

    No typos were harmed in the makin gof this reply :)

    ReplyDelete
  8. While I have yet to construct a player model, my working team algorithms suggests the following goals scored:

    ARS 1.34
    AVL 0.75
    CHE 2.13
    EVE 1.70
    FUL 2.14
    LIV 1.87
    MCI 2.30
    MUN 1.22
    NEW 1.28
    NOR 0.96
    QPR 0.98
    RDG 1.05
    SOT 2.27
    STO 0.84
    SUN 0.22
    SWA 2.29
    TOT 1.87
    WBA 2.16
    WHM 1.26
    WIG 1.56

    I can say that I am currently not adjusting for individual home/away performance (have a default for the entire league) nor I am adjusting for what you call Conversion/Resilience (lovely defensive names, btw). I am presuming that all teams convert at about the same ratio. I know that is not correct, but I am not yet confident I can predict what is correct so feel better simply removing individual team adjustments for conversion.

    Looking at these numbers, they all look "pretty" good. I am a bit concerned with the Wigan prediction and feel the Spurs have more than 2 goals in them. With that said, a couple interesting thoughts I have on the week:

    1. I am 98% confident City finally gets a clean sheet this week. So much so I am seriously considering pulling Zabaletta in for a point hit (I really need some backups with Guthrie, Tevez rotation doubts and I think Toure may be rested). Plus, their upcoming fixtures look fantastic: I have them conceding more than 1 goal in only 1 of the next 6 weeks based upon current ratings.

    2. Southie and Fulham should be a blast to watch, especially if Berbatov plays. I am seriously considering captaining Lambert over Suarez, my current captaining choice (on my team, not among all players).

    3. Chelsea is going to have a monster game. Unfortunately, I am not confident in any of their attacking players. If I had to take somebody, I would probably take Hazard, but there is simply too much unknown with the 3 10s to make a prediction right now. Another thought...I know RVP wasn't seriously linked to Chelsea, but imagine him in that lineup? Torres just isn't good right now.

    4. I would take those Swansea numbers with a huge grain of salt based upon current form. Swansea has looked awful and Reading looked much, much better last week (they were easily the better team against Newcastle). Still though Michu still looks good to my untrained eye; I wouldn;t drop him if I owned him.

    5. Norwich is going to have a fantastic holiday season (Snodgrass has quietly been brilliant and will be on my team by the end of the month). Seriously. With that said, they have virtually no shot at an somewhat overrated Chelsea team.

    ReplyDelete
  9. What do you think of West Brom? I have almost completely ignored them for some reason (I think I've seen about 30 minutes of them all season). You feel their offense is for real?

    ReplyDelete
  10. @JohnDoe

    I agreee with your pts 1- 5 in the main. I've not looked too far ahead at Norwich so cannot comment. I'm with you entirely on the City clean sheet and would love to bring in CClichy but don't know if he'll start with Mancini's latest comments.

    Here my projected GF for the weekend, alongside yours...

    ARS 1.3 1.5
    AVL 0.8 0.8
    CHE 2.1 3.0
    EVE 1.7 1.6
    FUL 2.1 2.0
    LIV 1.9 1.4
    MCI 2.3 2.3
    MUN 1.2 1.3
    NEW 1.3 1.8
    NOR 1.0 0.8
    QPR 1.0 0.6
    RDG 1.1 1.3
    SOT 2.3 2.4
    STO 0.8 0.7
    SUN 0.2 0.4
    SWA 2.3 2.4
    TOT 1.9 2.0
    WBA 2.2 2.3
    WHM 1.3 1.4
    WIG 1.6 1.7

    0.91 correlation with the largetst difference being I picking CHE to score 3 to your 2, and you've got LIV to 1.9 where I have 1.4 (and hence why I have omitted Suarez from this weeks top picks, where as may not have).

    My predictions are bsaed on SoT only with all team's scoring an average 1 goal to 3 SOT. I have not yet tried to incorporate my new form indicators, like Creativity & Penetration., I'll have a stab at doing this before GW7, but for GW8 dfeinietly.

    If you want to see how I'm positioning each team's form on a scale see my Team Form guide.

    As for WBA - I have them as pretty much performing at average offensively, which is 1.6 GPG. This visit of QPR is really the reason for Lukaku in this week's pick.

    ReplyDelete
  11. I've continued this discussion on predicting the GW results in the forum. http://fpl-analytical.1086037.n5.nabble.com/GW7-GPG-predictions-tp24.html

    ReplyDelete