FPL Prediction Tool: Gameweek 2

Welcome to our new weekly series where we’ll be presenting the results of our FPL prediction tool to show you the clean sheet and goal probabilities of the players for Gameweek 2

Table explanations

Clean sheet percentage (CS percentage) – percentage probability that a team for which the player plays will keep a clean sheet; calculated from average odds from bookies

One goal percentage – percentage probability that a player will score exactly one goal; calculated from average odds from bookies

Two goals percentage – percentage probability that a player will score exactly two goals; calculated from average odds from bookies

Hat-trick percentage – percentage probability that a player will score exactly three goals; calculated from average odds from bookies

Clean sheet expected points (CS exp) – cs percentage multiplied by points awarded for cs for players position (4 for GK and DEF, 1 for MID)

One goal expected points (One goal exp) – goal percentage multiplied by points awarded for a goal for a players position (4 for FWD, 5 for MID, 6 for DEF)

Two goals expected points (Two goals exp) – two goals percentage multiplied by points awarded for a goal for a players position (4 for FWD, 5 for MID, 6 for DEF) and by number of goals (2)

Hat-trick expected points (Hat-trick exp) – hat-trick percentage multiplied by points awarded for a goal for a players position (4 for FWD, 5 for MID, 6 for DEF) and by number of goals (3)

Clean sheets and goals expected points (CS & goals exp) – sum of clean sheet expected points, one goal expected points, two goals expected points and hat-trick expected points

First team expected points (First team exp) – 2 points if was in the line up last week and is expected to be in the next one

Total expected points (Total exp) – sum of clean sheets and goals expected points and first team expected points

Top 5 Goalkeepers 

GW2 GK

Top 10 Defenders

GW2 DEF

Top 10 Midfielders

GW2 MID

Top 6 Forwards

GW2 FWD

Full spreadsheet link with all the players

Download yourself a copy here >>  fpl 18-19 gameweek 2

Summary

Defensive perspective

This week it is Manchester City (71.94%), and far behind them Tottenham (54.43%) in second place, that provide the best chance for a clean sheet.

Ederson is a great option but takes one of the spots reserved for City players and we’ll likely already have 3 spots used and if not, Mendy would be the best choice from a defensive perspective.

Lloris looks good, especially if you have no Spurs defenders, with him not subject to rotation, but Davinson Sanchez caught the eye after his display saw him have 3 goal attempts, one of them ending up in an assist for Verthonghen.

Elsewhere, Man Utd (49.93%) look good for a clean sheet against a Brighton side that looked extremely lacklustre going forward on the weekend.

Embed from Getty Images

 

Despite Chelsea only having a 37.80% chance at a clean sheet, Marcos Alonso features in the top 10 defenders for expected points this week. Even in Sarri’s new system, which sees him operate in a back 4 rather than as a wing back as part of a 5, he maintained his attacking threat on the weekend which was evidenced by his assist and the fact he hit the bar with an attempt.

Attacking Perspective

Embed from Getty Images

 

Mohamed Salah (7.28) comes out with the highest expected points return of any of the categories, rivalled only by Sergio Agüero, who runs him very close with an expected return of 7.20 – the main differentiator coming down to the fact that Salah gets extra points over Agüero simply for being a midfielder, despite Agüero actually being more likely to score than Salah, and being more likely to score a hat-trick than Salah.

In the top 10 midfielders, City feature 4 players, with Sterling (ranked 2nd), Mahrez (ranked 3rd), Bernado Silva (ranked 4th) and Leroy Sané (ranked 10th) all predicted to return well against Huddersfield – Sané being the slightly more risky pick after him missing out last Gameweek.

Elsewhere, the Spurs boys are well represented in the form of Kane (ranked 3rd overall), Eriksen (ranked 7th), Lucas Moura (ranked 8th) and Alli (ranked 10th) who play Fulham and are predicted to do well. Kane appeared to struggle against Newcastle, but with that game and another weeks training under his belt, he should be approaching 100% match-fitness and this is a great fixture for Spurs’ talisman.

Alexis Sánchez (ranked 9th overall) is the only one outside of Spurs, Liverpool and Man City to feature in the top 10 midfielders and is set to return well against Brighton according to the bookies.

Outside of Kane and Agüero for the forwards, it seems that Jamie Vardy could provide his owners with decent returns and should be in line to start after coming on last week and notching, and with another weeks training under his belt.

Predictor Tool Team

We have built our team basing on our expected points calculations. We wanted to incorporate an element of realism with this, as we know we’re unable to create a new team every week to take advantage of the results of our tool, so we are managing it ourselves, but the spine of the team was built based on the results of our predictor tool.

We will make transfers each week as normal and will place the captain armband on whichever attacker (MID or FWD) has the highest expected score, as per the results of our tool.

Gameweek 1 score

So, we got 97 points in GW1 which placed us inside the top 25k – not a bad start!

Our front line were the only players not to deliver for us, with Agüero and King both blanking. Fortunately the rest of the squad did better than was expected – almost every one of our defenders got some attacking returns.

The midfield supplemented the defence nicely with 4 goals, 1 assist and 3 bonus points collectively.

Gameweek 2 team

This week, we have some of the top choices in each position.

Salah is expected to get the most points this week, so he is our captain choice. Vice captain is Agüero with second most points expected this GW.

King was disappointing last week, but according to the tool, he should be picked before Jimenez and we hope this week he will show what he is capable of.

According to our calculations, that team should get at least 54.95 points (not including saves, assists and bonus points, but also not including cards and some other factors like penalties saved/missed).

We will analyse our teams’ result next week – let’s see if it does as well as it did this week!

How it’s all calculated 

Basically, we are calculating expected points for every player for the upcoming Gameweek using the odds from the bookies.

As we know, points are mostly based on playing time, clean sheets, saves, assists, goals, bonus points and yellow/red cards. There are also some other factors like saved penalties, but they are more rare and harder to predict.

As our tool is still in development, at first, we would like to provide you with expected points for playing time, clean sheets and goals.

We base our calculations of the above on line-ups used in the last GW as well as average bookies odds published for each game (clean sheets odds and goalscorer odds). We first convert those odds to percentages and then use them to calculate expected points for every single player.

In our opinion, bookies odds are the best measure as they are the most reliable – if anyone was able to build a better tool to predict scorers and clean sheets then they would just exploit the odds and get rich. So bookies odds can be considered the best benchmark.

Clean sheets points are based on odds for wining to nil & 0-0 draw. Goalscorer points are based on anytime goalscorer odds, two or more goals odds and hat-trick odds. That means probability that a player will score during the match, that a player will score at least two goals and that a player will score a hat-trick (exactly three goals).

We first use hat-trick odds, convert it to percentages, multiply it by number of points for a goal per position and by number of goals (3). Then we use two or more goals – probability to score 4 or more would be quite low, so we assume that a player will score maximum of three goals during the match. In this case odds that a player will score exactly three goals are equal to odds to score two or more decreased by odds for a hat-trick. That way we have percentage probability that a player will score two goals, then we multiply it by number of points for a goal per position and by number of goals (2).

In the last step we use anytime goalscorer odds, convert it to percentages, decrease it by percentage for two or more goals. That way we have percentage probability of scoring exactly one goal, multiply it by number of points for a goal per position and by number of goals (1).

All the data used for calculations were taken from: oddschecker.com; if some odds were not provided, then data was taken from WilliamHill.com

Line-ups were based on last GW line ups. 

If you want to get in touch with our data analyst or you have any further questions about this article or the tool, then please follow Piotr on Twitter here >> https://twitter.com/Mikhatharahttps://twitter.com/Mikhathara

Food_for_thought_visual_identity-800x400

 

 

6 thoughts on “FPL Prediction Tool: Gameweek 2

Add yours

  1. Useful analysis, but I see the absence of taking assists into account as a flaw. Doesn’t seem right to have Laporte and Stones ranked above Mendy and Walker,for example.

    Like

    1. Hi, thank you for your insight! I agree with you and that is why we are trying to continue developing our tool! Unfortunately there are no bookies odds for assists and other factors so we need to build our own model first – that is why this data is not ready yet. However we’ve decided that we want to share with you at least some of the calculation we are able to do as for now – in our opinion it is still really helpful!

      Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

Blog at WordPress.com.

Up ↑

%d bloggers like this: